Author Archives: Lex Fridman

#462 – Ezra Klein and Derek Thompson: Politics, Trump, AOC, Elon & DOGE

Podcast: Play in new window | Download

Subscribe: Spotify | TuneIn | RSS

Ezra Klein is one of the most influential voices representing the left-wing of American politics. He is a columnist for the NY Times and host of The Ezra Klein Show. Derek Thompson is a writer at The Atlantic and host of the Plain English podcast. Together they have written a new book titled Abundance that lays out a set of ideas for the future of the Democratic party.
Thank you for listening ❤ Check out our sponsors: https://lexfridman.com/sponsors/ep462-sc
See below for timestamps, transcript, and to give feedback, submit questions, contact Lex, etc.

Transcript:
https://lexfridman.com/ezra-klein-and-derek-thompson-transcript

CONTACT LEX:
Feedback – give feedback to Lex: https://lexfridman.com/survey
AMA – submit questions, videos or call-in: https://lexfridman.com/ama
Hiring – join our team: https://lexfridman.com/hiring
Other – other ways to get in touch: https://lexfridman.com/contact

EPISODE LINKS:
Abundance (new book): https://amzn.to/4iZ1S8J
Ezra’s X: https://x.com/ezraklein/
Ezra’s Instagram: https://instagram.com/ezraklein
Ezra’s YouTube: https://youtube.com/EzraKleinShow
The New York Times: https://nytimes.com/by/ezra-klein
Derek’s X: https://x.com/dkthomp
Plain English (podcast): https://www.theringer.com/podcasts/plain-english-with-derek-thompson
The Atlantic: https://theatlantic.com/author/derek-thompson/

SPONSORS:
To support this podcast, check out our sponsors & get discounts:
Call of Duty: First-person shooter video game.
Go to https://callofduty.com/warzone
LMNT: Zero-sugar electrolyte drink mix.
Go to https://drinkLMNT.com/lex
AG1: All-in-one daily nutrition drinks.
Go to https://drinkag1.com/lex
Shopify: Sell stuff online.
Go to https://shopify.com/lex

OUTLINE:
(00:00) – Introduction
(03:17) – Sponsors, Comments, and Reflections
(10:33) – Left-wing vs right-wing politics
(19:54) – Political leaders on the left and the right
(44:29) – Internal political divisions
(47:29) – AOC
(58:50) – Political realignment
(1:10:32) – Supply-side progressivism
(1:17:42) – Wealth redistribution
(1:27:50) – Housing problem
(1:44:09) – Regulation and deregulation
(2:00:43) – DOGE, Elon, and Trump
(2:59:46) – Sam Harris
(3:09:24) – Future of America

PODCAST LINKS:
– Podcast Website: https://lexfridman.com/podcast
– Apple Podcasts: https://apple.co/2lwqZIr
– Spotify: https://spoti.fi/2nEwCF8
– RSS: https://lexfridman.com/feed/podcast/
– Podcast Playlist: https://www.youtube.com/playlist?list=PLrAXtmErZgOdP_8GztsuKi9nrraNbKKp4
– Clips Channel: https://www.youtube.com/lexclips

Transcript for Ezra Klein and Derek Thompson: Left vs Right Politics, Trump, Elon & DOGE | Lex Fridman Podcast #462

This is a transcript of Lex Fridman Podcast #462 with Ezra Klein and Derek Thompson.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Episode highlight

Ezra Klein
(00:00:00)
Democrats still think the currency of politics is money and the currency of politics is attention. And that’s a huge difference between the two sides right now.
Derek Thompson
(00:00:07)
I think the steel man is very easy to make here. Department of government efficiency. That sounds like an organization that’s needed if government is inefficient. And one of the themes of our book is just how inefficient government can be, not only at building houses, building energy, often at achieving its own ends. Building high-speed rail when it wants to build high-speed rail. Adding affordable housing units when it wants to add affordable housing units. I love Ezra’s line that we don’t just need to think about deregulating the market. We need to think about deregulating government itself, getting the rules out of the way that keep government from achieving the democratic outcomes that it’s trying to achieve. This is a world in which a department of government efficiency is a godsend. We should be absolutely obsessed with making government work well, especially if we’re going to be the kind of liberals who believe that government is important in the first place.
Ezra Klein
(00:01:02)
In my lifetime, the Democratic Party has never been as internally fragmented and weak, leaderless, rudderless as it is right now. Now, it won’t stay that way. You cannot change American politics. You can’t change the Democratic Party if you’re not willing to upset people. Donald Trump reformed the Republican Party by willing able to fight Republicans. He ran against George W. Bush, against Jeb Bush, against Mitt Romney, against the trade deals, against a bunch of things that were understood to be sacred cows. Somehow this guy ran right after Mitt Romney and John McCain, while attacking Mitt Romney and John McCain. The Democratic Party does need to change. It needs to attain a different form because the Obama Coalition is exhausted. It’s done. It’s not going to be able to do that. If doesn’t have standard bears who are willing to say, “We were wrong about some things. We have to change our views on some things. We have to act differently and speak differently.”
Derek Thompson
(00:01:58)
When Elon takes over Tesla, when Elon is at SpaceX, when Elon’s at X, I would imagine, and you know this better than me because you know him, and maybe most importantly for the purposes of this part of the conversation, you know the people who work for him. I’ll bet if you ask the people who work under Elon at X, Tesla, SpaceX, they say, “I know exactly what Elon wants. This is his goal for the super heavy rocket. This is his goal in terms of humanoid robots. This is his goal in terms of profitability of Twitter, and the growth of our subscription business and how we’re going to integrate new features.” There’s probably a really clear mind meld.

(00:02:32)
Right now, I have no sense that there’s a mind meld. And in fact I have the exact opposite sense, that rather than an example of creative destruction, which would be a mitzvah of entrepreneurship, we have an act of destruction, destruction. We have destruction for the sake of destruction. It’s much cleaner to me from an interpretive standpoint, to describe ,Doge as an ideological purge of progressivism, performing the job of efficiency rather than a department of actual efficiency itself.

Introduction

Lex Fridman
(00:03:08)
The following is a conversation with Ezra Klein and Derek Thompson. Ezra is one of the most influential voices representing the left wing of American politics. He is a columnist for The New York Times, author of Why We’re Polarized and host of the Ezra Klein Show. Derek is a writer at the Atlantic, author of Hit Makers and On Work, and host of the Plain English Podcast. Together, they’ve written a new book simply titled Abundance, that lays out a kind of manifesto for the left. It is already a controversial, widely debated book, but I think it puts forward a powerful vision for what the Democratic Party could stand for in the coming election.

(00:03:55)
If I may, let me comment on the fact that sometimes on this podcast, I delve into the dark realm of politics. Indeed, politics often devises and frankly, brings out the worst in some very smart people. Plus to me, it is frustrating how much of the political discourse is drama, and how little of it is rigorous, empathetic discussion of policy. I hate this. But I guess I understand why. If the other side is called either Hitler or Stalin online by swarms of chanting mobs, it’s hard to carry out a nuanced discussion about immigration, healthcare, housing, education, foreign policy, and so on. On top of that, anytime I talk about politics, half the audience is pissed off at me. And no, there is no audience capture. I get on equally by different groups across the political spectrum, depending on the guest. Why? I don’t know. But I’m slowly coming to accept that this is the way of the world. I try to maintain my cool, return hate with compassion and learn from the criticism and the general madness of it all.

(00:05:07)
Still, I think it’s valuable to sometimes talk about politics. It’s an important part to the big picture of human civilization, but indeed, it is only still a small part. My happy place is talking to scientists, engineers, programmers, video game designers, historians, philosophers, musicians, athletes, filmmakers and so on. So, I apologize for the occasional detour into politics, especially over the past few months. I did a few conversations with world leaders and I have a few more coming up. So, there will be a few more political podcast coming out in part so I can be better prepared to deeply understand the mind, the life and the perspective of each world leader. I hope you come along with me on this journey into the darkness of politics, as I try to shine a light in the complex mess of it all, hoping to understand us humans better, always backed of course by deep rigorous research and by empathy.

(00:06:06)
Long-term, I hope for political discussions to be only a small percentage of this podcast. If it’s not your thing, please just skip these episodes. Or maybe come along anyway, since both you and I are reluctant travelers on this road trip. But who knows what we’ll learn together about the world and about ourselves. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description. And now, dear friends, here’s Ezra Klein and Derek Thompson.

Left-wing vs right-wing politics


(00:06:38)
You are both firmly on the left of the US political spectrum. Ezra, I have been a fan of yours for a long time. You’re often referred to, at least I think of you as one of the most intellectually rigorous voices on the left. Can you try to define the ideals and the vision of the American left?
Ezra Klein
(00:06:56)
Oh, good. We’re starting small here.
Lex Fridman
(00:06:58)
And maybe contrast them with the American right.
Ezra Klein
(00:07:00)
Sure. So, the thing I should say here is that you can define the left in different ways. I think the left has a couple fundamental views. One is that life is unfair. We are born with different talents. We are born into different nations. The luck of being born into America is very different than the luck of being born into Venezuela. We are born into different families. We have luck operating as an omnipresence across our entire lives. And as such, the people for whom it works out well, we don’t deserve all of that. We got lucky. I mean we also worked hard, and we also had talent and we also applied that talent. But at a very fundamental level that we are sitting here is unfair, and that so many other people are in conditions that are much worse, much more precarious, much more exploited is unfair. And one of the fundamental roles of government should not necessarily be to turn that unfairness into perfect equality, but to rectify that unfairness into a kind of universal dignity so people can have lives of flourishing. So, I’d say that’s one thing.

(00:08:11)
The left is fundamentally more skeptical of capitalism and particularly unchecked forms of capitalism than the right. I always think this is hard to talk about because what we call unchecked capitalism is nevertheless very much supported by government. So, I think in a way, you have both. Markets are things that are enforced by government. How you set the rules of them is what ends up differing between the left and the right. But the left tends to be more worried about the fact that you could get rich building coal-fired power plants, belching pollution into the air, and you could get rich laying down solar panels. And the market doesn’t know the difference between the two. And so, there’s a set of goals about regulating the unchecked potential of capitalism that also relates to exploitation of workers. There’s very fundamental questions about how much people get paid, how much power they have. Again, the rectification of economic and other forms of power is very fundamental to the left when you think about what the minimum wage is.

(00:09:11)
I am a successful podcast host when I go into a negotiation with the New York Times, I have a certain amount of market power in that negotiation because other firms want to hire me. When you are a minimum wage worker, the reason we have a minimum wage is in part to rectify a power problem. A lot of workers do not have market power. They do not have a bunch of job opportunities. They’re not working with firms. And by the way, without certain kinds of regulation, those firms would cartelize and make it so they can hold down wages anyway. So, trying to rectify power imbalances is, I think, another thing folks on the left take more seriously. That would be a start of things that I think broadly unite the… maybe let’s call it the intuitions. I want to say that’s a podcast answer, not a book. I’m sure I left a million things out here, but I’ll start there.
Lex Fridman
(00:09:57)
I mean there’s a lot of fascinating things there on the unfairness of life that could be the inter-person unfairness, so one person getting more money than another person, or more skills or more natural abilities and another person. And then there’s just the general unfairness of the environment, the luck of the draw, the things that happen. All of a sudden you cross a street, and the car runs a red light, and runs you over and you’re in the hospital. So, that unfairness of life. And in general, I guess the left sees there’s some role or a lot of role for government to help you when that unfairness strikes. And then maybe there’s also a general notion of the size of government. I think the left is more comfortable with larger government as long as it’s effective and efficient, at least in its idea.
Derek Thompson
(00:10:42)
That’s certainly true in the last 200 years. It was new deal liberals who enlarged the government in the 1930s. It was Republicans who acquiesced to that larger government in the 1950s. And then starting the 1970s, 1980s, it’s typically been conservatives who’ve tried to constrict governments. Sometimes they failed while liberals have typically tried to expand, certainly taxing and spending. But one thing that I was thinking as Ezra was talking, and I was just writing this down because I thought Ezra’s answer was really lovely, but at a really high level, I thought… maybe you disagree with this. I thought about distinguishing between liberals and conservatives based on three factors, what each side fears, what each side values and what each side tolerates. I think liberals fear injustice, and conservatives often fear cultural radicalism or the destruction of society. And as a result they value different things.

(00:11:37)
Liberals I think tend to value change. And at the level of government that can mean change in terms of creating new programs that don’t previously exist. It’s typically been liberals, for example, who’ve been trying to expand health coverage, while conservatives have tried to cut it back. Just in the last few years, it was Biden who tried to add a bunch of programs, whether it was infrastructure, the Chips and Science Act, the IRA, and then Trump comes into office and is unwinding it.

(00:12:02)
And then I also think they tolerate different things. I think liberals are more likely to tolerate a little bit of overreach, a little bit of radicalism in terms of trying to push society into a world where it hasn’t been. Well, I conservatives are more likely to tolerate injustice. They’re more likely to say there’s a kind of natural inequality in the nature of the world and we’re not going to try to overcorrect forward with our policies. And so, I think that even at a layer above what Ezra was articulating with the policy differences between liberals and conservatives, there’s almost like an archetypal difference between what they fear, and value and tolerate. Liberals fearing injustice, seeking change, tolerating sometimes a bit of what people might think of as overreach, while conservatives fear that overreach, value tradition and often tolerate injustice.
Ezra Klein
(00:12:54)
The only thing I would say is that I do think this sort of the left likes big government, the right like small government oversimplifies. The left is pretty comfortable with an expansive government that is trying to correct for some of the imbalances of power and injustices and imbalances of luck I talked about earlier. The right is very comfortable with a very powerful police and surveillance in national security state.

(00:13:18)
I always think about the sort of George W. Bush era, although right now with ice agents hassling all kinds of green card holders, you can think about this moment, too. But the right’s view that on the one hand the government is incompetent. And on the other hand we could send our army across oceans, invade Afghanistan and Iraq, and then rebuild these societies we don’t understand into fully functioning liberal democracies that will be our allies, was an extraordinary level of trust in a very big government. I mean that was expensive. That took manpower. Compared to we’re going to set up the Affordable Care Act in America, that took a lot more faith in the US government being able to do something that was extraordinarily difficult. But the left is more confidence in the government of the check and the right has more confidence in the government of the gun.
Lex Fridman
(00:14:06)
You’re right, there’s some degree to which what the right, when the right speaks about the size of government, it’s a little bit rhetoric and not actual policy because they seem to always grow the size of government anyway. They just say small government, but they don’t… it’s in the surveillance state, in the foreign policy in terms of military involvement abroad, and really in every program, they’re not very good at cutting either. They just kind of like to say it.
Derek Thompson
(00:14:38)
Cutting is really hard. Government spends trillions of dollars and if you cut billions of dollars, someone is going to feel that pain and they’re going to scream. And so, you look at defense spending under Reagan, you look at overall under Reagan. Reagan might be one of the most archetypally conservative presidents of the last 40, 50 years. He utterly failed in his attempt to shrink government. Government grew under Reagan. Defense grew, all sorts of programs grew. So, I think that one thing we’re sort of scrambling around in our answers is that at a really high level, there are differences between liberalism and conservatism in American history. But often at the level of implementation, it can be a little bit messy. Even Bush’s foreign policy that Ezra was describing from a big sense of American history, is very like Wilsonian, this sense of it’s America’s duty to go out and change the world.
Ezra Klein
(00:15:31)
Or to use a current example, McKinleyan.
Derek Thompson
(00:15:32)
Or McKinleyan. Right. And a lot of people compare Donald Trump’s foreign policy to Andrew Jackson. This sense of we need to pull back from the world. America first. We need to care about what’s inside of our borders and care much less about what’s outside of our borders. Sometimes the differences between Republican and Democrat administrations don’t fall cleanly into the lines of liberal versus conservative because those definitions can be mushy.

Political leaders on the left and the right

Lex Fridman
(00:15:59)
All right. So, to descend down from the platonic ideals of the left and the right, well, who’s actually running the show on the right and the left? Who are the dominant forces? Maybe you could describe and you mentioned democratic socialists, the progressives, maybe liberals, maybe more sort of mainstream left, and the same on the right with Trump and Trumpism.
Ezra Klein
(00:16:24)
So, on the right, it’s pretty straightforward at the moment. And the right is composed differently than it was 10 years ago. But the right is run by Donald Trump and the people who have been given the nod of power by Donald Trump. So, that is right now Elon Musk, but Elon Musk’s power is coming from Donald Trump. That is maybe in some degrees JD Vance, maybe in some degrees, Russ Vought, maybe sometimes Homan’s over at DHS. The right beneath that, the Republicans in Congress are extraordinarily disempowered compared to in other administrations. They are sort of being told what to do and they are doing what they are told. Republicans in Congress, Senate Republicans, they didn’t want Pete Hegseth. They didn’t want Kash Patel. They didn’t want Tulsi Gabbard. They didn’t want RFK Jr. Nobody got elected to be a Republican in the Senate hoping that they would confirm Robert F. Kennedy Jr, a member of the Kennedys, a Democrat who is pro-choice and running as a Democrat two years ago for HHS. But Donald Trump told him to do it and they did.

(00:17:23)
So, the right has developed a very, very top-down structure. And one of Trump’s talents, one of the things that makes him a disruptive force in politics is his ability to upend the sort of coalitional structure, the interest group structure that used to prevail. The Koch brothers were the big enemy of the left 10, 15 years ago. The view was in many ways they set the agenda of the right. The Koch brother network is much less powerful Donald Trump because he just disagrees with them and has disempowered them. Not to say none of their people or none of their groups are meaningful at all. They are, but you wouldn’t put them at the forefront in the way that you might’ve had another time. Right this second, we’re using the left, but Democrats are in fundamental disarray. There is no leader. Democrats, Senate Democrats decided to vote for the containing resolution, avoiding a shutdown or a critical mass of them. Then Hakeem Jeffries is the leader of the House Democrats and Chuck Schumer, the leader of Senate Democrats are in bitter disagreement over whether or not they should have done that. Democratic leadership isn’t even united on the single biggest point of leverage they might’ve had. They disagree over whether or not it was even a point of leverage. Outside of them, the party has no leader, which is fairly normal after a pretty crushing defeat. But there isn’t the next in line.

(00:18:40)
So, you go back and it was pretty clear that after Barack Obama, it was going to be Hillary Clinton. After Hillary Clinton, it was either going to be Joe Biden or Bernie Sanders. Bernie Sanders had come in second in the primary. Joe Biden had been the vice president. You often have a presumptive next nominee who the party can look to for kind of leadership. Even after 2000, Al Gore was still giving big speeches. There was a question about Al Gore running again. There is no presumptive in the Democratic Party right now. You can’t turn around and say, “Oh, it’s going to be Pete Buttigieg. It’s going to be Josh Shapiro. It’s going to be Gretchen Whitmer.” Parties are given force, modern parties, which are quite weak by historical standards. Modern parties tend to be given force by a centralizing personality. Donald Trump being a very strong example of that on the right, but Barack Obama was the person who held together the Democratic Party for a long time.

(00:19:33)
In my lifetime, the Democratic Party has never been as internally fragmented and weak, leaderless, rudderless as it is right now. Now, it won’t stay that way. There’s a rhythm to these things. There’ll be a midterm, they’re probably going to pick up a bunch of seats in the midterm. If that means Hakeem Jeffries becomes speaker after the midterm, he’s going to have a much louder voice because he’s going to have power. It’s going to be a harder road for Schumer to get back to the majority because of the Senate map, and then we’ll start having a primary on the left. And you’ll begin to see voices emerge out of that.

(00:20:09)
But right now, the Democratic Party, it doesn’t have points of power. There’s simply outside of at the national level, there is no Democrat who wields control over a branch of government. They don’t have the Supreme Court, they don’t have the House, they don’t have the Senate, they don’t have the presidency and they don’t have a next in line. So, you’re looking at an organization without any of the people in the position to structure it, and the head of the DNC, the new head, Ken Martin, doesn’t have power in that way. So, they’re pretty fractured.
Lex Fridman
(00:20:41)
You got a lot of criticism for this, but you were one of the people that early on said that Biden should step down. Why is the Democratic Party at this stage in its history, so bad at generating the truly inspiring person? To me personally, AOC is an example of a person that might be that person.
Ezra Klein
(00:21:06)
You should have her on the show. I would watch that.
Lex Fridman
(00:21:07)
Definitely. I really try to, and we’ll talk about this, I try to do two, three hours and there’s a hesitancy on the left, especially to do these kinds of long programs. I think it’s a trust issue. I’m not exactly sure what it is. 80% of the people on this show are left wing. I’m pretty good faith and I try to bring out the best in people.
Ezra Klein
(00:21:29)
Have you invited her? Is that what you’re saying?
Lex Fridman
(00:21:30)
Yeah, yeah.
Ezra Klein
(00:21:31)
We’ll see what happens when people get closer to 2028.
Lex Fridman
(00:21:35)
Sure.
Ezra Klein
(00:21:35)
Maybe people begin taking that kind of risk.
Lex Fridman
(00:21:36)
I hope so. Bernie’s up there in age so he can’t do it anymore. Why is the Democratic Party so bad at generating this kind of power?
Ezra Klein
(00:21:45)
I don’t think it’s so bad at generating them. I think that it turned out to be bad at generating them this year. Look, as you mentioned back in February, 2023, I was somebody who came out and said, Biden can’t run again. This isn’t going to work. And my view, and that was really what that set of pieces was about, was about the argument that even though Biden was clearly going to win the primary, that there was still time for Democrats to do something the parties had done in the past and have an open convention. And you could structure the lead up to an open convention in a number of different ways. You could have something like a mini primary. But basically you’d have Democrats out in the media out giving speeches. And their ultimate audience would be the delegates at the Democratic National Convention. And my hope was through that you would find the person for this moment.

(00:22:29)
The thing for Kamala Harris, it was really difficult was she was for another moment. She’s picked by Joe Biden in 2020 amidst just a very different political equilibrium, a sense that you had a transitionary moment between two versions of the Democratic Party. Maybe Joe Biden reaching a little bit back to the past to these sort of lunch pail, blue collar Democrats. Joe from Scranton was a big part of the Joe Biden appeal. But also Biden never has a chance if he’s not Barack Obama’s vice president. And so, you have this sort of weird set of historical factors operating at the same time. There’s a desire for stability and experience amidst the chaos of Donald Trump in the pandemic. There is Biden as Obama’s vice president who nevertheless did not run in the election after Obama.

(00:23:17)
I think a lot of people look back at 2016 and think, you know what? If Biden had been the candidate, he would’ve beaten Trump and we would live in a different reality. And then Biden chose Harris as an effort to shore up his own, at least assumed weaknesses. He’s a white man in the Democratic Party at a time when the Democratic Party is diversifying. And when the view of how you win elections is you put back together the Obama Coalition. And the Obama Coalition is young people, it’s voters of color and it’s enough working class white voters and then college educated white voters. That’s the Obama coalition. And so, Biden picks Harris for different reasons. My view at that time was I was sort of a Tammy Duckworth person and thought he should have picked Tammy Duckworth. But there were different people out there.

(00:24:05)
And then the kind of moment that Harris was running in just sort of dissipates. First, she has a particular background from California where she’s tough on crime. Her book is called Smart on Crime Prosecutor. But she runs in the Democratic Party at time when it’s turned on that kind of politics. People want a lot from her personally, but they don’t want a sort of prosecutorial character. So, she sort of abandons that and never I think really finds another political identity, certainly before she begins running in 2024. That works. But she’s a talented debater. She’s a very talented performer on the stump, but she doesn’t really have a theory of politics and policy that she’s identified with. But she’s a way for Biden to signal that he understands that him being in 2020, a 78-year-old white guy, he understands the future is not him, or at least not just him. And he’s sort of trying to make a coalitional pick that speaks to his own potential weaknesses.

(00:25:02)
I think by 2024, you have two problems. He only steps down, what is it, June? They’re weeks from the DNC. They don’t have time anymore for an open convention. You know how the Biden position is very unpopular for a number of reasons, but particularly inflation and cost of living. So, now you have Kamala Harris running with a sort of anvil of being associated. It’s a Biden-Harris administration. She doesn’t really have a lane on cost of living. It’s not something she’s known for working on in the Senate. It’s not something she has a bunch of great ideas about. It’s not something she could write a talking about. It’s probably not the candidate you would pick for a cost of living election.

(00:25:40)
And she’s had no time to build that out. Maybe if she had been running in a primary for a year and a half, having to fend off Elizabeth Warren, and Bernie Sanders, and Pete Buttigieg and whomever else, she either would’ve figured out how to do it. Primaries or periods of education and learning for the candidates too, or they would’ve found somebody else who could do it. But she doesn’t get any of that. She’s thrown into the game with three months to go. So, they picked the candidate in 2020 who won. Whether you think Biden’s inspiring or not, he was a reasonable pick for that moment. He should have never run for a second term, and he sort of implied to a lot of people that he wouldn’t. And then the handover to Harris was a very difficult handover to a candidate who didn’t go through any kind of selection process for the moment in which he was running. We’ll see what they do in 2028, but the consequences of what they did in 2024 have been severe.
Derek Thompson
(00:26:29)
There’s two really big questions on the table that I think click together in an interesting way. You asked one, why did Trump win? And two, why do Democrats have this certain communication style that might make them less interested in coming onto an unstructured three-hour conversation with you? Let me try to tell a story that connects them. I think Trump’s victory in 2024 was over determined if there are a lot of factors here. Number one, if you look internationally, incumbents lost all over the world. They lost in the US, they lost in Europe, they lost in pretty much every developed country at rates that we really haven’t seen in 50 years. And that’s largely because the inflation crisis that came after COVID created an absolute disaster for incumbent establishment power. People couldn’t bring prices down. Voters were furious and they were destroying establishment orders all over the world. Democrats happened to be in power, and as a result, they got the brunt of it. That’s number one.

(00:27:21)
Number two, if you look at elections over the 21st century, two things are true. One, almost every election is unbelievably close. For reasons that I’m not sure I entirely understand, the parties have gotten really good, historically, bizarrely good at getting each group to come to the polls with about 48% such that every election is a battle over the next 1.5%. And in a world like that, little thermostatic swings are very important. And what we’ve seen over the last few years, and there’s this theory about thermostatic public opinion in American politics, that says that what often happens in politics is one party has a very compelling message of change, they become the establishment and then they become the victims of exactly the weapon that they marshaled. That then the next out-group party says, “We have a theory of change and we’re going to throw out the bums.” And the next party comes in and they overreach and then they lose.

(00:28:13)
In a world where you have thermostatic change and every election is very close, you tend to have elections swinging back and forth. So, I think that also explains why Democrats and Republicans have struggled to hold onto power for six-year, eight-year, 12-year terms the same way they did say in the 1930s or 1960s. But finally, you have to look at what kind of character Donald Trump is and what kind of a media figure he is. We were just talking off camera about how every age of communications technology revolution clicks into focus a new skill that is suddenly in critical demand for the electorate. The world of radio technology is a world in which Franklin Delano Roosevelt can be powerful in a way that he can’t be in the 1890s. And then you have the 1950s. Dwight Eisenhower, 1956, I believe is the first televised national convention. Famously the 1960 presidential debates between JFK and Richard Nixon take an election that is leaning toward Nixon and make it an election that’s leaning toward JFK because he’s so damn handsome and also just electrically compelling on a screen.

(00:29:21)
We have a new screened technology right now, which is not just television on steroids. It’s a different species entirely. And it seems to favor, it seems to provide value for individuals, influencers, and even celebrities and politicians who are good at something like live wire authenticity. They’re good at performing authenticity, as paradoxical as that sounds. Trump is an absolute marvel at performing authenticity even when the audience somehow acknowledges that he might be bullshitting. He’s just an amazing performer for this age. And it speaks to the fact that he seems to be, to borrow Ezra’s term, remarkably disinhibited in front of every single audience. There doesn’t seem to be this sort of background algorithm in his head calculating exactly how to craft these message to different audiences. He just seems to be like a live wire animal in front of every audience. And I think that compares very distinctly to the democratic character of bureaucratic caution in our age.

(00:30:26)
And there is a really important distinction between this vibe of the Trumpian ruler and the vibe of the rule follower. And the vibe of the bureaucratic rule follower is a little bit afraid of unstructured conversation, is always performing the background algorithm of how do I communicate in a way that balances all of the coalitions on my side? Because if you look at the Democratic Party right now compared to the Republican Party, I mean in 2015, I think there were four political parties in America. There was MAGA, there was the Center Right, there was the Bernie Wing, and there was the Biden Clinton Obama wing. And what happened is that Trump killed and skinned the Center Right and is now wearing it as a hat.

(00:31:10)
The entire Republican Party is Donald Trump wearing the skins of the old center, the Romney wing, and the Democratic Party is still a fight. It’s exactly what Ezra described. It’s a jungle. And maybe there’s something about that jungle nature of the Democratic Party that is making some of its leaders perform this sort of coalitional calculation when they’re communicating, such that it makes them less interested in appearing in settings that might cost them, that might not benefit them in exactly the sort of pre-calculated way they have to get their message across. And so, there’s not necessarily a whole lot of empirics to that theory. I’m a little bit going on vibes here, and maybe Ezra see some flaws to theory.
Ezra Klein
(00:31:50)
It’s an age of the vibe.
Derek Thompson
(00:31:52)
It’s the age of the vibe. Yeah, exactly. I’m trying to perform the live wire authenticity that I’m describing, but I do think that might begin to explain why you, Lex, might be picking up on a difference between the political vibes. An eagerness and a willingness and the one hand to have kind of unstructured and even chaotic conversations, and a care on the other side about not letting conversations become too unstructured or too careless.
Ezra Klein
(00:32:16)
Can I build on that? I know we’re supposed to talk about abundance, but I want to talk about this.
Lex Fridman
(00:32:22)
There’s an abundance of time.
Ezra Klein
(00:32:24)
An abundance of time. We’re on the Lex Fridman Show. So, two or three things. One is Democrats still think the currency of politics is money and the currency of politics is attention. And that’s a huge difference between the two sides right now. So, what did Kamala Harris come in and do? She came in and raised a shit ton of money, like a billion dollars in record time, basically. She had more money than Donald Trump did, and used it to try to buy attention. What it meant for Democrats to be good at social media, is to have a good social media team. People in your office somewhere in your campaign headquarters who put out cool things on social media, good memes, and good advertisements-
Ezra Klein
(00:33:00)
Put out cool things on social media, good memes, and good advertisements and so on. What it means on the right to be good at social media is to be you personally good at social media, your Vivek Ramaswamy, you’re JD Vance, your Donald Trump, your Elon Musk, and what you understand is you are the product. What it means to be good at attention is you are good at attention. Now, Harris, I think was actually better at some dimensions of this. They were just slightly older dimensions and people always gave her credit for, hell of a performer on the stump. She was way better on the stump than people realized she would be, and a good debater, she’d always been a good debater. She trashed Donald Trump in that debate, but she does not do social media herself at any level because she’s not going to take risks Democrats.

(00:33:41)
Most Democrats still live in a world where the thing that they’re optimizing for and attention is to not get negative attention. And what the Trumpist wing of the Republican Party understands, and this is truer for them than it probably would be for Democrats because for them the media is the enemy, or at least the mainstream media is etcetera, but is that attention. A volume of attention is itself good and you can only get a critical mass of it if you’re willing to accept negative attention. Agenda control doesn’t come from positive attention. It comes from conflict. You get agenda control by doing things the other side disagrees with, so they enter into functioning agreement with you to keep the thing you’re doing at the front.

(00:34:23)
Now, that doesn’t make you highly popular. Donald Trump is the most unpopular modern president at this stage of his presidency except for Donald Trump’s first term. It took, I think Nate Silver said it was 221 days for Joe Biden’s net favorability to go negative. It’s taken something like 55 days for Donald Trump to do the same. So what Donald Trump is doing does not optimize for favorability. It does not optimize, by the way, for big wins. Democrats feel like they got trashed in 2024, and in a way they did. But Trump’s popular vote victory was the smallest popular vote victory since 2000 when Al Gore beat George W. Bush by 17 dogs and three old men or whatever it was. And so, attention works really differently.

(00:35:06)
And while I think some of the Rogan of the left discourse has been frankly overstated, because honestly, the most parsimonious model of 2024 and 2020 is in 2020, you have a 4848 nation, something like that. Or maybe you have something that’s more like a 49 Democrat, 47 Republican nation. And in 2020, because of the pandemic, Donald Trump suffered a, let’s call it a 2. 5-point incumbent penalty. People are mad about the pandemic, they’re mad about things being chaotic, so he loses 2.5 points that given the natural split of the electorate, Joe Biden, a 4.5-point popular vote victory. In 2024, people are mad about inflation. They’re somewhat mad about the border. You have a 2.5-point penalty applied to the incumbent administration. Now it’s Harris, and you get a 1.5-point popular vote victory for Donald Trump. I genuinely don’t think, and this held internationally too, right? I generally don’t think you need a lot more to explain the election right now than that, but you do need something more than that to explain.

(00:36:06)
Donald Trump’s now since 2016, almost decade long dominance of all attention in American politics. Starting when he came down the golden escalator in 2015, American politics from 2015 to 2020 when Joe Biden won was about Donald Trump. And then from 2020 to 2024, when Joe Biden was president, it was about Donald Trump. And then from 2024 on, it was about Donald Trump. Joe Biden was an attentional void, be it his age, be it their strategy. They agreed that the topic of the nation should be Donald Trump, right? When he went back to begin his campaign in 2024, he goes to Valley Forge and gives a speech about January 6th and Donald Trump, right? It wasn’t about his own achievements, it was about Donald Trump. Joe Biden didn’t do the Super Bowl interview in 2023. That’s when I did my thing about this is not going to work. Probably because at that point, he was not capable of good extemporaneous interviews.

(00:37:02)
I mean, I think that was my view of them, that the revealed thing here was that they didn’t trust him to do interviews. I didn’t have some inside information about anything, I just looked at what they were doing, what they weren’t doing. They’re behind in the polls. They weren’t doing things like the Super Bowl interview. If you can’t turn your candidate into the product, if you don’t trust your candidate to be the product in an election, you’re fucked, right? And so that was to me, the tell, but attention is the coin of the realm. Now, there are better and worse ways of doing it. I don’t think Donald Trump is doing himself huge favors right now. I think they had, there was a path they could have walked to be a majority party.

(00:37:37)
I think that if he was more restrained, more inhibited, if he was able to not do a bunch of things that are mobilizing opposition to him, you could talk about what they would or wouldn’t achieve that way, but I think they could be in a much stronger political position that would make them stronger for the midterms. It would eventually make JD Vance stronger as a successor. I think they’re running a very high-risk strategy that has a very reasonable chance of if they don’t make what I would call an autocratic breakthrough, they might completely blow up their own movement. It’s all very high risk. So, for them, for everybody, for everything, what makes them good at politics is also what makes them bad at politics.

(00:38:14)
But for Democrats, the caution, the bureaucratic culture, the fear of saying anything that will make anybody mad, it is optimized for a different attentional era. One of the things I’m watching when you’re saying about leaders, one of the things I’m watching in the people coming up, the ones who want to run in 2028 is who seems like they have adapted to this era not in the way Trump did or Vance did or Musk did. I think they’re going to need something different. They fully represent the Twitter era of politics. I mean Musk bought Twitter. I think sort of all in on what politics right now, what online politics feels like. I think the thing that will come next is someone who’s able to synthesize both the lessons of it, feeling that we all have that it’s kind of sick and poisoned, right? That Twitter’s not a good place, X is not a good place. TikTok politics is not a good place that we’re all being turned on each other somehow.

(00:39:08)
You need to be authentic and authentically angry at what we’ve all become in the way that Obama ran as a political reformer who hated the red and blue cut of America, who hated what political consultants and pollsters were doing to us. You’re not going to have somebody who just echoes. There’s not going to be, there’ll be no Joe Rogan of the left. There will be no Donald Trump of the left because the left is different than the right, but it’ll have to be something authentically of this era, but also authentic to the backlash to it, which I think as we enter into this period where the president and everybody around him fully embraces this attentional economy, I think people are going to want something different from this attentional economy in four years.
Lex Fridman
(00:39:45)
And be okay with a negative attention that comes with being authentic.
Ezra Klein
(00:39:49)
You’re going to have to have some of it, right? You cannot change American politics. You can’t change the Democratic Party if you’re not willing to upset people. Donald Trump reformed the Republican Party by willing able to fight Republicans. He ran against George W. Bush, against Jeb Bush, against Mitt Romney, against the trade deals, against a bunch of things that were understood to be sacred cows. Somehow this guy ran right after Mitt Romney and John McCain while attacking Mitt Romney and John McCain, right? If you are not, the Democratic Party does need to change. It needs to attain a different form because the Obama Coalition is exhausted. It’s done. It’s not going to be able to do that if it doesn’t have standard-bearers who are willing to say we were wrong about some things. We have to change our views on some things. We have to act differently and speak differently.

Internal political divisions

Lex Fridman
(00:40:34)
Is there a degree to which the left uniquely attacks its own more intensely than maybe other parts of the political spectrum?
Derek Thompson
(00:40:47)
It’s possible. You go back to the model that I gave you of 2015 where there used to be these four large parties, MAGA, center right, center left, and left. Right now, the Republican Party is all MAGA, so there is no coalitional fight to be had. It’s all Donald Trump. And if Donald Trump wants to name a former left-wing environmentalist to be the HHS secretary, everyone says, “Okay, that sounds like a fantastic idea. That’s exactly who we were going to nominate too. Thank you, Donald. That’s wonderful. It’s at the tip of my tongue.”

(00:41:16)
On the Democratic side, there is a fight and it’s happening right now and our book is trying to win a certain intra-left coalitional fight about defining the future of liberalism in the Democratic Party. So, I’m not of the left. I’m certainly not of the far left. I have center left politics and maybe even a center left personality style if we can even call it that, but I do not begrudge the left for fighting because there’s a fight to be had. In many ways, I think sometimes they see, I’m not endorsing this, I’m describing it. I think they see their near-term opposition as not always the Republican Party, but as the forces, the Democratic Party that are in the way for them controlling one of the two major parties in this country.

(00:42:10)
And so, they do have an oppositional style and maybe that’s personality based. They are fighting the center left. They are criticizing the center left consistently, but I want to be good faith about this even though I don’t share their politics and say that they’re doing it because they’re trying to win power on the left of center, and so that’s why they’re criticizing the way they are. Now, our book and much of my writing is an attempt to do a little bit of a very specific dance as we touched on this I think really beautifully. We’re in an era right now of anti-institution politics, anti-establishment politics, and Democrats are at risk right now as being seen as the party that always defends institutions, the party that always defends the establishment status quo, and that is an absolute death knell, I think for this century’s angry anti-establishment politics.

(00:43:05)
So, what we’re trying to do is essentially say, here’s a way to channel the anger that people have at the establishment, but toward our own ends, right? We believe that we have answers on housing and energy and high-quality governance and science and technology, really good answers that are fiercely critical of the status quo in Democrat-led cities and Democrat-led states. We’re trying to be oppositional in a way that’s constructive rather than just destructive.

AOC

Lex Fridman
(00:43:34)
Just to put a nice pretty bow tie in the whole thing, let me ask for advice. What do I need to do for AOC to do a three-hour interview with me? Ezra, from your throne of wisdom.
Ezra Klein
(00:43:49)
I don’t think I know how you get AOC herself to do it. I would not pretend to know her offices or her particular views on this. I do think though that you can see different Democrats taking on different kinds of risks. Right now, we’re sort of in the age of Gavin Newsom. I mean, Gavin Newsom is the Governor of California and he’s spending some percentage of his time doing a podcast with Charlie Cook and Michael Savage and Steve Bannon. Gavin Newsom realizes that one lane for a Democrat is to be high risk and talking to virtually everybody. I think Pete Buttigieg in a different way, somebody who wants to take media risks. Now I think he’s going to, my gut on him is he’s going to hold his powder a little bit, so he’ll probably want to do the Lex Fridman podcast assuming he runs in 2028, in 2027.
Lex Fridman
(00:44:39)
Buttigieg?
Ezra Klein
(00:44:40)
Buttigieg, right. I think a lot of them are trying to figure out what is the lane for right now, and there’s a lane for the next two years and there’s a lane for the two years after that, and you’re going to see a lot of people begin to blanket media in the two years after that. Now, it’d be interesting, I would be curious to know, would Hakeem Jeffries come on your show right now? That’d be interesting. I mean, would he do it for four hours? I don’t know. The four-hour ask, the three to four-hour ask as somebody who also books politicians is hard. I have trouble, I like to book people for 90 minutes to two hours, and I tend to get negotiated down to, I try not to go under 75 or 65. But even as somebody I think well regarded in that world, it’s very, very, very hard for me to get politicians to sit for two hours.
Lex Fridman
(00:45:23)
I don’t have the sense that the three-hour ask is a big ask because of scheduling. I think it still is grounded in the fear of saying the wrong thing.
Ezra Klein
(00:45:32)
I just think they’re used to something else, right?
Lex Fridman
(00:45:32)
Used to, yeah.
Ezra Klein
(00:45:34)
I think that when you talk, I mean they are scheduled by schedulers. If you talk to them yourself, if you end up having a personal relationship with Wes Moore of Maryland and he wants to do your show, he will tell his scheduler, “I want three to four hours to do the show.” But the scheduler is used to a world. The staff is used to a world where nobody gets three to four hours for the boss. Reporters don’t, donors don’t, policy staffers don’t. So then, when some interview comes in and they say, “Hey, I want three to four hours.” The answer is no, because culturally it’s not done. You need Donald Trump himself, Pete Buttigieg himself, AOC herself to say to their staff, “No, no, no, we’re making time for this.”
Lex Fridman
(00:46:15)
Right.
Ezra Klein
(00:46:15)
Because it’s not how they make time for things normally. I don’t know how much it is fear. I do think they’re unused to it, but I suspect a lot of it is simply booking culture. I run into it too. They’re not used to saying yes to three to four hours for anything. It’s not that they don’t have it. They have three to four hours if their kid is having a graduation. I mean, they’re human beings, they can make time, but it would have to come in a way from them. My sense is part of the Rogan, it’s very unclear because they’re very differing stories on what happened in the Rogan-Harris negotiations, but it does seem that time was one of the sticking points.
Derek Thompson
(00:46:48)
It’s also possible that you’re going to find as you try to interview democratic politicians, that the exact same thing that happened with tech CEOs is going to happen among democratic politicians. You interviewed some tech CEOs and then they did a great job and their friends were like, “You were fantastic on a Lex Fridman podcast. That was such a great thing that you said in minute 97.” And then there becomes a bit of a meme that you can create really high value moments for yourself if you appear on Lex’s podcast and then it becomes less risky for the next marginal CEO to say yes. And I think right now what we’re talking about fundamentally is not physics, it’s culture, it’s just norms. I think there’s a fear of low expected value.

(00:47:28)
If you’re a high-ranking democratic politician and maybe you do a podcast like this, what if I say the wrong thing and it goes viral and my bookers and my agenda, people and myself just feel terrible for the next few weeks because all we see on Bluesky and X is just people hating. But if you get one interview that goes well, if you talk to Wes Moore, let’s say, and there’s a five-minute segment where he articulates vision of liberalism in the 2020s that everyone says, “My gosh, that is the best possible articulation of what the Democratic Party can stand for the next four years that I’ve ever heard.” Suddenly what’s happened is that appearing on the show becomes massively de-risk. And in fact, the risk valence entirely switches. We are leaving dollars on the floor by not appearing on this guy’s podcast, AOC. I’m a sensational communicator. If Wes Moore can do it, I can do it even better.

(00:48:19)
And so, I think to a certain extent, there’s a little bit of a riot theory phenomenon here. The person who throws the first stone in a riot has to be very courageous. The person who throws the 101st stone in a riot doesn’t need to be courageous at all, and there might be a little bit of that going on that people need to see proof of high expected value before it breaks what we’re acknowledging to be a bit of a communications norm on the left.
Lex Fridman
(00:48:43)
That’s a really convincing and powerful theory. I think I want to push back on it. So, what’s going to happen, for example, this very conversation, you’re both going to come off brilliant.
Ezra Klein
(00:48:56)
Well, we’re barely like 50 minutes into a four-hour.
Lex Fridman
(00:48:59)
Yeah, it’s going to be… Let’s see what happens.
Derek Thompson
(00:49:00)
I have no more material. Lex, it’s all straight downhill…
Lex Fridman
(00:49:03)
But people will get tired, they’ll listen to this…
Ezra Klein
(00:49:03)
It’s sloppy.
Lex Fridman
(00:49:06)
And they’re like, “Well, that’s Ezra. He’s brilliant.” They’re going to be worried about their own candidate. If AOC comes off brilliant, they’re not going to be thinking, “Oh, this is a place to be brilliant.” They’re going to be like, “Well, because AOC is brilliant, but my candidate is not.” It’s still boils down to the caution. I’ve had a lot of Republican, high profile Republican people reach out to me, they don’t give a shit. They’re just like, “Whatever, I’ll come.” People on the left, I’ve had two people who I respect deeply and admire, expressed caution about the previous people I’ve interviewed and not wanting to come on.
Ezra Klein
(00:49:45)
Well, that’s the thing the left has to get over. Yeah, that’s a very important cultural. They began to do a thing where spaces were verboten because it had platformed so-and-so, and I think that culture is changing. I think they realized, I mean, that they abandoned huge, vast swaths of significant culture because they wouldn’t go places. It’s crazy. The idea that you only go where people agree with you is genuine lunacy. I don’t want to act as your booker here and 2020 candidates are what they are and they each have their own press strategy. But I would say I’m doing this myself on my show. I don’t think the best people to get right now in the Democratic Party are the seven people who lead the polls. I’m looking for people who will think out loud.

(00:50:35)
I just had Jake Auchincloss on. He’s not that well-known house member from Massachusetts. I just think he’s a guy who thinks out loud. I am booking someone else right now two more people like that frankly, who are neither of them. They’re like [inaudible 00:50:51]. I could get a bigger name. I’m more interested in people who are thinking. One reason I think Bernie Sanders always did everything is to him, the thing he was really doing was he had something to say. The point of almost everything for Bernie Sanders is to be in a place where people can hear the thing he has to say, and a lot of politicians don’t have that, right? The point of anything is what it does for them in the polls, what it does for them with the donors, how it repositions them.

(00:51:18)
I’m interested right now at this moment because it’s not 2028, it’s not 2027, there’s no democratic primary happening right now. The idea that you’re going to pre-run the Democratic Party primary is stupid. We don’t even know what the, we might be in World War Three. We have no idea what, we might have AGI, the things that the 2027 primary might be about, the kinds of scandals that might’ve erupted by them are totally different. Whereas, the reason I have people on is because they are saying something. They’re a live mind in the moment that has a perspective on this that you want to hear. And so, I would look for that.

(00:51:56)
And I think a lot of people who are not there are people who are trying to protect something, protect a standing in the polls, protect a sort of coalitional set of allies they currently have, and then there are people who are trying to just be heard. Again, a lot of Bernie Sanders’ culture, the way he does media is now Bernie Sanders is Bernie Sanders. He didn’t used to be. He wasn’t in 2015 even. When Bernie Sanders ran for president in 2015, it wasn’t initially a big deal. It was like, “Oh, bummer.” Elizabeth Warren didn’t run for president was a feeling of most people on the left. And so, Bernie Sanders was the guy who’s been saying the same thing for decades, but in the wilderness and nobody was listening.

(00:52:37)
And now, he still has the instincts of somebody who understood that the most important thing was to find a place where people were listening, where they would let you talk and even better let you talk for a while. I think the candidate who’s going to do well in 2028 is going to have an instinct like that, but even right now, I think the question is, one of my big questions is I’m booking for my show is just I want someone who has a perspective on this moment, who feels like they have had a thought that is about right now and who we are right now and what the story of America is right now.
Derek Thompson
(00:53:11)
I think what’s really important is what Ezra said, it’s about having something to say. We wanted to talk to you and talk together about how much we wanted to talk to you because we got something to say. We wrote a 300-page book about how much we have to say. We love going on podcasts and television shows and radio and then doing live events to tell people what we have to say. We think this idea of abundance isn’t just important for redefining what the American left means. We think that the outcome of thinking abundantly about housing and energy and science and technology is what politics is all about. It’s about people, the good life, and we think this is the path toward it.

(00:53:49)
So certainly, one thing that’s profoundly motivated us is having something very concrete that we just want to get out there in the world. My sense as a writer, as a thinker is I want my software running on as many pieces of hardware as possible. I want to get my ideas out there as much as possible and who gets credit for them and where the idea goes, and that’s all secondary. I believe on ideas because they’re important. And so, I want to talk to people who have large platforms about those ideas because how else is the idea getting into the mainstream except through those large platforms? That’s how broadcast technology works in the first place. So maybe one thing that you’re touching on is a little bit of ideological ambiguity. Maybe a part of this is this sense that people don’t know exactly what it is they have to say for three hours. All I can say for sure is that we know.

Political realignment

Ezra Klein
(00:54:42)
There’s also a reality, and I mean the book is trying to enter into this reality. I think one thing you’re saying is that people have coded you. And so, Donald Trump is really excited to do it and maybe left with politicians are not. One thing that we think is that we’re in a period of realignment. The last chapter of the book, we talk about an idea that is picked up from a historian named Gary Gerstle, which is an idea of political orders. And political orders are periods that have a sort of structure of consensus and a structure of a zone of conflict, but it’s more or less agreed on by the two sides, even if only tacitly. So, you have a new deal order, new deal order is founded by FDR. It is entrenched when Dwight Eisenhower accepts the New Deal as part of the US proving that it can treat workers better than the Soviet Union. So those are sort of right there. The three ingredients typically have an order.

(00:55:40)
You have a party that starts it, opposition party that accepts key premises. Dwight Eisenhower doesn’t come in and say, “We’re going to roll back the whole New Deal,” and it’s often held in place by an external antagonist. In that case, the Soviet Union. You then you have in the 70s stagflation, the Vietnam War, series of problems that the New Deal order no longer seems able to handle. So, you have the rise of what he calls the neoliberal order. And the neoliberal order is if you’re going to choose a founder, it’s going to be Reagan on that one. It’s much more about markets. It’s very concerned with things like inflation, and it really is entrenched by Bill Clinton, the era of the government is over. And partially, it’s entrenched also by the fall of the Soviet Union.

(00:56:20)
The fall of the Soviet Union is like this proof point that the capitalists are right, that markets are the way of the future, government does not know what it’s doing, and that becomes the governing set of assumptions. And so, there are arguments about what the markets should be doing. Obamacare is about creating markets and health insurance. You can use markets very progressive ends. We want to use markets for lots of progressive ends, but the neoliberal order basically collapses amidst a financial crisis and climate change and China. And those are the three things that grow a little, but also separately we think kill it, which is the neoliberal order does not have an answer to the financial crisis, and it botches in many ways.

(00:57:01)
The answer to the financial crisis puts too little demand into the economy, lets a recession linger and a very slow recovery linger for too long, it doesn’t know what to say really about climate change. Markets have made a lot of people rich by doing a lot of things that are very, very damaging for the environment, very damaging for the future of the human race potentially. And you have the rise of China and the neoliberal order said you integrate China into the global economy. You bring them into the WTO, you trade with them, you help them build their industrial base, you help them pull their people out of poverty, which that part is good, and they’ll become more like the West. They will liberalize. They’ll have a free press. The richer we make China, the more China’s going to become like us, and that proves totally wrong.

(00:57:46)
China becomes more authoritarian over time, but it also develops an industrial base. It becomes as it does not become more like us, becomes dangerous, at least in our view. You don’t want to ever have a conflict with another country who you’ve outsourced your key industrial base to. And so, you have the sort of fall of that order. And then again here, things that would’ve been ridiculous at one point in American politics then become possible. Bernie Sanders is one of them, right? The idea that you’d have somebody, a self-described socialist running for president and coming anywhere near the Democratic nomination, that was unthinkable in 2004. And by 2016, it almost happened.

(00:58:24)
And Donald Trump is another thing. Donald Trump runs like headlong into the failures of neoliberalism in the Republican Party. He runs against trade. He runs against a sort of Paul Ryan more open immigration. George W. Bush and John McCain were both very big on liberalizing immigration policy. He runs against the Iraq war and foreign adventurism, and there’s an isolationist instinct that coexist very awkwardly now within a territorial expansionist instinct, but at least to 2016, it was more isolationist. And so, Donald Trump and his sort of reimagining of the Republican Party as a right-wing populist, more like sort of some Christian democratic parties in other countries up in that quadrant of socially conservative, economically populist, that becomes something that’s possible.

(00:59:13)
But nothing has found an equilibrium, right? Nobody’s agreed to the other side’s premises. There are certain ones that people are agreeing on, both the Republican and Democratic parties a very different view on China now. Biden kept a lot of Trump’s policies on China and actually strengthen them, and now Trump is building on that aggressively again. But in terms of the other things, there isn’t agreement about what the next period in American politics should look like, and that’s one reason I think it’s very dangerous, both as a question of media strategy, but also is a question of politics to code people, places, platforms too tightly. Republicans and Democrats aren’t going to get along in Congress. That has to do with, I think the incentives of Congress.

(00:59:54)
My first book is called Why We’re Polarized. It’s about those almost hydraulic incentives for partisanship. But in terms of what is the meaning of my podcast, of Derek’s, of yours, of Joe Rogan, of Theo Vaughn, of Call Her Daddy, of a million different places that are not well coded, that’s I think are very up for grabs. I mean, Elon Musk was an Obama-era liberal in 2012. I mean, I think his personal process of radicalization is not going to unwind itself, but a lot of the people who Democrats are like, “All these billionaires are right-wing now.” No, people are just uncertain. I mean, some of them are a little bit afraid, but people are uncertain. They’re moving back and forth. The sort of texture of it is unsettled and it’s going to take time, these transitionary periods.

(01:00:41)
I mean, they can go very badly too, but they take time and I think people who are clinging to old certainties about what tells you which side folks are on, my sense is a lot of people who are very open to MAGA in 2025 are going to feel very differently about it in 2028 depending on how they do, right? If they do great, then they’re going to entrench. But if they don’t, then a lot of people became MAGA curious are not going to be MAGA curious anymore, but they’re not going to want the last Democratic Party either.

(01:01:11)
I was making this point to someone the other day about why the Democratic Party’s embrace of the Liz Cheney style independent didn’t work. Liz Cheney, of course, being Dick Cheney’s daughter, a Republican. But what Liz Cheney, the Never Trumpers were a way of reaching out to who the Democratic Party thought the independents were. But the key thing about an independent to a political party is not that they don’t like the other party, it’s that they’re an independent because they also don’t like your party. And so, finding a bunch of people who are meant to be messengers to them about why they shouldn’t like the other party, it’s fine, but what you need to do is explain why they should like your party. You need to have some message. You need to accept some fault. You need to think about what it was about you that drove them away.

(01:01:58)
One of our deep views about politics right now and not politics policy, the texture of the economy of the country is that the last period in American politics, in the economy was about demand. The fundamental problem coming out of the financial crisis was demand. We had too little demand in the economy. Behind that too little demand in the economy was this other thing that was building up, which was a cost-of-living crisis. Housing was getting super expensive, healthcare in certain ways, energy. But energy is more complicated in ways we can talk about elder care, child care, higher education, right? This is a point, my wife is a journalist at the Atlantic, Annie Lowrey with Derek, and she wrote this piece on in early 2020. The pandemic that went very viral called the Affordability Crisis.

(01:02:44)
And it sticks in my head because she’s writing at a time when people were saying, “The economy’s great. Everything’s great.” You looked at measures of consumer confidence in 2020, February of 2020, terrific. She’s like, “So, how come if the economy’s so great? Everybody I talked to is so upset.” And she’s like, “Look, people are making more money than ever, but it’s getting eaten up and eaten up and eaten up by these things they really need.” They keep getting more expensive, even as consumer goods get cheaper, then the pandemic hits, the problem becomes COVID, but then you have inflation and inflation moves the problem of the economy, the fundamental problem everybody’s paying attention to.

(01:03:20)
From the demand side, how do we get more people at work? How do we get them to spend more money to the supply side? We don’t have enough. We have a constriction of semiconductors, of used cars, and then eventually everything, everything is getting more expensive. We’ll see what happens with tariffs. We do get, by 2024, the rate of inflation under control, but prices are still much higher and now people are paying real attention to prices. And the affordability crisis, which again is a cost-of-living crisis, which had been growing for a very long time, is now at crisis levels and it becomes the substance of politics. You had all these Democrats saying, I don’t know what the problem is. Inflation has come down to whatever it was, 3 to 4% in 2024, and they’re right about that. But one, the price level of everything remained high. Two, people were now like, “The housing so expensive. I’m never going to be able to afford a home.” My parents went to public university, debt free. I could never do that. And what we’ve done is fail.

(01:04:19)
I mean, Democrats in this case, Republicans haven’t done that great on it either. But in blue states, Democrats have failed on cost of living. The reason California, Illinois, New York are losing hundreds of thousands of people to Florida, Texas, Arizona, Colorado, is that they failed on cost of living. It is too expensive to live there, and the reason they failed at cost of living is supply. They did not make enough. They actually made it too hard in many cases to make enough of the things people needed. Some of those are straightforward. We didn’t let people build enough homes. Some of them are more like we’ve made it too expensive to build public infrastructure like high-speed rail or the Second Avenue subway. Some of it has to do, I think, in the long run with innovation and the relationship between Democrats and technology.

(01:04:58)
But one of our views is that there are other things in politics that will matter too, but we are in a period where the cost-of-living supply affordability is the fundamental economic question. Donald Trump himself has said he won because of the price of groceries. He’s got this very funny quote where he is like, nobody said, I don’t have a Donald Trump impression, but he’s like, “Nobody ever used the word groceries in politics before I did.” Well, it’d be good if he then wasn’t making it more expensive, but Democrats believe his weakness is cost of living. They’re probably right, but they don’t have a strength on it. The key question our book is trying to refocus politics on is how do we make more of what we need? How does the government either organize itself or organize markets to create more of what we need, and how do we admit as liberals times when we’ve made it so the government makes it too hard to make more of what we need?

(01:05:47)
I’ll say one last thing and this is a pretty long answer. I thought one of the most important things that has come out recently is a piece in Foreign Affairs by Brian Deese, who’s a former head of Joe Biden’s National Economics Council, and Deese helped negotiate every major Bill Biden passed. It’s a very-
Ezra Klein
(01:06:00)
He’s helped negotiate every major bill Biden passed. It’s a very straightforward piece about what it is Democrats have not done to make it possible to build at the level of their goals. And he says things like, “We should just remove federal funding from cities that have highly restrictive home zoning codes.” He says, “We should have a goal for how much nuclear we build in the next 10 years. We should be trying to reach a goal of new nuclear capacity.” It’s a very, very important piece because Deese is right at the center of Democratic policy. Instead of retrenching, he’s like, “Okay, we didn’t get there. What do we do now to make it possible to get to the place we promised you we can go?”

Supply-side progressivism

Lex Fridman
(01:06:37)
And we should say that the book you’ve mentioned, which I’ve gotten a chance to read, and I think it’s incredible, highly recommend, it’s called Abundance. I think of it as a kind of manifesto for what the left would represent in the coming years. So I think people should read it from that angle. And both of you have been writing about this topic from different angles for a while. I think in ’22, Derek, you wrote an article on this topic of abundance titled A Simple Plan to Solve All of America’s Problems. And Ezra, you wrote an article in ’21, a supply side progressivism titled The Economic Mistake the Left is Finally Confronting. And you’ve just described, laid out this more progressive perspective on supply side economics that you’re presenting in Abundance. I was wondering if you could give the broad, high-level explanation of this idea of supply side progressivism?
Derek Thompson
(01:07:44)
Well, my piece about the abundance agenda, which I wrote in 2022, started with me standing outside, waiting for a COVID test. And this was a period where two years after the pandemic started, COVID tests were still being rationed and it was like 21 degrees outside. And I was getting very, very frustrated about the fact that still, we seem to have a scarcity of COVID tests. And as I’m sitting outside, just freezing my ass off and just getting really mad, I’m thinking, “It’s not just COVID tests we’ve had scarcity of. We also had a scarcity of COVID vaccines early on in the rollout, which created this really discombobulated scheme for distributing the early COVID vaccines.”

(01:08:20)
And then also, you go earlier into March and May of 2020, and we had a shortage of PPE equipment for our doctors to remain safe as they were taking care of a pandemic. And I thought it’s interesting that this entire experience of the pandemic has essentially been defined by this concept of scarcity. And as I zoomed out a little bit, I thought it’s not just the pandemic. It’s really so much the 21st century economy that’s been defined by scarcity. Ezra beautifully described the degree to which housing unaffordability has become the economic problem of our time.

(01:08:54)
In the history of political orders, each political order is in part defined by the internal crisis, the Great Depression springs neoliberalism, stagflation springs neoliberalism. Now we’re in this molten moment where we’re waiting for the new political order to emerge and it’s going to emerge because of the lever, because of the power of housing affordability. You have to solve that problem if you want to solve the problem of American anger about prices, and part of this is just pure arithmetic. If you look at any family’s budget, the biggest part of their budget in any given year is the part that goes to rent or mortgage. It’s housing, housing, housing, and housing connects to everything else. It connects to innovation. You want cities to agglomerate, to bring smart people together. Housing relates to all sorts of other affordability. Like if you care about the cost of eldercare, you want to make it cheaper to house institutions, buildings that can care for children, which means you want to bring those rents down.

(01:09:55)
And so I thought as I’m zooming out on this concept of scarcity in the 21st century, we have chosen to make housing scarce. In some of the most productive cities and states, often run by Democrats, we have rules, zoning rules, historic preservation rules, permitting processes, environmental reviews, laws that we created that have gotten in the way of making abundant the most important material good there is, which is housing. And as I kept working myself into a lather and getting mad about the world, I thought it’s actually not just housing. It’s clean energy too. There’s lots of environmentalists who are on my side and believing very fervently in climate change who’ve made it very difficult to site solar panels or site solar farms or to raise wind turbines or to advance geothermal or to accept nuclear power. We have chosen to make clean energy scarce as well.

(01:10:51)
And then finally, the ultimate boss of scarcity was the pandemic itself, which constricted the supply of all sorts of goods around the world, setting the price of everything to the moon. And that’s why inflation wasn’t just an American phenomenon, not just a North American phenomenon. It was a global phenomenon. And I thought what we need to solve for this crisis of penumbral scarcity is an abundance agenda, an approach towards solving America’s problems that puts abundance first. And Ezra and I have a very focused definition of abundance. We believe, we say in the first page of the book, America needs to build and invent more of the things it needs. We believe that housing is critical. We believe energy is critical. We talk a lot about science and technology, but we really put government effectiveness at the heart of this because one really deep vein of our book is a criticism of where liberalism has gone wrong in the last 50 years, where liberalism has gone from in the New Deal era, a politics of building things.

(01:11:55)
FDR and the progressives transformed the physical world, not just with infrastructure projects, but with building roads, the highway system under Dwight Eisenhower, we changed the physical world during the decades of 1930s to the 1950s. But in the last half century, liberalism has become very good at the politics of blocking rather than the politics of building. And if you look at the way that liberals define success in the last few decades, it’s often about success defined by how much money you can spend rather than how many things you can actually build.

(01:12:31)
You look at the fact that, for example, in the book, we have so many examples, California authorizes more than $30 billion to build a high-speed rail system, which basically doesn’t exist. Just last week, the Mayor of Chicago bragged that they spent $11 billion building 10,000 affordable housing units. That’s $1.1 million per affordable housing unit. That’s absolutely pathetic. We have a story in the book about a $1.7 million public toilet built in San Francisco, $1.7 million for a toilet because of all of the rules that get in the way and raise the price of building public infrastructure like public bathrooms in San Francisco and California.

(01:13:10)
So liberalism, I’m worried over the last 50 years has become so good at the politics of blocking and the politics of associating the money authorized as success rather than what you build in the physical world that we’ve lost sense of material abundance, of how important outcomes are and not just processes. And so this is a book that’s trying to nudge the Democratic Party back to what we think are in a way, historically its roots. Thinking about what Americans need and making it easier for government to act efficiently to provide them, and that really does, I think, begin with housing and energy.

Wealth redistribution

Lex Fridman
(01:13:47)
Is there a tension between the left, the progressive wealth redistribution kind of ideas with the idea of building that’s primarily getting out of the way and letting the market get the job done?
Ezra Klein
(01:14:09)
Let’s say two things on that. So one, we think there’s a real tension between equality, redistribution and constricting the supply of specifically housing.
Lex Fridman
(01:14:18)
So housing by the way, I would love to understand is that’s the big problem of-
Ezra Klein
(01:14:24)
Housing and energy, I think are the two most significant that we focus on in the book. Right? Housing and clean energy. We don’t have housing, we don’t have enough clean energy. I would add things to that, public infrastructure. We don’t really focus that much on education, but we could and we could talk about that. Immigration is probably there for me too, and we talk about that a little bit in the book and we do talk a lot about how to pull innovation forward from the future. But when you ask about redistribution, I really think this is an important point because there’s a great new paper by David Schleicher, and I’m so sorry because I’m forgetting his co-author. They’re law professors though, and they talk about the victory of gentry law. We used to have a law that was very dynamic when it came to property and land. It was very different than how things were in Britain. And over time, back half of the 20th century, we moved American law to be much more for what they call the gentry. We moved it much more towards protecting those who currently have things. Right?

(01:15:15)
And we do that through 1,000,000 things, covenants and HOAs and all these contracts we make people enter into so they can’t even build on their own land. But one of the things that just happens when you constrict the supply of housing is that people who got in when the getting was good, it’s a classic story, in New York and LA and in SF, you bought a place in 1977 for $220,000, and now it’s worth $2.7 million, and maybe you’ll pass it on to your kids or you’ll sell it, but the working class families can’t afford to live there anymore. Right? So that’s not even a question of redistribution. Sometimes what you need in order to create the possibilities for opportunity and mobility is enough supply of the thing.

(01:15:56)
At the same time, we don’t think that redistribution is the problem here. I’m pro redistribution, I’m pro more redistribution than we currently do, but to give one example the way these can be great tastes that go together, Derek tells in the book at some great length the story of Operation Warp Speed. And here you have in the mRNA vaccines, technology that was critically funded by public money, specifically DARPA at different points, then hastened after COVID. Government through Operation Warp Speed under Donald Trump really tried to clear out regulatory cruft, move these things really fast, but the demand on the side of the public for having funded so much of this, for having made so much possible was that when these vaccines hit, they were going to be free. Maybe the most important medical advance of that entire era, and it wasn’t going to be like Ozempic say, where it’s $15,000 for a year of doses. Right? It wasn’t going to be only available to the richest people at the beginning. We were going to try to give it to everybody, sorted by need to the best that we could and it would be free.

(01:16:58)
Now, you’re not going to do that with everything, right? There are places for the price signal to actually function and where it can function to then bring on more supply later. And there’s all the econ 101 stuff that we all know, but there are a lot of places where redistribution and supply increases go hand in hand. Another good example, I’ve done over the course of my career, a huge amount of work on health insurance reform and universal healthcare. And let’s say you got Bernie Sanders had become president in 2016 and had swept in a huge Democratic majority and they passed Bernie’s single-payer-for-all plan, which was by the way, much more expansive than any existing single payer plan in the world. It covered much more than Canada’s or the UK’s or anybody else. If you had done that, what you would have needed immediately was a huge supply increase in healthcare because you would have had a huge demand increase.

(01:17:44)
What happens if you make healthcare free? People are going to use more of it. If you make insurance much more widespread, people are going to go to the doctor more often. Well, if you don’t have enough doctors, you don’t have enough nurses, you don’t have enough surgeons, you need more. We constrict the supply of all those things using residency rules, using nursing rules, immigration rules, who can practice as a nurse practitioner? What can a nurse practitioner actually do? You have to be attentive to the supply side even if what you’re doing is aggressive redistribution. Now, there are places where these things don’t conflict, like I’d like to see a much expanded child tax credit, and I don’t think that has a big supply side implication one way or the other. But on a lot of the things we’re talking about, even if what you want to do, and it is often what we want to do, to do more redistribution, if you’re redistributing some kind of thing that gives you access to a good or a service, you need to expand the good or the service.

(01:18:32)
We do rental vouchers, giving people rental vouchers in the San Francisco housing market, unless you build more housing, just creates something that our friends at the Niskanen Center call cost disease socialism, where you are increasing demand for good at which you’ve constricted supply. If you do that, you’re just going to drive the price up. To some degree, that’s at least part of the story of higher education. We give people Pell Grants, we give people all kinds of subsidies for higher ed, but we have not done nearly enough to increase supply or regulate the way in which colleges just then pocket part of that money. And so they’re building these fancy gyms and they’re competing with each other, but they’re not actually increasing the supply of slots. It’s certainly not the level we want them to.

(01:19:12)
So there’s a lot here where I think it scrambles traditional categories. You cannot do effective redistribution in ways that we would like to see them done, and many people on the left would like to see them done if you’re not taking supply of the thing that you’re subsidizing seriously.
Derek Thompson
(01:19:26)
If you think about it from a first principle standpoint, what if what you wanted to do was to bring American poverty as close to 0.0% as you possibly could? You got a bunch of smart people into a room. You said, “What can we possibly do?” In my opinion, the answer would include a lot of what are sometimes called demand-side policies, a lot of redistribution of income. The child tax credit, I think would be essential. Expanding the earned income tax credit would be essential. Expanding cash welfare might be essential. Certainly, redistributing income from the rich to the poor would be essential. These are demand-side policies, they’re tax and spending policies. But if you only approach this subject through the demand side, you will utterly and categorically fail. Because like we said, housing is the biggest part of a typical family’s budget. And if your only policy on housing is to increase housing vouchers without increasing the supply of housing, macroeconomically speaking, there’s only one direction for housing prices to go, and it’s straight up to the moon. You’re not actually bringing housing prices down. You’re just subsidizing a constricted market, and therefore creating enormous inflation.

(01:20:41)
You have to solve some of these problems on the supply side.And one of the conceptual scoops that Ezra and I are trying to work out for the left, for liberals in America is to get people to ask the question, “How do we solve this problem with supply? Housing is a crisis. How do we solve it with supply? Energy is a crisis. How do we solve it with supply? Medical innovation, scientific discovery is a massively interesting phenomenon. We don’t even understand how it exists, really. How do great discoveries actually happen? Is there a supply-side policy for that as well?” That’s the question we’re asking over and over in the book.

(01:21:19)
So at the end of the day, demand-side progressivism and what Ezra calls supply-side progressivism really are peanut butter and chocolate. They are two flavors that go beautifully well together, but here’s the problem that I think your question was putting your finger on. I think a lot of Americans don’t believe that these things work together because what they see is a liberalism that just taxes and spends and Americans don’t see the benefits of that spending. The Biden administration authorized $42 billion to build rural broadband in America. Practically, none of it was built. Authorized $7.5 billion to build EV charger stations in America. Practically, none of it was built. How many tens of billions of dollars have been authorized to build high-speed rail in California? Practically, none of it exists. You can’t ride it.

(01:22:05)
So the problem is when the reputation of a tax and spend liberal makes contact with the fact that people don’t see the results in the physical world, like where’s my money going? I have in my head something like this idea of what I call Equinox liberalism, which is to say there’s some forms of liberalism where it’s very expensive, but you see what you’re getting. Like when you spend $270 to go to Equinox for the month, right? It’s a really expensive gym bill, but people who go there seem to love it. They’re like, “The equipment is always free, everything is clean. I go into the locker room, there’s a bunch of Kiehl’s lotions to put on my face after I shower. I am getting exactly what I’m paying for. Yes, I’ll pay out the nose for a gym because I love seeing that money going to work.” And in places like Sweden, Denmark, citizens seem very happy. They’re paying much higher taxes than people are in America, but they’re seeing where the money’s going to work.

(01:22:57)
The problem with the liberalism that blocks rather than builds is that people don’t see the money going to work. All they see are the dollar signs being spent by government, and then they walk out of their house and they see collapsing infrastructure and they see crime and they see housing prices going to the moon. And so they think, “Wait, this social contract is broken down. You’re asking for Equinox prices, but you’re giving me a shit-ass gym and that’s unfair.”

(01:23:25)
And so what we’re trying to say is… In a very serious way because I’m not trying to be flippant about gyms. In a very serious way, what we’re trying to say is a part of this problem is that you Democrats have looked so hard at the demand side of the ledger that you’ve forgotten how powerful the levers are on the supply side. And if we can just pull those levers on housing in particular, we can bring down the cost of living and people might even support the tax and spend model more because they’ll feel like they’re participating in Equinox liberalism and not its opposite.

Housing problem

Lex Fridman
(01:23:55)
Can we zoom in on the housing problem? Can you explain the housing problem and what’s the importance of housing in the quality of life, in the flourishing of the nation of the United States in general, and what is broken about it?
Derek Thompson
(01:24:12)
Let’s take the second part of the question first. What’s so important about housing? We’re talking about life, and even at a higher level, we’re talking about freedom, like the freedom to live where you want to live, the freedom to feel like the good and achievable life is actually good and achievable. This is profoundly a question of housing. There’s this great paper that was written several years ago called The Housing Theory of Everything by Bowman, Southwood, and I forget the other guy’s name, but they made this really beautiful point that no matter what you care about in terms of public policy or politics, housing probably makes contact with it. If you care about innovation, innovation, as I said, it’s about getting people together in a city where they can work together. That’s about housing density. If you care about childcare costs, that’s about bringing down the cost of buildings. That’s also about housing. If you care about being able to live near your friends and family, this is also profoundly a question of housing.

(01:25:05)
So when we talk about housing over and over, we’re not just talking about the four walls and roof and floor, we’re talking about what housing means to people because housing is life. And right now, what we very clearly see in the data is that Americans are leaving expensive cities and states that tend to be run by Democrats and they’re moving to areas that are sunnier, that are cheaper and are often more likely to be run by Republicans. And this is, I think because starting in the 1960s, 1970s, you had this era of blockage points in housing. We started seeing zoning regulations, we started seeing historic preservation rules. We started seeing laws that made it easier for citizens to sue to stop a new development from going up. Essentially, new tools were invented to empower people’s natural conservatism.

(01:25:57)
For hundreds of years, people might’ve always felt like, “I want my neighborhood to stay the same.” That might’ve been something that the Neanderthals were feeling, but only in the last 50 years have we really outfitted human beings with a weapon to go along with their biological preference for the familiar such that they can utilize it to stop the change of the physical world around them. And so we’ve seen the rise of what is commonly called NIMBYism, Not In My Backyard, the rise of a movement to stop the development of new housing around where they live. And so what you see in the data is according to one study that was reported on by Yoni Applebaum, my colleague at The Atlantic, if there’s a city with a 10% increase in the progressive vote share, there’s a 30% decline in the number of houses that are permitted. For some reason, it does tend to be these areas that are more populated by progressives that have more of these choke points.

(01:26:50)
Some of this is just historical quirk, but some of it is, I think maybe the character of a certain kind of liberal, maybe often an older liberal, who believes that preserving the physical environment is the good, that the best thing you can do for the planet is to stop things from changing around you, stop things from being built. And that’s an old-fashioned version of environmentalism that we’re trying to turn the liberals against because really, we see the challenge with climate change is not how do we get everybody to stop using electricity? How do we get people to stop using any power or electrons that come from natural gas and oil? We want people to recognize that if the people want to lead modern lives and they’re going to continue to demand modern lives, and so we have to use clean energy technology to allow them to live those lives. That means you have to build an absolute shit ton of new clean energy, solar and wind and geothermal and nuclear, and maybe in the near future, even fusion.

(01:27:52)
It requires an attitude of being excited about building new things, rather than a liberalism that seeks always to block changes to the physical environment around them. So I think that what we’re trying to do is to allow people in a way to meet processes and outcomes. I think liberals want housing abundance in some part of their brains, in some part of their hearts. I think they want the price of housing to go down, but there hasn’t quite been a really clean articulation of a mindset or paradigm-shifting argument that gets them to see how the processes that we’ve created go so dramatically against the outcomes that liberals want. And we’re trying to say, “Here’s a new process. Here’s a new question you can ask yourself. How can we solve the housing problem on the supply side?”
Lex Fridman
(01:28:41)
And also, solving the housing problem is one of the mechanisms to lift people out of poverty. So there’s a lot of goals that align well with the liberals. You write about cities, “Cities are where wealth is created, not just where it is displayed. They are meant to be escalators into the middle class, not penthouses for the upper class.” So the housing problem we’re talking about specifically, or most importantly is in urban areas.
Ezra Klein
(01:29:06)
Yeah. This is something that we spend a lot of time thinking about for the book. There’s a lot of land in the United States, and sometimes you’ll hear people say something like, I’ve heard this a lot, “Well, not everybody can live in San Francisco or Venice and Los Angeles or Manhattan or Brooklyn or whatever.” And fine, right? Not everybody can. That’s true. The problem is that cities are engines of opportunity and economic dynamism. As you know better than we do, the frontier AI labs in the US, basically all of them outside of China exist in 50 square miles in the Bay Area. There’s not one in New York, there’s not one in Dallas, there’s not one in Chicago, there’s not one in Austin. There’s not one anywhere but in the Bay Area, but the key thing is that there is huge spillover from these hyper-dynamic industries, be it finance in New York, be it tech in San Francisco, be it culture in Los Angeles.

(01:30:14)
You make a lot more money being a barber near Google than you do being a barber in rural Arkansas, and this is the ancient pathway to mobility. People who are poor and work in the service sector, they move to richer areas where the productivity is higher, and thus the money spreads around. And if you look at American mobility and opportunity over the 20th century, about a third, roughly on some calculation, just comes from this people moving to richer areas. And if you look in the last 20-ish years, 30-ish years, you see this extremely steady process of income convergence as people move towards richer areas begin to throw itself into reverse because it used to make sense for the janitor and the lawyer to move to New York City, but now it only makes sense for the lawyer to and the janitor leaves because you can’t support yourself and your family and live in a home and have a reasonable commute being a janitor working in Manhattan. Now, obviously some people do it and it’s often very tough and they live very far. It’s very hard in San Francisco.

(01:31:27)
You can’t look at cities as, “Well, that’s just where the rich people are,” and that’s a problem that many Democrats now have. Michael Bloomberg famously described New York City as a luxury good and luxury goods cost luxury prices, and that’s true in the sense that New York became a luxury good, but that’s a problem. That is a terrible problem and an inversion of what the city is. There’s this famous advice, possibly Apocryphally from Horace Greeley who’s a early American newspaper publisher and political candidate, but he says, “Go west, young man, go west.” Right? The opportunity is out in the west where the lands are open. It’s never been true, including for him. That guy moved from a rural area to New York City, and that’s how he became famous, and he ran a newspaper and he ran against Ulysses S Grant in a presidential election.

(01:32:16)
The cities are the frontier. The cities have always been the frontier, not of the land, but of the economy because the frontier of the economy is where ideas are produced and ideas, even now, even the age of remote work are produced in the big cities where people live together and they compete with each other and they cooperate with each other. And so if you gate the cities, if you make it impossible for someone making 50,000 bucks with two kids to live in the city, then what you’ve done is you’ve actually closed the American frontier. You have forced them into lower productivity places. Their children are less likely to grow up around the inventors in the cities.

(01:32:50)
There’s really amazing research from Raj Chetty and others, basically showing that kids, no matter what income class they’re in, they’re likelier to grow up innovating and patenting in the innovations of the place around them, right? Smart kids don’t just grow up and innovate in anything. If they grew up in the Bay Area, they innovate in technology as Steve Jobs did, and Wozniak did, right? Because they just lived around those people. And that is true in many, many, many different things.

(01:33:16)
And so when you gate the cities, housing is almost too small of a thing for what we’re talking about. We’re talking about if you can live next to economic opportunity. We are talking also about if you can put the people together who will create the next era of economic opportunity. Right now, the Bay Area is still in some ways drafting off the back when it was cheap. A lot of the people we’re talking about who have made amazing things, they started in the Bay Area when you could afford to live there. It wasn’t always like this, and it wasn’t that long ago that it wasn’t like this. And now, fine, you can go there if you have money or you have a great job offer from Google or Apple or whomever, but over time, you need the ferment. I have a personal interest in reading memoirs and noticing that the memoir is really a housing story. One of the ones I like about this, Moby, the electronic musician, his first memoir, which is great, it’s a memoir of a certain era in New York, it’s a housing story. He was just living next to a bunch of other musicians in cheap-ass housing. I just read Meet Me In The Bathroom by Lizzy Goodman, a sort of oral history of the Aughts Rock Revival in New York City, a housing story. They could afford to live here. Right? I’ve read a bunch of these in San Francisco too, where people just like, they’re actually squatting, right? The San Francisco ferment, its queerness, its openness and tolerance of new ideas, its home of the psychedelic counterculture that intermixed with the defense culture that created Silicon Valley, right? You’ve read probably… Is it Frederick Turner’s From Counterculture to Cyberculture? That was a story of cheap housing. You need to allow people to be around each other, to mix each other. If only one type of person can afford to be there, it becomes a monocultural over time. And so it’s not just housing. This is about the geography of economic innovation and opportunity. That’s why it’s so fucking important.
Lex Fridman
(01:35:07)
Yeah, that’s so brilliantly put that the economic dynamism and the intellectual dynamism that makes America, I think the greatest nation on Earth is at the core, a housing story. It’s like you have to live near the place where there is turmoil and turbulence intellectually and economically speaking. And so you want to be able to move there and be part of that culture, part of that economy and part of that, how you raise your kids and culturally what you want them to study, what you want them to do. Yeah, that’s fascinating. So how do we solve the housing problem? Is it just remove as much regulation as possible, like get out of the way?
Derek Thompson
(01:35:51)
There’s certainly a lot of regulation that you want to get out of the way. You look at California, for example. You were just giving a beautiful summary of just how important it is for people to be able to move next to economic opportunity. If you look at the number of houses that have been permitted in the State of California, over the last 40 years, it’s basically just a squiggle line down. The tragedy here, just to put a really fine point on it, is that the city wasn’t gated by geography or by destiny.
Ezra Klein
(01:36:18)
Or by technology. We know how to build apartment buildings.
Derek Thompson
(01:36:19)
Yeah. Yeah. Elisha Otis invented the elevator in like the 1850s. This is not exactly a breaking technology. We chose to do this. We wrote these laws. We people are filing these lawsuits. We judges are accepting these lawsuits and determining that this building can’t be built. This is entirely self-inflicted. And it’s why over and over again in this book, we call it a manufactured scarcity, which is like a little bit of a funny term. How do you manufacture the absence of something? No, a manufactured scarcity means you didn’t have to do this. This is a human-made rule whose purposeful goal was to make it harder to add to the supply of something. And in this case, especially since the 1960s, we have made it purposefully difficult to add housing supply, and the outcome just follow the process. You see housing supply decline often in the richest cities and these states that are governed too often by progressives. How do we undo it? Yes. A huge part of it exists at the layer of law. Right? California is already trying to change its laws, right?

(01:37:22)
And single-family zoning, make it easier to add accessory dwelling units called ADUs. We’re changing this at the level of law, but one thing I’m very fixated on is making sure that we also communicate to people that it has to be changed at the level of mindset, and even at the level of political courage because here’s like a model of what often happens. You’re the Mayor of some city and you want to add a housing development of let’s say 500 units. Right? 500 apartment unit building is going to be set up by some developer, and there’s a City Council meeting to determine whether or not this apartment building is going to be built.

(01:38:02)
Because of loss aversion and because the people who tend to go to City Council meetings are older and richer and homeowners, guess what the overwhelming volume of reaction is in those City Council meetings? It’s a lot of people who feel like they have something to lose, saying, “You cannot put up this building, you cannot add these apartments. It’s going to ruin the character of the neighborhood. It’s going to create traffic. It’s going to be like an eyesore because I don’t like buildings that are that tall.” Any number of excuses can come out of like the Pez dispenser of excuses. You take care of one, there’s going to be another that comes to the surface.

(01:38:37)
And so what often happens is that these Mayors or these people sitting in the City Council will look out into this room of 50 people saying, “No, no, no.” And they’ll say, “I’m going to represent the feedback that I see and I’m going to vote no on the addition of these housing units, the addition of this apartment building.” Where political courage comes into play is the ability to say, “You know what? I want everyone here to know that I hear you. I’m listening to you and…”
Derek Thompson
(01:39:00)
I want everyone here to know that I hear you. I’m listening to you and I represent you. And so, I’m grateful that you showed up to the city council meeting. But for every one person here who says they see a benefit to adding a new apartment building, I know that there are 10,000 people in the city that didn’t have the wealth or the knowledge to be here at this meeting. And they’re going to benefit from more housing because that’s going to bring down their housing costs, and I represent them as well.

(01:39:28)
I don’t just represent, you could say the circle of care that I can see right now. I also represent the circle of voters that we can’t see right now, who are in the city or who are in the state or might even want to move to the state, but currently can’t because housing prices are too high on account of housing supply being restricted.

(01:39:48)
And so, one thing we’re trying to get liberals to have is a sense of political courage to stand up against this very visible NIMBYism and say, “We represent interests that aren’t necessarily visible at this city council meeting. We represent the larger interest of housing abundance.” So, we’re going to always default to saying yes rather than default to saying no, just because the people who happen to come to these meetings are NIMBYs.

Regulation and deregulation

Ezra Klein
(01:40:14)
I want to add a wrinkle on regulation and deregulation. So, we were talking about coding earlier, who gets coded as right wing? Who gets coded as left wing? Deregulation is a word that is highly coded as right wing. The right wing wants it to deregulate, right? They want the government to stop regulating the market. It’s fine. In many cases, the government should deregulate parts of the market. In many cases, it should regulate parts of the market more. What we don’t talk about enough is how much the government regulates the government and how badly it needs to deregulate the government. So, I have many more left friends and they’ll come to me or they’ll critique me and they’ll say, “This all is fine, but what we really need in this country is public housing,” or it’s been rebranded social housing. It’s fine, Singapore, huge amount of social housing, right? They do a great job with it.

(01:40:59)
One of the things we go into, and this book is a manifesto on some level, but something we really try to do is take you into the gritty, grimy, frustrating details of how policy plays out on the ground, what actually happens after a bill passes, and why we get the outcomes we do. Because often, it’s like a bunch of decisions made after everybody stopped paying attention.

(01:41:21)
So, one of the things we pay some real attention to is the kind of housing that people on the left all agree on is affordable housing through government grants, right? The government says, “Oh, we’re not just going to have market rate developers coming in, building more luxury condos for the children of the upper class. We’re going to build affordable housing for people who should be in this city, but otherwise couldn’t pay enough to be here.”

(01:41:47)
California. I can give you two different examples, but let’s look at Los Angeles. I’m from outside LA and they have something… I always forget if it’s Measure H or Measure HHH. But California, LA voters pass a bond measure about a billion dollars a little bit more, I think it was to build affordable housing. Six years later when I’m writing about this, they have built a couple of thousand units at an average cost of $600,000 to $700,000 a unit.

(01:42:13)
So, it’s costing more to build housing under this affordable housing bond measure, which they have agreed to pay for than it is to buy a home market rate in Denver. Denver is a nice place to live. So, why? What had happened? Well, it turns out that when you trigger the public money in various cities, these are city ordinances often, but not always.

(01:42:35)
And you use these public grants and you cobble together the different grants you need to build an affordable housing complex. What you’ve done is layer onto yourself a huge number of rules that the market rate developers don’t have to follow. You’re either using union labor or paying prevailing wage. You’re building to higher green building codes.

(01:42:52)
Oftentimes, by the way, to get through the kind of planning board meetings that Derek is talking about, you’ve made a bunch of concessions on the design. Is there going to be parking in it, security? Things like that. You have to often agree to who’s going to be in the home. So, you’re getting… There’s a thing in the Measure H stuff where, well, they wanted it to not just be the taxpayer money. They also wanted nonprofit grants so the money would go further.

(01:43:14)
So, you’re trying to get these other grants, but these grants are to house homeless veterans. So, now to open the thing, you need to find these homeless veterans. You need to go through extra disability accessibility reviews. And, of course, we all want these things to be accessible, but they already had to comply with the American Disability Act. But now, you’re doing another disability act review in the city, and they come in and they say, “Well, your doors are a little bit not as wide as we think. So, you got to make all your doors 3 inches wider before you can open up,” and that adds time and that adds cost.

(01:43:39)
You have these subcontractor rules, right? This is, now I’m using an example from San Francisco. But you had a preference initially for minority-based subcontractors that became illegal. It became small businesses, but that meant it had a preference against the bigger contractors who were more efficient at building housing.

(01:43:54)
In order to use public money, and then very much in order to build public housing directly, it ends up being more expensive and slower than market rate, and that is a choice. We do not have to try to solve every problem in society through an individual housing project. Building affordable housing is hard enough. It’s not impossible, but it’s difficult to do. You do have to talk with the neighbors, right? It’s never going to be trivial to build a 500-unit apartment building.

(01:44:23)
But instead, we layer on all these external and additional agenda items. I call this everything bagel liberalism because you sprinkle just enough on the bagel and it’s great, but if you saw everything everywhere all at once, you put everything on the everything bagel, and it becomes a black hole from which nothing can escape.

(01:44:39)
And so, the need to deregulate government, right? Why? I saw some of Elon Musk’s marathon appearance on your show, and he talks about high-speed rail in California. And the thing he says there that it’s functionally illegal to build high-speed rail in California. I got 100 disagreements with Elon Musk right now. That’s not one of them. It’s functionally illegal to build high-speed rail in California.

(01:45:01)
Like I went out, I toured the high-speed rail. I’ve done a lot of reporting on that. You can’t build it affordably. You just can’t on time. They have no way to get the money to build the rest of it. It’s not going to happen and it’s not going to happen, not because you can’t build high-speed rail. Europe builds high-speed rail. Japan builds high-speed rail. China builds high-speed rail. And it’s not because of the private market, like we’ll only build luxury high-speed rail. It’s because we have so heavily regulated a public project that you can’t finish it, and you definitely can’t bring it in on time and affordably.

(01:45:31)
And so, I really would like to de like, uncode, deregulation because yeah, there’s places where I would like to regulate more. I don’t want you to be able to build a coal-fired power plant in America. I just don’t want it to be possible. I think it’s bad. But I do want it to be possible to build high-speed rail and affordable housing, including through the public market.

(01:45:49)
And one thing, my friends on the left, I think really underrate is how hard they’ve made it for the government to act. They believe in government, but if you believe in government, then you have to make it possible for government to complete projects. You have to make it possible for the people who work for government to apply their own genius and initiative. They have to have agency.

(01:46:09)
I always say that like we’re sort of trapped right now between a party that wants government to fail and a party that won’t make government work. And like, we are trying to push this idea that the thing we want is a capable government, a strong government, a government that when it promises it will do something, it actually does it and gets it done. And that requires not just deregulation of the private sector, though sometimes it does require that. It requires deregulation of the government itself.
Derek Thompson
(01:46:34)
I love this question of deregulation. But I also sometimes find it very frustrating, because sometimes I find that people’s sense of regulation is so specifically coded. Regulation is just rules. If you change the word from regulation to rules, I think it’d be easier for people to see.

(01:46:51)
Some rules are good and some rules are stupid. We all understand that in life. That’s what regulations are. They’re just rules. And sometimes, they give us exactly the outcomes we want. And sometimes, they give us the outcomes we would never hope for. And here’s a good example. You go back to the 1940s and 1950s, America was fucking disgusting. The air and the water was horrifying.

(01:47:15)
In 1943, residents of Los Angeles woke up to a smog that was so black. They thought the Japanese had launched a chemical attack against America. In New Hampshire, in the rivers next to textile mills. Sometimes, the rivers themselves would run green and purple and red depending on what textile colors were being dumped into the river. You had the Ohio River on fire in the 1960s, 1970s. The world of mid 1900s America was truly sickening, and it made people sick.

(01:47:45)
And so, we passed a raft of environmental rules, and some of them achieved exactly what we wanted. The air we breathe and the water we drink is cleaner because of the Clean Air and Water Acts. We did extraordinary things with this era of regulation. Some of these regulations were about outcomes that the Clean Air and Water Act regulates specific pollution levels in the air and the water or the emissions coming out of tailpipes.

(01:48:08)
Some of the regulations though, were about process. NEPA is the National Environmental Protection Act, and it among many other things, empowered individuals and citizens to sue the government and organizations, businesses, to stop construction or fill out environmental reviews to prove that their construction would meet muster wouldn’t degrade the environment too severely.

(01:48:33)
This opened the door to an infinitude of lawsuits to wrap up any effort to build anything in red tape forever. So, fast forward to 2021 the year, or sorry, the month after Newsom’s signs, Gavin Newsom, Governor of California signs the law to end single-family zoning in California. It seems like a massive win for the pro-housing YIMBYs of California.

(01:48:58)
The month after that law was passed, the board of supervisors in San Francisco decided to rule against an apartment building that would have added 500 total units and 100 below market rent units. So, something like public or social housing that was going to be built on a Nordstrom parking lot, just about the best possible place you could add housing in the world, a Nordstrom parking lot because the builders, the developers, didn’t file the appropriate paperwork under environmental review.
Lex Fridman
(01:49:28)
Right.
Derek Thompson
(01:49:29)
So, this is a world in which the most housing starved city in America is being starved even more of housing because of the expression of, or the power of the instrumentalization of a rule past the 1970s that has allowed people to sue to stop the physical world from being changed. And I think it goes back to this idea that like sometimes the solutions of one era can become the problems of the next generation, right?

(01:49:54)
It was really good to pass the set of environmental bills that we did in the 1960s and 1970s because it addressed the extremely real problem of air and water and land being degraded by industry. But we’re in a new world, and the problem environmentalism today is in part a problem of global warming. And we have to build not only dense housing, but also clean energy. And the same rules that were designed to help the environment in the 1960s and 1970s are sometimes ironically used in a way that hurt the environment in the 2020s. And that’s one reason why, we, as liberals need a paradigm shift.
Lex Fridman
(01:50:32)
Okay. You said a lot of really interesting insights there. So, one, if I understood correctly that regulation of outcomes is more a good idea than regulation of process, because regulation of process is where you can breed a lot of, basically the lawyers show up. And the other insight is, if we get rid of 99% of lawyers, the world would be a better place. There’s a lot of jokes around that. I don’t know if you agree, but that…
Ezra Klein
(01:50:57)
I definitely wouldn’t say 99%.
Lex Fridman
(01:50:57)
98…
Derek Thompson
(01:51:00)
Yeah, yeah, 82.3% to be absolute.
Lex Fridman
(01:51:02)
In the 80s, okay.
Derek Thompson
(01:51:03)
No.
Lex Fridman
(01:51:03)
Majority of them, no, they’re simply there to take advantage of the rules.
Ezra Klein
(01:51:10)
We talk a lot about a book. I don’t know if you’ve run into this one. But it’s by Mancur Olson, who’s sort of a founder of public interest economics, and it’s called The Rise and Decline of Nations. This is a very famous book. Libertarians love this book, and I love this book. It’s not right about everything, but its fundamental question is, how come after World War II did the completely destroyed, bombed out countries of Germany and Japan? You would have thought they would be just screwed.

(01:51:38)
Instead, they both become growth miracles. They both do much better than the UK, which was on the winning side of the war. Why? And Olson’s argument is that affluent, stable societies over long periods of time, develop something that is very difficult to develop and very important to develop, which is bargaining organizations, right?

(01:51:56)
Collective action is hard. It is hard to form an organization. It is hard to make that organization persist. But if you can do that in an atmosphere of stability, then over time, that organization will tend to entrench its power, right? Think about AARP or the Chamber of Commerce or certain unions or the National Manufacturing Council, et cetera, the business roundtable. They’re not that powerful when they start, but over long periods of time, they become really powerful, and they start to pass laws and make themselves more powerful and so on.

(01:52:26)
There’s a lot you can say about this insight, but my favorite part of Olson’s book, and one that I don’t think people emphasize enough, is this insight he has, which is that over time, countries will begin. Every country has a kind of form of natural selection within it, and that form of natural selection will select for people with the skills to navigate, best navigate the kind of economy that country has.

(01:52:48)
So, if you’re in China right now and China’s in its developmentalism phase, you really want to be in civil infrastructure, right? It is great to be a civil engineer in China, great to be working on semiconductors in China, great to be doing all this stuff in the physical world. But America as a kind of price of our affluence and our success, we’ve become a country, and this happens to a lot of countries where negotiation is really important. And a country in which negotiation is really important is going to, over time, start developing a preference for lots of lawyers, people in finance, management consultants, because it is a society that requires continuous, what he calls complex bargaining.

(01:53:27)
And I think this actually explains a lot. Patrick Collison, we quote him in the… Who’s the founder, CEO of Stripe, brilliant tech guy. Here’s his point that he made in an interview with Noah Smith, who’s a blogger and economist where they were talking about high-speed rail, and sort of Patrick makes this point. He’s like, “It’s just tough to be a high-speed rail engineer in America. You’re going to have a much easier time working in the digital space.” And so, the digital space becomes a kind of frontier of last resort.

(01:53:54)
And so, people want to build things, go into bits and bytes, not into atoms, right? To use the old kind of Peter Thiel cut. And I think there’s something to that. It’s not that lawyers are bad. Some of the people I love and most in this world are lawyers. And many lawyers do amazing, amazing work. But one reason we’ve selected for so many lawyers and America has a lot of lawyers, is we became a society that needs a lot of lawyers because we are a society that is stable, affluent, and we’ve become very into bargaining.

(01:54:21)
Donald Trump is a real estate developer. That’s a relational business. The way real estate development works in this country, the reason you don’t have all that many huge firms building housing in many, many states simultaneously is it requires a lot of relationships in the individual city you’re in.

(01:54:35)
And so, if you read, say, Maggie Haberman’s, great biography of Donald Trump, Confidence Man, which focuses a lot on his time as a real estate magnate in New York. You see Donald Trump doesn’t build a ton of stuff in other states. He actually builds in other countries sometimes or more to the point he puts his name on things built in other countries. But really, what he did was build here and he built here through his relationships with the New York political system.

(01:54:57)
And so, over time, societies that make building and construction and creation something that is the output of negotiations rather than very clear standards and rules where if you’ve done it, you can just do it. You get a lot of lawyers, you get a lot of management consultants, you get a lot of finance people because that’s what society is selecting for. That’s what you’ve made it possible to have agency and freedom in. That’s what you need to do, to do big things.

(01:55:23)
This is coming from Nick Bagley, who’s a University of Michigan law professor. But between Walter Mondale and Tim Walz, there’s not a single person on a Democratic presidential ticket who didn’t go to law school, not one, like the Democratic Party in particular is a party of lawyers. And lawyers look at things through the legal perspective. And the legal perspective is that government is legitimate by following process.

(01:55:46)
And Bagley’s point is that government attains legitimacy, at least in part through outcomes. And when you prize process so high over outcomes, you think you’re acting legitimately, but actually it what would have made you legitimate in the view of the people is that you built the thing you told them you were going to build, you made it possible to live affordably in the city. And so, you have this sort of movement then over time to populist strong men who say, I alone can fix it, because they’ve kind of given up on this procedural liberalism like they didn’t deliver for them.

(01:56:12)
So, you keep telling them, we’re the ones who know how to run government and they don’t see it. And eventually, somebody else comes and says, “I’m going to bust through the walls of this thing like the Kool-Aid man,” and they win.
Lex Fridman
(01:56:22)
So, speaking of the Kool-Aid man…
Ezra Klein
(01:56:25)
I can see what you’re about to do.
Lex Fridman
(01:56:30)
Yeah, I’m so transparent. It makes me think I’m a robot built in a lab somewhere.
Ezra Klein
(01:56:36)
Sam Altman once said to me, he said, “Well, aren’t you just a reinforcement learning system with energy running through it?”
Lex Fridman
(01:56:44)
Yeah.
Ezra Klein
(01:56:45)
It’s like, I’d like to think not, but maybe.

DOGE, Elon, and Trump

Lex Fridman
(01:56:48)
So, on the Kool-Aid man, I was wondering if you could maybe steelman the case for and against Elon Musk and DOGE, because you mentioned all of these regulations, all of these complexities that get in the way of building, it seems like a bold human like Elon is required to break through the regulation, put that on one side and the other side. I read somewhere that abundance is kind of the anti-DOGE, which to me, it seems like there is conflicting ideas, but there’s also alignment. So, maybe can we break all that?
Derek Thompson
(01:57:31)
I think the steelman is very easy to make here. Department of Government Efficiency. That sounds like an organization that’s needed if government is inefficient. And one of the themes of our book is just how inefficient government can be, not only at building houses, building energy, often at achieving its own ends, building high-speed rail when it wants to build high-speed rail, adding affordable housing units when it wants to add affordable housing units.

(01:57:57)
I love Ezra’s line that we don’t just need to think about deregulating the market. We need to think about deregulating government itself, getting the rules out of the way that keep government from achieving the democratic outcomes that it’s trying to achieve. This is a world in which a Department of Government Efficiency is a godsend.

(01:58:17)
We should be absolutely obsessed with making government work well, especially if we’re going to be the kind of liberals who believe that government is important in the first place. So, that to me is the sort of pillbox version of a steel case for a Department of Government Efficiency.
Ezra Klein
(01:58:32)
Wait, before you do the anti-case, can I offer a different steelman?
Derek Thompson
(01:58:34)
Please?
Ezra Klein
(01:58:35)
I think the steelman case for what DOGE is, right, rather than what it pretended to be, is that the government is an interest, the bureaucracy, the deep state, the rules, the regulations, and it’s not about efficiency, never was. You wouldn’t do this if it was about efficiency, that it’s zero-based budgeting, that you’re breaking the thing. You’re turning it on and off. You’re firing massive parts of it.

(01:59:04)
Because the only way to make change within it possible is to delete what currently exists, whether it was efficient or not. You would never actually know that if you had it all come and present its case for efficiency or something, you’d know you’d get turned around, whatever. The only way like, the problem with the government is there is no actual competition.

(01:59:25)
The Department of Education doesn’t get outcompeted by the agency of education, which has started up three years ago or something. And because of that, the only way to make possible radical change is to bulldoze the thing that currently exists. And then, once that is done, you can begin to rebuild. If you’ve fired half of the Department of Education, then you can start rehiring your people and they will actually do what you want. If you have shown that you can delete every regulation or just not follow it, then you can begin deciding which ones to actually follow. If there is no department of USAID and you’ve moved back under state, then you can tell state what really to fund in terms of foreign aid, right?

(02:00:05)
There’s a theory here, I think that was never about efficiency. It was about deletion. He’s not trying to make things run a little bit better. He’s not trying to lower the overhead cost of government. That the theory is that in the first term, the bureaucracy impeded Donald Trump. It didn’t listen to him. Bureaucracy is supposed to be limbs of the President. The only way to make the federal government a neural link of Donald Trump himself, is to destroy the federal government. And then, rebuild it as that thing.
Derek Thompson
(02:00:34)
I think if you talk to people at DOGE or talk to people who are authors of Project 2025 who are at Heritage, who are chiefs of staff, and the people working for Heritage, if you have a truth serum conversation with these folks and you say, “Defend what’s happening?”
Ezra Klein
(02:00:52)
Oh, that’s not up. They’re excited to tell you.
Derek Thompson
(02:00:52)
This is what they’re saying. They’re saying, something metastatic has grown inside of government, not just over the last few years under Joe Biden, but over the last few decades, maybe going all the way back to FDR and even Woodrow Wilson. We have allowed an administrative state to accumulate like barnacles on a ship around the executive branch. And it’s keeping the executive branch from being able to translate. The Democratic will into policy because there’s never any president who’s purely elected by the people.

(02:01:21)
They’re elected into an office that is already pre-contaminated by the bureaucracy itself. And we’re trying to take all of that away. We’re trying to, I mean, this is very just explicitly the case they’re going to make, trying to make government more democratic, not less, by allowing the democratic, the elected president guide or lead in a pure way.
Lex Fridman
(02:01:44)
Okay. So, as you said, two things. One is the steelman, and then at the end there, there was a non-steelman. So, the first part is cutting, removing as much as possible to see what’s actually needed. That seems like one of the ways to figure out what’s actually needed is by removing it. And then, there’s the second thing is that you mentioned so that you can install the people that follow your policy, whatever.
Ezra Klein
(02:02:12)
Oh, I disagree here that that wasn’t the steelman. I mean, I think you have to listen to them to do the steelman, right? I don’t think the steelman is imaginary. If you read the OMB regulation that froze funding, it said explicitly that the government has to reflect the will of the people as expressed through their choice of president. If you read anything, Russ Vought has ever said, they have the unitary executive theory.

(02:02:37)
I’m not saying it’s right or wrong. If you asked me to do the non-steelman, which I would love to do too. It is not my view that the steelman case of this is to make the federal government fully responsive to Donald Trump. But the steelman case for what they are doing as expressed by them. And I think a steelman reflects, I mean, I have talked to these people, right? I’m telling you what they tell me on some level. The steelman cases, they believe, like as Derek said a minute ago, this has become non-democratically responsive, and the way it becomes democratically responsive is deeply responsive to the president. The president represents a different politics than Joe Biden’s politics. And so, if you have a state filled with liberal civil servants who don’t want to do what he says, that is a violation of small democratic principles of how the government should be run.

(02:03:25)
I think if you don’t like it, that’s fine. But I do want to say, I actually think that is the steelman. They’re not doing this for no reason. They have an intention here, and I think whether you like it or not, so it depends on whether or not you like their view.
Lex Fridman
(02:03:40)
Okay. Well, I don’t like it because it cuts my ear in a certain way that one of the criticisms I have for Donald Trump and the Trump administration is, there’s a natural circles of sycophancy that forms, and every president has their personality and psychological quirks, and I think he is one of the people where favoritism is more likely to develop.

(02:04:02)
So, when you choose who to install as the head of whatever organization, so if you fire everybody and rehire, the rehiring process is more likely to have people that just said nice things about Donald Trump in the past versus a meritocracy-based system that these people are really good.
Ezra Klein
(02:04:20)
Lex, AOC should definitely come on the show.
Lex Fridman
(02:04:23)
Well, she has her own. But the ideal, the steelman to me about not maybe what Elon is doing is, in order to have the best people in the world doing a great job at every part of government, you have to figure out, I mean, his first email of like, “What have you done this week?”

(02:04:47)
If we just steelman everything he’s been doing, which is like, let’s find the productive people that show up to work that love what they’re doing that are actually sort of amazing at what they’re doing. I mean, how would you solve that problem from his experience in business? It’s painful, but effective to just fire almost everybody. And then, rebuild from there and continuously do that.

(02:05:09)
And as the result, as long as everybody is aligned in a mission is you’re simplifying over and over the system. So, it’s more and more and more effective. Now, how else would you approach… We could start now criticizing, how would you make government more efficient, if not by the way that they’re doing it?
Ezra Klein
(02:05:34)
Yeah, can I steelman to the other side now?
Lex Fridman
(02:05:36)
Sure.
Ezra Klein
(02:05:37)
Okay. You wouldn’t do this. You wouldn’t do this.
Lex Fridman
(02:05:43)
Yeah.
Ezra Klein
(02:05:43)
So, let me say a couple of things. One is that efficiency only makes sense when yoked to a goal. So, when… If Elon Musk came in and did this at Tesla, Musk had a goal for Tesla. It was to build the electric vehicle of the future. SpaceX, he had a goal for SpaceX. We know what the goal for SpaceX is in some long-term way, go to Mars, but in some short-term way, cheaper orbital travel, cheaper orbital shipping, reusable rockets, et cetera. We know what the goal for SolarCity was.

(02:06:16)
In a way, and I do think sort of the purchase of Twitter is late Elon, which becomes a much more political set of goals. But nevertheless, in a way, at least there’s an expressed idea at Twitter, which is the return of free expression. I think it’s really important if you are steelmanning or any kind of manning is really asking and listening to what people are saying their goal is, because efficiency does not exist in a vacuum, right?

(02:06:41)
A state that is efficient at building dirty energy and a state that is efficient at burning clean energy, they can be opposite versions of each other, right? You have to be optimizing towards something. You’re an AI guy. You have to optimize towards something. And the goal… I find myself really consistently frustrated by conversations about DOGE that treat efficiency as some free-floating thing. Or I sometimes will hear people say to me, very smart people have said to me, “Oh, what DOGE is really doing is stripping the government down to studs, so it can put AI into it, because Elon believes in AGI is coming soon, and you need to make a government capable of using AGI.

(02:07:18)
And I always say to them, “Okay, AI towards what? Towards what value function? Towards what prompt are you inserting at the base level?” Sorry about that. And I think a few things become very clear. One is that it is towards Donald Trump. It is his movement. Elon Musk serves at his pleasure. Elon Musk has said that he would like to himself chisel Donald Trump’s face into Mount Rushmore. He said he loves Donald Trump as much as a man can love any other man. He believes in Trump, right?

(02:07:48)
I think you have to take him at his word, or maybe you think he’s very cynical and he’s saying all that to curry favor in a sycophantic way with Trump. But at some point, Elon Musk does not serve with any kind of independence. If Trump says, “Your power is gone,” his power is gone. He’s functionally an outside advisor to the government.

(02:08:02)
So, then you really do, I think, have to listen to what Donald Trump has said, what Russ Vought has said, and they believe that Trump represents something fundamental. His movement represents something fundamental that has been suppressed in American life, okay. I think if you were doing something… The best cases I’ve seen for Elon are a zero-based budgeting case.

(02:08:19)
You’re trying to break everything down to studs, and then it needs to rejustify itself. If he did this at Twitter, the engineers had to come in and justify to his people what they were actually doing. But that wasn’t how those emails worked. We all know this, right? He doesn’t have a staff capable of seriously working through, and then following up on emails from 2 some million government employees about what they did that week, and then really checking, well, did they do it well? Was the thing accomplished, right? That’s not how you do that like what? He’s got 50 people at DOGE? It’s nothing.

(02:08:53)
If you want to do zero-based budgeting, it has to be against a criteria. So, coming in and gutting USAID, some of the things USAID does are just extraordinary, right? PEPFAR, nobody anywhere thought PEPFAR was not an efficient program. PEPFAR is maybe the highest value thing that we have ever done in the US government to save human life through foreign aid. I mean, just bluntly, it’s a George W. Bush program. It saved a generation of people from dying of HIV AIDS.

(02:09:19)
They just turned it off, and they didn’t have somebody come in and justify the PEPFAR funding. They just turned it off as they turned all these different things off and have never given people a serious effort to come back in and say, ” This is what we do.” They just fired half of the people at DOE. DOE is, if you look at it, DOE has the lowest staffing level of any agency, but the fourth highest appropriation, and it administers the more than $1 trillion higher ed. It is in overhead terms, one of the leanest of all the agencies alongside Social Security Agency.

(02:09:51)
They didn’t tell those people what they wanted out of a DOE of half the size. They didn’t let those people come in and argue for their jobs. They’re just cutting things, and they’re not explaining at any point, at any level what they want to do with them. And then, I will also say what they’re doing is probably illegal, right? Just this week or last week, a judge said that about 26,000 people need to be rehired by the federal government because their firings were illegal.

(02:10:17)
And now, that’s going to go up to the Supreme Court and we’ll see what John Roberts and Amy Coney Barrett say, right? That game is not done yet, but they have Congress. I am a person who believes we needed civil service reform. Too hard to hire, too hard to fire, too hard to manage. And they could have gone to Congress.

(02:10:37)
And frankly, at the beginning, Ro Khanna and a bunch of Democrats that they wanted to work with DOGE, Democrats were defeated. If you had wanted to build some amount of like political bipartisanship, it was there to do, right? Like the Democrats voted for the Laken Riley Bill, the hardcore immigration bill. They were ready to deal, but they didn’t want to do that because it is fundamentally for them about political control. And this is my steelman case for the action. This is a case, I believe.

(02:11:02)
What Elon did at Twitter was not make Twitter more efficient. It’s not like a better product now. It’s a different product. What he did was he made it controllable. What he did at Twitter was he went into something where he thought the people disagreed with him, and he broke it until he had actual operational control of the thing, and they would do what he wanted them to do, and then he can move it in his direction.

(02:11:20)
What they’re doing in the federal government, and this is what Chris Rufo wanted to do, it’s what Russ Vought wanted to do, is make the thing controllable. And what they want to do with that control, I don’t even think they fully know, but their view is that the federal government is like a… It is the capital city that they have conquered.

(02:11:37)
And now, they need to turn it into something that they can actually use. And in order to do that, I mean, Russ Vought has said, “You need to traumatize the civil servants,” right? That was his word, traumatize the civil servants. When Elon Musk talked about USAID, he said, “It’s all worms. No apple. We’re going to feed it to the wood shipper.” This isn’t the language of, let’s see who’s really doing a good job here and who’s not. This is the language of conquering. This is the language of destruction, and they have not-
Ezra Klein
(02:12:00)
This is the language of destruction and they have not articulated new better goals for it and nor are they starting with the parts of the government where I think you would start with. What do they want the education department to do? I’m a professional political journalist. I talk to people in the Trump administration. I can’t tell you what they want out of it, aside from the fact that they don’t like it and they think it’s like a hotbed of other people’s power.

(02:12:22)
I can tell you what some conservatives have written about it, but the administration itself has not articulated goals. So what I think their view is right now is they’re breaking the thing and trying to make the people who would be oppositional either leave or go into hiding and then they’re going to figure out what they want to do with it.

(02:12:38)
But right now they’re in the period where you have to sack and then later on you can build yourself your monuments and turn the civil service to your own ends. I don’t think it’s still… I just… That is what I think is going on.
Lex Fridman
(02:12:52)
I think that’s, from a certain perspective, accurate and insightful description. If we just look at Twitter, you said that he didn’t make it more efficient, he made it more controllable. Let me just describe, just because I know the engineers inside Twitter well, which we could argue that the government is very different than a company. That’s an important argument there. With Twitter, the culture that’s there now, people are really excited to work there and to be productive there, engineers. Elon is really good at finding people that are there, extremely good at what they do and are there to give everything.

(02:13:37)
And the resulting thing this machine that is created, all barriers removed, you create beautiful stuff. It’s not like move fast and break things. It’s very cynical perspective on it. No, everybody loves what they do, are good at it and are really rapidly figuring out all the multitude of problems that arise in that. It’s a different culture, a culture of productivity, and it’s not like some negative boys club or there’s all kinds of negative perspectives you can take on those things. No, it’s a positive culture of productivity.

(02:14:15)
Controllable side, also really important to understand and that could be taken advantage of in a really negative way, especially in government. But you should say on some of the most successful software projects, there is a level of control required like open source projects, Linus Torvalds, that heads up the Linux kernel. There has to be what’s often called the benevolent dictator for life. There has to be a controllability to the machine of this extremely productive, efficient team to work together.

(02:14:50)
Now, those are engineering projects and the problem with government is you get elected new benevolent dictators for life that come.
Ezra Klein
(02:15:00)
They’re often not benevolent.
Lex Fridman
(02:15:01)
And they’re often not benevolent and they all convince themselves they’re benevolent. I’m sure Hitler thought he’s doing the right thing, the good thing, the benevolent thing in his worldview. Every single dictator does that, yes. So that’s the problem, but in order to be efficient, there has to be some level of controllability and there has to be, I think, to go back to the steelman, you have to have a culture of productivity that I’m not sure I’ve had quite a bit of experience with DOD and DARPA from the academic side.

(02:15:36)
Just everything is so bureaucratic and slow. That’s a different culture. So part of the destruction is in reestablishing culture and yes, make it more controllable so you can be efficient.
Derek Thompson
(02:15:50)
We still need to define what the outcome is.
Lex Fridman
(02:15:52)
Yes, that’s a-
Derek Thompson
(02:15:55)
Productivity and service of what?
Lex Fridman
(02:15:55)
In service of what?
Derek Thompson
(02:15:56)
Efficiency in service of what? And I think the juxtaposition of the private and public sectors is really useful here. Private companies, publicly traded companies have the benefit of tight and quantifiable feedback loops. Are you making more money this quarter than you were last quarter or less? How’s consumer feedback? Do people say they like the product more? Are you getting 4.8 on your reviews or 4.6? The feedback loops are very quantifiable and they’re also very, very quick.

(02:16:22)
You know exactly when you’re doing a good and bad job. And so the KPIs are so cleanly drawn that people are aligned in their sense of we are moving forward and moving back. I don’t think government has done this at all under DOGE. In fact, I think in some ways goals that you could clearly identify as being good are being torn down. And this is why again, it’s so useful to be able to articulate goals in politics because without an articulation of goals, you don’t know whether the job to be done is to take something away or to add something.

(02:16:53)
And right now the answer from DOGE is to take away, take away, subtract everything, break down to studs as Cesar said. But let’s take a really, really concrete example here. The FDA, right? DOGE just laid off dozens of probationary employees. So upwardly mobile, recently hired, often very young, nicely credentialed employees at the FDA as a part of its slashing of the federal workforce.

(02:17:16)
One of the big problems of the FDA is slow approval of phase three clinical trial drugs. Everyone in this country who believes in science and technology wants the most lifesaving drugs to be brought into the public marketplace as soon as possible and competed against with other drugs, the price comes down and helps to extend people’s lives and health spans. I think everyone agrees that is an outcome worth fighting for. And if they don’t, certainly, as you and I willing to say that’s an outcome that we want. That’s what we want from our science policy.

(02:17:44)
What happens if you cut probationary employees at the FDA? The FDA doesn’t become more efficient. It becomes less efficient because the same amount of work spread over fewer people means longer delays in terms of approving phase three clinical trial drugs and deciding whether or not to approve them for public consumption. So in this really, really clear and very specific example, I think we can see the problem with not having articulated goals. You don’t know whether the job to be done is to take away employees or to add them. I think if instead what DOGE had done is come in and say, “You know what, Elon Musk and a bunch of other people from Silicon Valley, one thing we take very seriously is the importance of scientific and technological progress.”

(02:18:28)
Because if you look back over decades and centuries, what distinguishes our generation from every other generation in terms of its health and its power is science and technology. And we want to infuse government with a sense of science and technological progress. And to that end, one thing we want to do is to have a smarter and more efficient FDA so that people can experience life-saving medicines.

(02:18:49)
I think what you would do is research the bottlenecks that exist to American science and pharmaceutical policy and say, “We should hire more people at the FDA to accelerate drug approval, decide which drugs are to be rejected and which drugs are to get the FDA label.” That is the opposite of the DOGE approach, and this really, I think puts a fine point to the problem of a DOGE without goals.

(02:19:16)
When Elon takes over Tesla, when Elon is at SpaceX, when Elon is at X, I would imagine, and you know this better than me because you know him, and maybe most importantly for the purposes of this part of the conversation, the people who work for him. I’ll bet if you ask the people who work under Elon at X, Tesla, SpaceX, they say, “I know exactly what Elon wants. This is his goal for the super heavy rocket. This is his goal in terms of humanoid robots. This is his goal in terms of profitability of Twitter and the growth of our subscription business and how we’re going to integrate new features.”

(02:19:48)
There’s probably a really clear mind meld. Right now I have no sense that there’s a mind meld and in fact I have the exact opposite sense that rather than an example of creative destruction, which would be a mitzvah of entrepreneurship, we have an act of destruction, destruction. We have destruction for the sake of destruction. It’s much cleaner to me from an interpretive standpoint to describe DOGE as an ideological purge of progressivism performing an act or performing the job of efficiency rather than a department of actual efficiency itself.
Ezra Klein
(02:20:24)
I want to say because I really want to emphasize that it has goals. The goals are clear on some level and they have to do with centralizing power. So let me take out something that’s not DOGE because this is, I think an important place where you see what the effort is. What is Congress for? Not just Democrats in Congress who are in theory right now the opposition party, but Republicans in Congress. Congress is this aggregation of information from different places in the country who have chosen representatives to represent them in Washington. That’s how the system works.

(02:20:56)
One of the things that has really cowed congressional Republicans, it is a huge sun-like gravitational force now on Capitol Hill, is that Elon has made it known that any house or Senate Republican who defies Trump on a key vote, cabinet nominees or the CR, that kind of thing, he will dump 50 to a hundred million dollars into a primary.

(02:21:20)
It’s no money at all for him. It’s lethal for them. This is well known. He has said this personally to some of them. It’s been well reported. This is probably why Joni Ernst voted for Pete Hegseth. Well, what’s achieved by this? I think it’s an interesting question because Republicans in theory are allied to Donald Trump, right? He’s the nominee of their party and they don’t all have literally the same view of him. But I think you might say from one perspective there is a value in the system and there being checks and balances.

(02:21:47)
There’s a value in the system, in this system having to absorb other kinds of information. Now, we already don’t know the checks and balances we once had or thought we would have in this country because we have nationalized political parties instead of branches of government that compete with ambition, checking ambition, and I’ve done whole shows on this and we can talk about that if we want. But we do have political parties and political parties are themselves institutions that aggregate different kinds of information. In order to try to come to some outcome that is a better outcome because more information is surfaced.

(02:22:19)
I think Donald Trump would be in a stronger position for him if Senate Republicans could have done what they actually wanted to do and not confirm RFK Jr., not confirm Tulsi Gabbard, not confirm Pete Hegseth, not confirm Kash Patel and Dan Bongino, right? That’s actually a huge amount of risk the Trump administration has taken on. If they had named a sort of normal figure to HHS and then there’s a measles outbreak. Well, measles outbreaks are tough. That’s a hard thing. If you name RFK Jr. And there’s a huge measles outbreak, you’re really going to get blamed for that because people are ready to blame you.

(02:22:52)
If there is a domestic terror attack after you’ve put Kash Patel who’s quite unqualified in charge of the FBI and they’ve launched a war internally against the FBI which they see as a hotbed of anti-Trump sentiment and the FBI in the internal chaos misses some things and you have deaths on American soil, that’s a huge amount of risk you’ve taken on. I mean, the guy they fired, Chris Wray, he was Trump’s appointee. It wasn’t some Democrat. Trump named him in his first term.

(02:23:18)
But Elon, what he did here was he created a death star of primary money, and he has said that, “If you cross Trump in Congress even as a Republican, you’re done. Between Trump’s control of attention and loyalty and my ability to outspend you, you’re toast.” I think what that reveals is that what he wants is for power to be centralized under Trump. We’ve been talking about the bureaucracy, but he also wants it to be true in Congress. I’m not a Donald Trump fan, but obviously other people are Donald Trump fans. Obviously, I think at this point Elon Musk is a Donald Trump fan, but much of the country thinks this guy is great and I think we should take what they’re doing at word and deed.

(02:23:58)
The point of Donald Trump is that Donald Trump is right about things. We should give him power and he should use that power. My sense is you have some mixed feelings about him, but not everybody does, and this very consistent application of authority across the people… JD Vance. The difference between Mike Pence and JD Vance is JD Vance said explicitly in the whole run up to the vice presidential sweepstakes that his view of what went wrong in the first administration is between the bureaucracy and the staff.

(02:24:25)
Too many people are trying to inhibit Donald Trump and that what he would do is tell Trump that he’s got to get rid of these generals and he’s got to get people who will do what Trump actually said. The view of Trump’s fans, the view of his allies is that the first term didn’t go well enough because they had too much opposition from Republicans in Congress who talked Trump into things they shouldn’t have talked him into and too much opposition from the civil service and even from Trump’s own staff.

(02:24:51)
I think a very simple heuristic of why these terms are so different is that the most important member of the Trump family, Trump aside in the first term was Jared Kushner and maybe Ivanka, and the most important member of the family in the second is Donald Trump Jr. right? Kushner brought in a bunch of inhibitors. He brought in mainstream figures like Gary Cohn and Kushner represented other parts of society that were mixed on Trump and they wanted him to do certain things and not others, and there was maybe a productive tension.

(02:25:19)
McConnell was majority leader. He was more powerful than Thune is. Paul Ryan was speaker. He’s more powerful than Mike Johnson is. In the second term, you have Donald Trump Jr. who brings in much more right-wing figures. They’re accelerators, not inhibitors. Accelerationists in many cases explicitly. You have Elon Musk who believes the likeliest problem is Trump doesn’t go far enough, fast enough. And you have a weak Republican Congress that is further cowed by Musk’s money.

(02:25:45)
It could be good or it could be bad, but the Curtis Yarvin take that we need a more monarch-like figure is clearly being tested out. The view, just as you said about software engineering things is that you often need a benevolent dictator. Now, I don’t think Trump is benevolent, other people do, but the view that what is being attempted here is something much more centralized in its power, I think is actually a shared view of what’s going on. It’s like a consensus reality we have, not like an argument over reality we’re having.
Lex Fridman
(02:26:12)
Do you think there’s some degree to where if you trust in the system of democracy, which I do, and there’s some people, and we’ll talk about that. One of the main criticisms or concerns of Donald Trump is he’s going to break democracy. But if you trust that democracy holds, isn’t this an interesting experiment of how when everybody’s aligned, aggressive cutting of regulation and the number of people is an experiment of like, “Okay, let’s see what this does to a really over bloated bureaucracy that’s become extremely inefficient.”

(02:26:48)
And by the way, so I have an optimism about it that matches the division of abundance in the book, that once you do the creative destruction, then you start to really be able to step in and have a clear vision of like, “Okay, housing. What are the policies to solve housing?” But I do think the first step that’s needed is the destruction.
Derek Thompson
(02:27:15)
As Ezra said, we’re sort of speed running a particular experiment here of what does executive power look like if we do away as much as possible with checks and balances. And I would submit that we’re already starting to get some feedback loops from the market. The stock market is not the economy, but it is a very clear voting mechanism. And what’s clear is that many institutional and retail investors think that the current economic regime is pushing us toward a recession that we don’t have to have, right?

(02:27:49)
Donald Trump has insulated himself from any feedback loop or any sense of criticism that his tariff policy might not be the best course of strategy for American industry or global relations. And as a result, what we have is I think fairly described as a purposeful chaos. I mean, what other term can you use if a tariff is being announced at 9:00 AM and then taken away at 3:30 PM, and then 10 days later announced at 9:37 and then renegotiated at 2:45? This is not a principled theory of the perfect tariff level on international trade.

(02:28:27)
This is, I think, much more parsimoniously explained by just an expression of Donald Trump’s personality. This is a New York real estate guy. He loves making big awesome pronouncements and then using those big awesome pronouncements in order to negotiate little one-on-one dealings where he can rest for himself a personal sense of power or money, or pride, right? Announcing these tariffs in a chaotic way, forces world leaders to get on the phone with him and say, “Donald, what can I give you to bring down the tariff?”

(02:28:59)
This is personality standing in for politics in a way that’s totally unmolested by anybody else’s sense of, “Hey, that’ll maybe chill it on the tariff policy. Hey, let’s maybe slow down on the deportations. I’m not so sure about this particular move over here.” Instead, you have an executive branch that’s just a full manifestation of Donald Trump’s mind. And I do think at the early returns, if you look at consumer sentiment, if you look at the stock market, if you look at the 10-year yield, you have a range of let’s called them aggregated economic information telling us that the economy, consumers, employers, investors, do not like what’s happening now, which is going to be a really interesting test case.

(02:29:40)
Donald Trump’s first four years in office, love him or hate him, were four rather successful years of economic growth. Low unemployment, steady growth, low inflation, pretty much every economic indicator in the green. We’re already in the red in many of the economic indicators that never even blinked yellow under his first administration. And I personally don’t think it’s a coincidence that you’re getting these red indicators at a time when Donald Trump is having an entirely different experience of being president where there is no Mnuchin to tell him, “Hey, maybe let’s cool it off in the tariffs.”

(02:30:13)
The one more thing I would add here is you said maybe Trump’s presidency is a kind of right-wing abundance, right? And I think that it’s a worthy question, right? Is Donald Trump just doing his own version of abundance and we should… even if we disagree with his process, ideologically root for the outcomes that are likely under it? Here’s why I don’t think so. Let’s say that you or just someone as a conservative shares our view that they want housing to be abundant.

(02:30:45)
I think what they should really root for is to reduce the tax for building and make it easier to add housing units cheaply. Well, houses are made out of materials. Two of the most important materials in home building are soft lumber and drywall. Soft lumber, we import from Canada. Drywall, one of the key ingredients we import from Mexico. One of the first things that’s going to happen if you raise 25% tariff on lumber from Canada and drywall from Mexico is that the cost of housing is going to go straight up.

(02:31:15)
And this isn’t my personal opinion, this was a March 7th memo sent by the National Association of Home Builders essentially in a controlled panic saying, “Please don’t do a tariff policy like this. You’re going to screw over home builders even though ironically, you Donald Trump were elected by Biden to Trump voters who were mad about the price of housing.” So I am not in the moment optimistic that his centralizing style is going to be economically useful for Americans, whatever their interests.

(02:31:45)
My sense of the early feedback and of the early returns is that he in fact does not have a very clear and beneficial economic agenda. He has a personal agenda. He likes taking phone calls from international leaders and working out little deals with them. I don’t think that’s in the larger interest of economic growth, and certainly I don’t think it’s in the specific interest of reducing housing prices.
Lex Fridman
(02:32:09)
Okay. There’s a lot to say there. So first of all, can we separate the abundance and those efforts from tariffs? Because I don’t know who agrees with the tariffs. Tariffs don’t make sense to me economically. Maybe you can explain who agrees or likes the tariffs on the right or the left or anybody.
Derek Thompson
(02:32:30)
But do you see my point that-
Lex Fridman
(02:32:31)
Well, that point also to comment on it, I mean, one of the mechanisms by which bureaucracy forms is a kind of polite civility and a structure and a process. There is a argument to be made, and I’m not saying Donald Trump is that person, but there’s some qualities there of picking up the phone and calling Putin. That goes against all the process. Some of the most successful peace negotiations throughout history broke process. I just spoke with Narendra Modi. There’s a process you’re not supposed to meet with Pakistan or whatever for India. You just screw it. “I’m going to go to a wedding, a Pakistan wedding of a high up official. I’m going to do these things that are very Trumpian, break the rules.”

(02:33:17)
Now, in a perfect world, some of that is good matched with principled policy that’s surrounded by a large number of experts that actually understand that policy. I can criticize Trump all day, but I’m just saying that there is some degree to the picking up the phone and talking to leaders and playing in the morning, say one thing in the evening another and that could be part of a principled chaos.
Ezra Klein
(02:33:43)
I think the important part of madman theory is that you’re not actually a madman. You just got to convince people you are.
Lex Fridman
(02:33:48)
Yes, yes.
Ezra Klein
(02:33:49)
Look, I don’t think… I’m in agreement that we should talk to everybody. I don’t know how many people followed the 2008 election closely who are watching this. There’s a big fight in that election between Clinton and Obama about should you negotiate with your enemies? And Obama’s view is, “We should. We should talk to anybody.” And Clinton’s view was more nuanced than that, right? Certainly, we shouldn’t at this juncture. And during his presidency, Obama did a deal with Iran on the nuclear question. He negotiated with Cuba. He had very direct negotiations with Russia.

(02:34:23)
He did it importantly unlike Donald Trump without alienating all of our traditional allies. One of the things that I think is important to say about Trump is that the difference between Trump and, say, Biden or Trump and Obama is not that Trump will negotiate with Putin and Obama wouldn’t. It’s that Trump is realigning our alliances.

(02:34:45)
He doesn’t really seem to want to negotiate with the Europeans, or at least he wants to do it from a more hostile position. He wants to make Canada the 51st state, not treated as a longtime ally. The thing here is not that Trump is negotiating with our perceived enemies. These are his perceived allies and he’s turning our traditional allies into perceived enemies. So with him, I think it’s important. I am very much on the view of you talk to everybody. And it was my view going back for some years that it was clear that the Biden administration needed to be pushing for negotiations over Ukraine.

(02:35:18)
There was not going to be some end game here where Ukraine got all of its territory back. But, but, but, but I don’t think it is reasonable to look at what he is doing and say that is the norm that he has broken. The idea that we should have negotiations. I mean, many different presidents have done many surprising things and he’s rolled back a bunch of those things. Republicans have been very unfriendly to the opening that began with Cuba, right? And that was a big deal in American foreign policy. Something Obama did against very heavy criticism, something that cost him and cost Democrats in Florida and among Cuban voters in the following election.

(02:35:56)
So I just don’t think that that’s that an unusual thing with Trump. I also just want to say with the tariffs, he wanted to cut the tariffs off from DOGE and cut the tariffs off from this broader agenda. And one reason I don’t is that I think the way to understand tariffs… Look, you can be accomplishing many different things with tariffs. One thing, which was a theory that you heard from some people around Trump and fit what he said on the campaign trail, which was Trump’s position on the campaign trail, was that he was going to lay down 10 to 20% tariffs on all imported goods and 65% tariffs on imported goods from China. So I think that’s a bad idea for a bunch of different reasons, but what that is, is a stable change in the cost structure for all corporations and all trade. And then different players can make different investment decisions in the long term based on this new cost structure. What he’s doing, as Derek said, and I’ve had these funny experiences where I’m literally doing a podcast with a tariffs expert and I’ll be talking about the auto parts problem, and then my producer will be like, “He just delayed the tariff on the auto parts.” What he’s doing with these tariffs that you move on and you move off and you negotiate over endlessly, and I’ve been told this again by people around him is their view is that America had leverage it wasn’t using. And tariffs in particular are a form of executive control. Trade deals need to be negotiated and ratified with Congress. Trade deals are something you have to do with the rest of the system, but tariffs are something Trump can do unilaterally. And one thing Trump wants is control. That is the through line of virtually everything in his politics. Tariffs are a leverage based form of foreign policy. They are America has the biggest economy in the world that anybody who’s going to get into a trade war with us is going to suffer worse than we will. And there’s something where the president can use the tariffs because of authority granted by Congress some time ago for other purposes. He can use them with a huge amount of discretion. So you can use them on the one hand to say, “Well, I think that we’ve lost too much manufacturing and the dollar is undervalued in other places, or overvalued rather. And so we want to use tariffs to make the cost structure differently so that people have to locate more of the supply chains and the intermediate manufacturing in the United States.”

(02:38:06)
But you can also use a tariff to say, Columbia has to take the people we’re deporting and they have to take them in chains. And Donald Trump is doing both things. And the reason I think it’s important to keep those in mind is that what Donald Trump wants, what connects a lot of different things is that he wants leverage over things. I think it connects to Eric Adams and you were sitting here talking to New York City. Eric Adams is a Democratic mayor. Donald Trump has no truck with Eric Adams, right? He did not support Eric Adams. Eric Adams is not his natural ally, but what he had over Eric Adams, he realized was a leverage that he could get the cases off of Eric Adams and then get Eric Adams in his pocket.

(02:38:43)
So Donald Trump stepped in to save like the corrupt Democratic mayor of New York City. Not because they have a deep ideological alliance, but because then Donald Trump would’ve power over New York City and its policy that he wouldn’t otherwise have. The thing that connects Trump, Trump is a very old school politician. He’s very relational. He’s very 19th century. He’s looking for the angle. He’s very zero sum. He’s looking for things to empower.

(02:39:08)
DOGE is a way of getting power over the bureaucracy. Tariffs are a way of getting power over the international financial system and foreign policy and breaking traditional alliances. Maybe in his mind, even getting Canada to become the 51st state, maybe getting Mexico to change its immigration policy, trying to weaponize the Justice Department against Eric Adams as a way of getting power over Democratic mayor. Some of what they’re doing with grants and money is a way of getting power over other institutions in American life.

(02:39:34)
Again, you could think it’s bad or you could think it’s good, but it is coherent. I think it’s bad, but it is coherent, and the people who think it’s good think it’s good for Trump to have power. Again, with the Eric Adams thing, they were very explicit about this, right? They said that it is worthwhile for the president to negotiate over his policy objectives and to trade things to get his policy objectives more fully carried out. And so Eric Adams went from saying New York City would be a sanctuary city. They’d aggressively carry out deportations.

(02:40:02)
You can think that kind of transactionalism is fine. I think in this case, given that it’s about corruption, it wasn’t. But the idea, Trump’s idea that the problem is that the president doesn’t wield enough power here the way he does in Russia, the way he might in India, the way he might in China. Trump has spoken very openly of envying some of the powers these other people have. I think it’s a consistent governing philosophy that needs to be taken as that. Right?

(02:40:28)
I think sometimes you describe… One of the difficulties Trump poses is sometimes if you just describe what he’s doing, I think quite neutrally, people say, “Oh, you’re criticizing him, you’re anti-Trump.” But I’m just describing what he’s doing. You can think it’s good or it’s bad, but the fact that he wants to arrogate power and find points of leverage and tariffs are one. What he’s doing through DOGE is another. What he did with the DOJ is another. It’s all very consistent. The question is really then just what your normative view is on Trump having that much power.
Lex Fridman
(02:41:01)
Yeah. I have a question about DOGE, but before that, I’ll just say that I don’t think Donald Trump should have that much power because I think to this day, my biggest criticism and concern is that he is a person that denied the results of the election. And so I’m unwilling. I like George Washington. I like people that have the skill, the ability, the track record to walk away from power. And a person who’s unwilling to accept reality and is willing to bend reality to maintain a grip on power, even if what they’re trying to do is really good, maybe making a government more efficient makes me very concerned.

(02:41:49)
But on the topic of DOGE is more on the Elon side. If we can combine DOGE and abundance as a topic, if DOGE succeeds the effort, let’s just forget DOGE, but the effort of making government more efficient, what does that look like before the next election and after the next election? If we can just look at success.
Ezra Klein
(02:42:12)
I know I’m jumping again, but I’ll do something quick and then pass to you. I’m not a huge fan of the experiment framing because amidst this experiment, people are… Through USAID, as my colleague Nick Kristof has documented and potentially further and otherwise will die, lose homes, be scammed by things. So I feel like experimenting with the government at that level is very dangerous, but if assuming democracy survives, Democrats, they’re going to make the government work again, are going to have to become less rule-bound than they were.

(02:42:47)
I have outlined that the problem with a personality type of the right is it’s autocratic now, and the problem with a personality type of the left is it’s bureaucratic. And I don’t want to see things going either to where DOGE is in terms of, I think, there’s a fundamental lawlessness to it, but I also don’t like the aims of it. But I don’t want to see things where Democrats were, which is such respect for process, that they will put the process ahead of getting their own things done.

(02:43:14)
They will listen to any lawyer with any super intense interpretation of any law, and that’ll turn them all the way back. What DOGE has demonstrated is that the room for movement inside the American government, inside the state, as much wider than anybody gave it credit for both Republican and Democratic administrations. It may not be as wide, it probably isn’t legally as wide as what DOGE is trying to do, and that’s why they’re losing all these court cases. But it is wide. And you could also use Congress. You could pass statute to make it wider, and you should.

(02:43:45)
There is no abundance agenda that works if the Democratic Party is as rule-bound and as process obsessed as it was in the last 10 years. It doesn’t work flatly. That is why so many of the stories we tell in the book are about process gone. It’s not process gone wrong because it’s doing exactly what it was intended to do, but it is process that has become antagonistic to the promised outcomes.

(02:44:10)
And so I do hope there’s a kind of thesis, right? All the institutions are great. Democrats are the defenders of government. Antithesis, the government is a corrupt cesspool. It’s a ball of worms, and it has to be destroyed and taken over by a benevolent or non-benevolent dictator. And synthesis, which is that the government’s process has made it not a functional state. It’s non-responsive in many important ways, and it needs to be reformed at a quite fundamental level, but in a way that is lawful, in a way that is thoughtful, in a way that is running experiments that gathers information, and then can make adjustments, in a way that is respectful of the human lives, that it touches and affects, in a way that is not hostile to the goals of good government, but is more committed to them than it is committed to the process…
Ezra Klein
(02:45:00)
…but is more committed to them than it is committed to the process the government has erected and evolved over time.
Derek Thompson
(02:45:07)
As you articulated the principle beautifully, and just to put some meat on the bones, abundance is about being incredibly concrete about what you want to accomplish in the world. Then it’s about understanding how to accomplish that. A stage of the process that I think Doge has entirely skipped. I don’t think we’ve reached the stage of understanding. We’ve destroyed and then maybe it’s sometimes promised to understand the thing that we’ve destroyed after the fact. You want to set goals, you want to understand how to meet those goals and then you want to meet them. And the problem with liberalism over the last few years, in the last few decades is that we’ve become disconnected from outcomes. Really, really crystal clear example. So Joe Biden in 2021 signs the bipartisan infrastructure law. And he and Pete Buttigieg call it truthfully, one of the most important infrastructure bills passed in the last few decades. $1.2 trillion to do exactly what Ezra and I want to do, to build in the world.

(02:46:04)
There’s a piece of that law that’s a $42 billion program called BEAD stands for broadband, equity, access and deployment, I think. $42 billion to build rural broadband. Fast-forward to 2024, practically none of it is built. We’re now four calendar years after it was passed and the program is at this point, it seems like it’s going to die and they’re just going to transfer the whole thing to Starlink. So why? We want to understand why does government fail to achieve its outcomes? And how can we learn from those failures to allow government to succeed at its outcomes? Well, you look into it and it turns out that in order to take these $42 billion and send it to the states, the states had to go through a 14 stage process in order to get the money.

(02:46:47)
First, the FCC had to draw a map of the places where America needed more rural broadband. And then there was a challenge period where people could question the map, and the FCC would remake the map and then the challengers would sue them to change the map again. Then the states had to file a letter of intent and a five-year action plan and a funding program, all of which could be subject to their own challenge periods. And in each of these challenge periods, the commerce department is going back to the states and saying, “We really like your five-year plan, but your workforce development program didn’t pass this matrix of equity. And you didn’t reach out to the right bidders over here. And if you try harder to reach out to more people who just aren’t white men to be your employees, that would be fantastic if you could put that into the new edit that you submit to us.”

(02:47:29)
And of course there are delays because every state has to do its own programs. The commerce department is backed up, yada, yada, yada. You get to a point where out of 56 states and jurisdictions that have applied to begin the 14 stage process, by the time the Democrats lose the election in November 2024, three out of the 56 have passed all 14 stages. And very little of the money has actually been spent because of all the problems that Ezra and I discussed about how hard it is to build in the physical world. $42 billion therefore dies upon contact with planet Earth. That’s not government achieving its goals. And a Doge that we were sort of duumvirate of in this parallel universe, is one that would try very clearly to A, articulate a goal. What are we trying to do here? We’re trying to build rural broadband. Why? Because we think connectivity is incredibly important to the economy of the future. It helps people’s health, it helps the economy of rural areas. Let’s build rural broadband.

(02:48:28)
Two, what are the roadblocks? What are the bottlenecks? What’s hard about taking a pot of money that exists in Washington and actually creating broadband networks in rural Kentucky? Let’s understand what those roadblocks are so that we can do two things. We can take away the things that need to be taken away to accelerate the program. And maybe we can add new policies that will accelerate the spending because bottom line, we want to make a difference in the world. That’s a world where government is, to borrow Ezra’s language, deregulated itself. It’s easier for the government to achieve its goals. I think that it’s really important at the level of a principle here that liberals fall out of love with this procedural fetish that has dominated the left of the last half century and fall back in love with outcomes. To be ruthlessly obsessed with how liberalism has failed and how these kinds of failures aren’t just technocratic stories to tell in a podcast.

(02:49:26)
I think this is fundamental to why Democrats are losing the communications war in an era of anti-establishment and anti-institutional. We find ourselves in a reflexive position of having all the cranks, so to speak, having left the Democratic Party and we’re the ones who defend all institutions. We’re the ones who defend the establishments. We’re the ones saying government can only do good. But as a result, we lose the ability to talk to people about how government fails and how they can see that failure. And how sometimes they’re literally leaving cities and states run by Democrats because that failure is so effing obvious to them. So this isn’t just about the BEAD program, it’s not just about 14-step programs, it’s not just about rural broadband. It really is about a higher level principle of political communication. How do you develop a liberalism that in an age of anti-establishment anger both reflects that anger and channels it for proactive purposes by not just doing destruction destruction but creative destruction.
Lex Fridman
(02:50:25)
So first of all, beautifully put and second is of the big thing that Doge did is make this a sexy topic to discuss. And then you can tear down Doge with the way they’re doing it, say it’s wrong, criticize. But then people are all of a sudden more and more caring about the efficiency of government and educating themselves, learning about it, and it’s creating a culture of transparency to the whole thing. So now you can swoop in with a book like Abundance and describe here’s actually how to have a clear mission and metrics, how to solve these problems. But that allows as opposed to have this culture of process, me personally, one of the things that really frustrates me about this world is that, the bureaucracy of that kind of process. Especially because everybody, at least in the United States is just all so fucking polite everywhere about the whole thing. They’re all so nice to you as they’re doing the process. About every single thing they’re just like, this is… And then you have to call from nine to five.

(02:51:28)
There’s hours and let’s schedule a meeting in three weeks from now to discuss this document so we can have another document. And all of a sudden the big dreams and the visions, the hopes that people have invested in building a project that’s an incredible project dies. It’s not just a waste of money, it’s the possibility of a beautiful thing that could have been built, never gets built.
Derek Thompson
(02:51:49)
You’re retracing Ezra’s lovely line about how the character of the right these days is autocratic, and the character of the left can be overly bureaucratic. I hope there’s a middle synthesis lane here where there’s a political identity that believes in efficient bureaucracies. Sometimes it actually does take a lot of people to get certain things done. But you really, I think need to have an eye toward institutional reform. This is a theme of our book that we haven’t talked about as much yet, but I think it’s so important for abundance to believe that each generation adopts, is passed down institutions that were created for different eras, different decades to solve different problems. That’s how you get an environmental revolution in the 1960s that solves the problem of dirty air and dirty water. But leaves us with a set of norms like NEPA that make it impossible to add clean energy in our generation, just a tragedy of unintended outcomes.

(02:52:43)
And to narrow this down to a world that I know you care a lot about, there’s a chapter in our book about science policy and the history of the NIH. And the NIH really comes into its own after World War II. And it is immediately this beautifully funded crown jewel of biomedical research, just the federal government irrigating university researchers studying cancer and heart disease and brain science and everything else. Practically every single scientific breakthrough in the US in the last 70 years at least bears the fingerprints of NIH. But the NIH is also a bureaucracy. And like every bureaucracy, it has accumulated a set of processes and habits that it’s the people most affected by it. Scientists in America told us and will tell anybody else who’s listening, has flaws.

(02:53:31)
According to some surveys, 40% of the time that scientists are working today in America, they are filling out grants and doing paperwork, not doing science. That is astonishing. As we say in the book, imagine if we discovered that one year there was a virus that broke out in the American academic scene, such that our scientists suffered from chronic fatigue disorder between January and June every single year. They just couldn’t work. We’d be like, “This is an absolute shanda, we have to fix this problem.” This problem exists. It’s our own bureaucratic rules. It’s the rules that we wrote that’s slowing down science. So what should we do? I don’t think we should tear down the NIH. I don’t think we should slash and burn grants as we’re currently doing in the administration.

(02:54:17)
I think we should understand what’s broken and be clear about our desired outcomes. My desired outcome for science is that we produce and pursue high risk, high reward science under the theory that almost no important breakthrough is going to be obvious before you discovered it. If it’s obvious, you probably already knew it. So why is it some new science? You should want to incentivize scientists to ask their biggest, most curious questions in a high risk, high reward environment. And right now the NIH doesn’t do that. The NIH for a variety of reasons, whether it’s aspects of the peer review process or just assumptions that scientists have of the system, is too incrementalist. It funds older researchers and it wastes a lot of scientists time with bureaucratic paperwork and kludge. We’d love to reform it. We have ideas for reforming it. But if liberals don’t use the language of and act on institutional reform, we will allow institutions to grow old and sclerotic and piss off Americans.

(02:55:16)
And they’ll vote for people who come in with a wrecking ball to tear it down. So that’s why I think it’s so important for abundance liberals like us to be very clear about our goals, our outcomes, and exactly what we want to accomplish because I think that’s the only way to really see how and why institutions that are all around us truly do need reform.

Sam Harris

Lex Fridman
(02:55:37)
Listen, I hope your book becomes the manifesto of the Democratic Party. It’s a beautiful vision. I’m a little bit skeptical because of the momentum of bureaucratic thought, but nevertheless remain hopeful. I have to ask as I’m a fan of both of you in the interest of time, Ezra, you had an intense debate many years ago with Sam Harris. Okay, so you’re, I would say, like I said, I’m a fan of both of you. You’re both intellectually rigorous people. So the debate, the contentiousness of it was both sad to me. But also as just a fan of you and since I admire your intellects, it’s just fun to watch, what is it Godzilla and King Kong fight. I wish there was more of it. I wish you would do this podcast again and debate it more on some other topic and argue. It’s just great. Anyway, some time has passed and-
Ezra Klein
(02:56:40)
Yeah, was this eight years ago now?
Lex Fridman
(02:56:42)
Yeah,
Ezra Klein
(02:56:42)
2018-ish, something there.
Lex Fridman
(02:56:44)
Yeah, the battles you have fought over the years, that’s just-
Ezra Klein
(02:56:47)
That’s a lot of chapters.
Lex Fridman
(02:56:49)
Yeah, this is like several chapters ago. But in the interest of camaraderie, what do you admire most about Sam Harris?
Ezra Klein
(02:57:00)
Oh, that’s not a hard question. If you go back, if you listen to that debate on my show at Vox, I intro that debate by saying, “Look, I disagree with Harris on like this specific conversation he had with Charles Murray about race and IQ. But he’s good on meditation, he’s good on psychedelics, he’s good on consciousness, he’s good…” I don’t know if I said AI back then though, but I think he’s good on AI. I always felt in that without going into the way back machine, that Sam really thought like he really somehow got where I was wrong. And I thought I offered like a lot of like tries for deescalation and he did me weirdly the favor of publishing our whole email correspondence. And I think if you read that you can see that I was not angling for a fight here.

(02:57:42)
What I admired about him since, I think the thing that he’s been good on is he’s been very independent. So Sam at that time, there was sort of the emergence around then of this thing that people then called the intellectual dark web. And it was like Sam and the Weinstein brothers and Ben Shapiro, and I forget who was part of it. And as a bunch of those people I think, it kind of split up over time, but Harris has done a good job not falling into conspiracy as the Weinstein brothers did. Not letting his anger at the left blind him to the failures of the right. He’s a guy, I guess both for better and for worse as we all are. But he is perfectly willing to stand alone in a crowd. Like I haven’t listened to that much of his stuff lately, but my sense is he’s been like quite clear-eyed from my perspective on Donald Trump. So my view is not that Sam Harris is in general a bad actor, it just isn’t.
Lex Fridman
(02:58:36)
Yeah, he’s been like difficult to categorize and fearless about it, meaning like he deliberately resists audience capture.
Ezra Klein
(02:58:49)
Yeah, which is a hard thing to do.
Lex Fridman
(02:58:50)
It’s quite difficult.
Ezra Klein
(02:58:50)
I think I saw a clip of him on this show sort of like going after Trump and it became like a big whole thing.
Lex Fridman
(02:58:55)
Yeah, he’s been very consistent on that. Let’s try to find the interesting thing here. I actually had a while back, a podcast with Richard Haier. He studies intelligence. And it was a very detailed non-policy discussion about IQ tests and all that kind of stuff. And there I tried extremely hard to be very nuanced because that felt like an uncomfortable topic back then. It doesn’t feel so uncomfortable now. Do you think the Overton Window has expanded? Do you think the kind of things we’re willing to talk about now… And this has to do with the Trump moment also.
Ezra Klein
(02:59:35)
I think the place where the topic seems to me to have changed in my ambient awareness of it is two things. One is the sense of the Flynn effect, which is yeah, it’s hard to individually change your IQ, but there is a very well-documented effect. And I talked to James Flynn before I talked to Sam. There is a very well-documented effect where IQs have been rising over time. There’s good evidence now that that has stopped. Again, this is ambiently my sense, I’ve read some things like in passing, I did not come to this podcast preparing to talk about IQ or I would have prepared very carefully. And then there was an interesting FT article just the other day that was like, Have We Passed Peak Human Intelligence? And my sense is that that wasn’t about IQ, although again, I don’t remember exactly what was in it. I kind of glanced at it and put it aside to read more later.

(03:00:24)
But the literacy scores are going down and a number of test scores are going down. And the sort of sense of the piece was that we are making ourselves stupider by endlessly staring at screens and social media, which I think is probably right. I’ve had this sort of line since becoming aware of that piece in my head that, we spend a lot of time talking about how to get smarter and not enough time talking about how to avoid getting dumber. But we do a lot of things that we think make us smarter, like I would say having social media on our phones because like, oh, you’re getting all this information all the time that in fact make us dumber because it’s not just that the information is bad. But the lack of concentration, the constant distraction, the sort of lack of focus.

(03:01:05)
And so my sense is that there is a different kind of conversation here. I’m not sure it’s really an IQ conversation, but that there is a sense that we are inflicting a possibly global, certainly societal cognitive wound on ourselves. That’s what John Height’s book is a little bit about, that’s sort of more coming at it from a behavioral standpoint. But we have the Flynn effect reflects… This is how James Flynn would describe it. The Flynn effect reflects society’s putting on, he, I think put it the scientific spectacles. You create societies where we prize things like reading abstract intelligence, symbolic logic. You teach people on them a lot and we get better and better and better at doing it. And reading changes the brain, it changes the physical structure of the brain. You’re hijacking parts of the mind and meant for other purposes to do this kind of interpretive work.

(03:01:59)
But it’s of course possible just as we became societally connected and then made more and more widespread technologies like literacy, that changed our brains and led to the increase in this thing we call G intelligence. And decreases in other things, like I’m really shitty at knowing which direction I’m going into. That sort of part of my brain that does navigation is terribly atrophied compared to somebody in a society that did not have Google Maps. It’s entirely possible to go too far in that. It’s entirely possible to move on to technology to begin to weaken that. I have this concern very profoundly about AI, by the way, which obviously you have a lot of your roots in. And there are parts of AI I’m very optimistic about, but my biggest concern about AI is we’re going to make ourselves much stupider without realizing it. Because the things that are easy to automate in AI, which is like getting the AI to summarize the reading of something or getting AI to write the first draft of something, that’s where all the intelligence happens in my view.

(03:02:59)
I think one advantage I have over a lot of other podcast hosts who superficially do what I do, probably this is true for you, I’m sure this is true for you, I know, is that I really do do the reading. No summary is equivalent to me doing the reading and sitting there and making the associations myself and spending the time in the book. And then sort of thinking about what it brought up in me. I write the first draft. ChatGPT cannot write the first draft of my book. The first draft is fucking hard to write. And it’s often hard because it’s completely wrong, but not because it’s narrowly wrong. ChatGPT will never tell you that the problem with what you’re trying to do is that you’re just trying to do the wrong thing. It’ll never tell you if you tell it to write a first draft, that’s the wrong direction for this draft. Some part of you has to know it. And often the problem is you just haven’t done enough reporting, haven’t done enough research and ChatGPT can’t tell you that either.

(03:03:47)
And my worry for my kids, my worry for society is creating technologies that make it incredibly alluring to automate the part of creation that is most difficult, most laborious, and most likely to lead to genuine insight and the sort of sharpening of your own mental acuity. We better hope the AI can autonomously make innovations because I really worry we’re going to stop being able to.
Lex Fridman
(03:04:14)
Yeah, and all of that is brilliantly put, I definitely think that social media is making me dumber. Like if I spend a week checking social media versus reading books, I’m just distinctly the quality of my thoughts-
Ezra Klein
(03:04:31)
Even the same content. I don’t read things anymore as much as I can on like a screen, I print everything out and I sit at a table and I read it. I can read it on my iPad, I can read it on my laptop. I print it all out because my attention is different.
Lex Fridman
(03:04:44)
And the same goes for AI, whether we’re talking about this kind of research, it’s more distinct and rigorous. And the other space that I do every single day is programming. I’m definitely becoming a worse programmer by using AI offloading because it actually works really well there. I’m becoming worse at creative thinking at what you’re saying, writing the first draft, which require that skill, that first little leap, that little mini leap into the creative genius that we all do every single day. AI is not able to do that, and it’s definitely doling that. All right, we covered a lot of ground today. But on the note of optimism, what gives you hope about the future of this great nation of ours, the future of America? Looking out in the next few years and the next few decades, centuries, when we colonize the solar system and beyond or we could just stick to the next couple of decades.

Future of America

Derek Thompson
(03:05:49)
Sure. Despite the fact that our book is a deep diagnosis of modern liberalism with a ton of criticism of the last half decade in politics, I am at root a profound optimist about everything just at a general personality sense. To a certain extent, a question like this is attempting to elicit a little bit of what can be considered analysis of the world. But you’re also eliciting what is fundamentally like personality, like what makes you optimistic? I’ve just always been a real optimist. I’m optimistic about science and technology, especially in the realm of biomedical science in a big way. I think if you look at what’s happening right now in mRNA cancer vaccines, in CAR T-cell therapy for redesigning T-cells to attack cancers. If you look at the GLP-1 drug revolution and some of these studies that have been done on the fact that GLP-1 drugs, while they were initially synthesized from lizard venom to help people with type two diabetes turned out to have these effects that seem to reduce body-wide inflammation. That not only rewires our minds and makes it easier for people’s desired sense of moderation to be actualized.

(03:07:04)
So people who want to eat more fruits and vegetables seem to find it easier to eat more fruits and vegetables when they’re on GLP-1 drugs. But also because there’s probably a lot of neurological issues that are fundamentally issues of inflammation, including maybe dementia and Alzheimer’s. We might have accidentally from the tongue of a lizard, a partial medicine for Alzheimer’s disease. The ability of science to connect these dots in the cosmos just absolutely thrills and fascinates me. And I hope that we get better at making those connections.

(03:07:39)
And while I have a lot of fears about AI, many of them shared by Ezra, I am really interested in the possibility of AI being useful for synthesizing large bodies of knowledge to allow people to make cross-domain comparisons. I think a lot of inventions in tech history are essentially ingenious recombinations of ideas. Like to a certain extent, you look at something as fundamental or archetypal as Thomas Edison inventing the incandescent light bulb. What did he do? He just tested 10,000 different materials and figured out that actually it was like a very special kind of bamboo that burned for the right amount of time and said, “Boom, I did it. I made the incandescent light bulb.”

(03:08:18)
The ability to have a machine accelerate the degree to which we understand what those 10,000 materials can do, or synthesize knowledge so people can combine it and say, we’re going to take a little bit of CRISPR over here and a little bit of cancer science over here to develop a gene therapy that targets a particular inhibitor that allows the immune system to attack a protein that is explicitly related to pancreatic cancer. I do absolutely believe that we might be on the doorstep of those kinds of breakthroughs. And that makes me incredibly optimistic because I think at base it’s like, what’s life about? What’s abundance about? Why housing and energy? Well, because we think they’re fundamental to living a good life. You need a place to live, and people deserve the freedom to live where they want to live.

(03:09:06)
Energy is what powers the entire economy. And more energy would power technologies that we can’t even imagine or are just beginning to, whether it’s supersonic flight powered by clean fuel or fusion technology that basically gives us infinite solar energy, the sun’s actual energy in a particular location. These are beautiful things that can happen. But also I think the good life is about health. It’s about health and it’s about the wealth that comes, and the freedom that comes from finding health in your own life. And I am incredibly optimistic that we might be at the cusp of a real golden age in taking all these little ingredients that we’ve spent decades on, whether it’s genomics and proteomics and a little bit of AI. We’re at a moment right now, I hope, where we can have this explosion of combinatorial intelligence. And that hopefully in an optimistic way, AI could be useful in accelerating us toward that future. But I also think to the point of this book, we have to get institutions right too.

(03:10:11)
These are discoveries that are going to happen inside of institutions. And how those institutions work and how they’re funded and the incentives that are created by law or by technology, really matter in terms of the world that we build. And so that’s why I think it’s really important to not only be obsessed with what the technology can do, but how it’s instantiated in the institutions that we have because it ultimately is institutions and individuals that build the world, not technology acting on its own.
Lex Fridman
(03:10:36)
Yeah. And I should say that you make a really great case for investing in weird science, meaning stuff that doesn’t on the surface make sense. You said lizard venom. There’s all this popular criticism of scientific projects that sound like a waste of money when in reality, at least in the scientific realm, projects that seem like they don’t have any positive effect might actually end up being the ones that transform human civilization as we know it because of the unintended discoveries that happen and all the eureka moments. All the special discoveries happen truly, when you’re just passionately pursuing a cool thing in science, that’s how scientific minds work. And so it makes sense to invest in exploring weird shit, the weird mysteries of the universe. Anyway, Ezra, what gives you hope?
Ezra Klein
(03:11:30)
I’m a less temperamentally optimistic person than Derek by a lot actually. And I would usually give a pretty similar answer to this question that Derek gave, which is technological advancement. We live in an age of marvels. I guess I’ll say from the realist perspective, that we live in a more liquid moment than many. That if one of the advantages of the ’90s of much of the life that I have grown up in, is that much of the structure of technology, of global governance was fairly stable. We could sort of like look at it and there were certain, not certainties but fairly reliable guardrails of what was and what wasn’t going to happen. And there were disruptions like 9/11 and the financial crisis. They weren’t small and slowly they broke that entire system. But I think one of the things people sort of miss, sort of feel about the way history’s accelerated is that things have just all gotten faster and less predictable. And they really have, I think gotten faster and less predictable.

(03:12:28)
We’re in an age it seems to be more like the early 20th century, late 19th century than like the 1990s. And that means the possibilities range very widely. When you think about AI, when you think about biotech advances, when you think about energetic advances, when you think about the sort of shifting nature of global alliances, of the technological and political systems, we are in the most high variance period that I think has happened in a very, very, very long time. And that carries tremendous peril. Things could go terribly, and it carries tremendous possibility. What if AI is an incredible boon for humanity? What if we do invent and deploy the clean energy technologies that deliver energy abundance. Not just a kind of answer to the worst of climate change, but genuine energy` abundance.

(03:13:20)
Unlocking new things like mass desalination. And what if we do… Like I’m an animal suffering person. I’m a vegetarian. I care a lot about animal suffering. What if we did figure out cultivated meat that we grow in breweries and on scaffolds, and we don’t have to kill tens of billions of animals that we’ve raised in unimaginable suffering every single year? What if we do figure out how to make government better and more responsive? What if that thesis, antithesis, synthesis thing does work out? But it’s not hope, it’s like that’s a future. You have to create a future. You have to call into being a future you’ll have to fight for. So it isn’t so much that I find myself hopeful about what America or the world would be like in 2030 or 2040. But I find that I believe in the possibility of enough remarkable outcomes, that it makes the present really worth being engaged in. And really worth trying to do your small part, to bend the arc of the time in the direction that you find more just.
Lex Fridman
(03:14:25)
Well, thank you to both of you for fighting for abundance and writing this manifesto for abundance, and for all the writing and the work you do in the podcasts and just being incredible minds in this world that I’m a fan of. So thank you for talking today.
Ezra Klein
(03:14:45)
Thanks very much, Lex.
Derek Thompson
(03:14:46)
Thanks so much.
Lex Fridman
(03:14:47)
Thanks for listening to this conversation with Ezra Klein and Derek Thompson. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Napoleon Bonaparte, ” In politics, stupidity is not a handicap.” Thank you for listening. I hope to see you next time.

Transcript for ThePrimeagen: Programming, AI, ADHD, Productivity, Addiction, and God | Lex Fridman Podcast #461

This is a transcript of Lex Fridman Podcast #461 with ThePrimeagen.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Introduction

Lex Fridman
(00:00:00)
The following is a conversation with Michael Paulson, better known online as ThePrimeagen. He is a programmer who has entertained and inspired millions of people to have fun building stuff with software, whether you’re a newbie or a seasoned developer who has been battling it out in the software engineering trenches for decades. In short, ThePrimeagen is a legendary programmer and a great human being with an inspiring roller coaster of a life story. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description. And now, dear friends, here’s ThePrimeagen.

Love for programming

Lex Fridman
(00:00:42)
What do you love most about programming? What brings you joy when you program?
ThePrimeagen
(00:00:46)
I can tell you the first time that I ever felt love in programming, or felt that joy or that excitement-
Lex Fridman
(00:00:51)
Sure.
ThePrimeagen
(00:00:52)
Which was in college. It was the second class, data structures, and the teacher that was teaching Ray Babcock, he was talking about linked lists. Now you have to learn Java at Montana State University when I went, and so he’s off there explaining this whole linked list thing and all that, and then he shows code. And in the code it’s abstract class node or whatever it was, I can’t remember what it was. And then it had a private member, and that private member was of type node, and I’ve never seen that before. It is a class that is called node with a member that is of itself. And for the first time ever I was like, “Oh my gosh. There’s no end. There’s no way to iterate. This is not a set of 10 items. This is a set of infinite items.” And so my mind kind of exploded in that moment, like, “What you can express is huge. I can see what memory looks like. I can see this hopping through space.”

(00:01:47)
And I just remember being just so blown away, because up until that point, everything was just, “All right, I have a list of 10 items. I have a list of 20 items.” Right? It was very rigid and small, and the things I built were really small and trivial, and all of a sudden I felt like I could build anything in that one moment. And it was so amazing. I just remember sitting in class for, I don’t even remember how long those classes were or anything, but I just remember being just completely profoundly impacted by this notion. And so I just sat there and I watched, and I had the exact same experience in heaven’s forbid by a software engineering class, when we talked about the decorator pattern, where you can keep on constructing these objects in this recursive way. Not that I think that’s actually a good idea to do, but just watching that and realizing there’s so many weird and unique ways you can solve problems, and anything your mind can think of, you can just create that. And I just remember getting just so excited about the possibility that anything is possible.
Lex Fridman
(00:02:41)
Yeah, let’s wax philosophical about a linked list. It is pretty profound. For people who don’t know, a node in a linked list doesn’t know anything about the world it’s in. It only knows about the thing it’s linked to, its neighbor. Maybe that’s symbolic. It’s a metaphor for all of us humans. There’s billions of us on this planet and we only know about our local little network.
ThePrimeagen
(00:03:04)
Yeah.
Lex Fridman
(00:03:05)
And it’s kind of beautiful. And you realize in that little simple data structure, you can construct arbitrarily large systems, and they’re like roots that go through memory. And then of course, that’s where you get all the programming languages that allow you to dump junk into memory and have memory leaks, and therefore create infinite pain as you try to figure out where that unfreed memory is. For me, yeah, probably… It’s so beautiful the way you put that. Linked lists are indeed beautiful. Recursion also for me, when I finally wrapped my brain around what it means to write a recursive function.
ThePrimeagen
(00:03:49)
What was the thing? What was one that taught you? Because I think we all… You probably did factorial, where you just do a quick factorial of it. It just doesn’t hit home. What was the thing that made it hit home?
Lex Fridman
(00:04:01)
I don’t remember the first.
ThePrimeagen
(00:04:05)
I remember my first. How do you not remember your first? It was magic.
Lex Fridman
(00:04:08)
I’ve had so many that it just…
ThePrimeagen
(00:04:11)
I mean, you are a Lisp guy. You’re probably pretty used to the recursion.
Lex Fridman
(00:04:14)
Yeah, all I remember is just surrounded by sea of parentheses. I mean that’s really, probably, when I… In high school, I think it was either Java or C++. Wow, how do I not remember that? It must have been C++. And then college, the generic bullshit software engineering classes were Java, but then the renegades, the cool kids, were all using Lisp. That’s when you’re doing the AI, the quote-unquote “AI” at that time, that that was Lisp. If you want to write a chess engine, you would use Lisp. And so for me, probably the moment I really fell in love with programming was Lisp, and writing Othello programs and chess engines, all kinds of engines that play a game, and then I could play against that thing and that thing would beat me. The joy of being destroyed by the thing you’ve created. And oh, game of life too. Cellular automata. That’s when I…

(00:05:19)
I built that, you know, all kinds of programming languages. That’s less about programming languages and more about the system you create. And that just filled me with infinite joy, having… Now similar to the linked list situation, creating a system where each individual cell only knows about its neighbors and operates in very simple rules. But when you take that system as a whole and allow it to evolve over time, you can create infinite complexity. So I just… Man, those are many pothead moments, where I’m just looking at the beautiful complexity that can be created with cellular automata. That filled me with just infinite joy, for sure. But yeah, all I remember is parentheses. So my memories of my first are drowned in a sea of parentheses.
ThePrimeagen
(00:06:11)
Oh man. Well, first off, mine was in Java, so my first was a little bit more rigid, kind of, you know, a corporate experience.
Lex Fridman
(00:06:20)
Yeah. Cold, meaningless…
ThePrimeagen
(00:06:23)
But… yeah. I was in a lab, everyone was using CentOS at that… or Cent OS or however you say. I always called it CentOS, the fresh maker.
Lex Fridman
(00:06:29)
Yeah.
ThePrimeagen
(00:06:30)
And so it’s just like I’m in this very cold…
Lex Fridman
(00:06:32)
That’s nice.
ThePrimeagen
(00:06:33)
Thank you.
Lex Fridman
(00:06:33)
Yeah.
ThePrimeagen
(00:06:34)
I’m in this cold, rigid environment with my Microsoft keyboard, programming away in Java. And I still have just such… this memory of despair, because I love programming, this was after the linked list, and I cannot figure out recursion. And so I go to the university store and I buy a book and it’s Deitel and Deitel learn Java and it has a section, Recursion, so I open it up and I start reading it, and it just doesn’t hit home. And I’m spiraling into this.
Lex Fridman
(00:06:35)
Yeah.
ThePrimeagen
(00:07:02)
Like, “Maybe I’m not a programmer. Maybe I’m not worthy enough to enter into this circle of people who can figure out what the heck recursion means.” And Deitel and Deitel’s, I still remember this, their exact phrase was, “Every young budding developer solves this recursion program,” and it was the Tower of Hanoi. And guess what? I don’t know if I can solve the Tower of Hanoi to this day. It’s a very hard recursive problem. And I just sat there and thought, ” Oh my gosh. I’m not going to make it.” And I sat there in the lab for eight hours, 10 hours doing these things, so worried. It’s the week of recursion, we have to do a lab assignment. “I’m not going to be able to do it.” And I just remember being genuinely worried about that. And then…

(00:07:47)
Because always my big problem was like, “Okay, do factorial.” Why not just use a for loop? “Okay, what about Fibonacci sequence?” Why not use a for loop? I don’t understand. What’s the purpose of recursion? I don’t understand it yet. It’s so powerful. Why? It looks like a really complicated for loop. And so I just could not understand it. And then lab came that day and it was, “I’m going to give you a 2D array you have to read from a file. This is what a starting position looks like. This is what an ending position looks like. This is what a wall looks like. I want you to find me a path through the maze.” So I just sat there like, “Okay, well I guess I can just go up and I can create a visited grid, so I know not to visit these places anymore.” And then all of a sudden it just started clicking. Like, “Well, wait a second. I don’t know the maze, but if I just go up, right, down, and left, and hop back every time I’ve been to that square, don’t visit it, it will just go forever.”

(00:08:38)
And I realized in that moment, I’m like, “I actually understand. I’ve understood recursion this whole time, I just never had a problem in which it actually made sense to use. And that was my big downfall, is that I was measuring my understanding with the problems that I had available, which were just list traversal, which is not a good use of recursion. And so I just remember that freeing… Oh, man. Recursion. It was a great moment in my life.
Lex Fridman
(00:09:01)
I mean it does require, to be fair, a leap of faith, because people will tell you, those conformist, dogmatic, Java instructors will tell you, that this is important, to understand recursion. But it takes a leap of faith that this is a different way of looking at the world, and it’s a powerful way of looking at the world. Actually, I think I remember my first now.
ThePrimeagen
(00:09:37)
All right.
Lex Fridman
(00:09:40)
I think it was [inaudible 00:09:42] first search for one of the games. Maybe Othello, something like that, and for that implementing recursion. Understand that you can search trajectories through the space of states and do that recursively? That was mind blowing.
ThePrimeagen
(00:09:57)
Yeah.
Lex Fridman
(00:09:57)
Just imagining the possibilities.
ThePrimeagen
(00:10:02)
You can just see it all. Yeah.
Lex Fridman
(00:10:02)
Yeah, just numbers flying. It was like The Beautiful Mind. And that’s when I also discovered conspiracy theories, and I just saw. I saw the truth. Okay, yeah. So what were we talking about? Oh, what was the most painful aspect of programming for you? What memories do you have of deep, profound suffering in terms of programming in the early days?

Hardest part of programming

ThePrimeagen
(00:10:27)
I would say the biggest one that I can really hold on to had to be one of two experiences. The first experience was when I was at a place called Schedulicity, and… Am I not allowed to say the place [inaudible 00:10:43]
Lex Fridman
(00:10:42)
You’re allowed.
ThePrimeagen
(00:10:44)
I’m not sure if they’re even operating still at this point, but they’re in-
Lex Fridman
(00:10:47)
There was something funny about the name. I’m sorry.
ThePrimeagen
(00:10:48)
Oh, Schedulicity? Yeah. Actually, the name was so bad that when you looked at their paid-for Google ad terms that they would make sure that they’re at the top of the list, the spellings were just insane, because no one knew how to spell the word Schedulicity, and so the Google optimizing for that is just hilarious. But okay, go back to the thing.

(00:11:09)
The thing that kills me the most about programming, what I actually considered the worst aspect of programming, is when you know everything. And so when I was at this job, it’s just every single day I’d come in, there were no surprises, there was no questions. I didn’t understand the code base, sure, that’s fair. I didn’t understand all the things about the code base. But I knew I was going to go in, I was going to generate some sort of object from the database. I was going to take that object from the database, and I was just going to map it over and just display it on the webpage. There’s no creativity, there’s nothing to it. It’s very almost factory line kind of work. And that was a very difficult moment for me, which is…

(00:11:46)
I didn’t enjoy programming, because I knew everything about it. I already knew exactly what I was going to do that day. I knew all the hurdles I going to have to go over. There was no unknown unknowns, if you will. It was just knowns at all times. And for me, that is the worst part about programming, is when you already know the solution and it’s just a matter of how fast you can type and get it out from your head to your hands.
Lex Fridman
(00:12:10)
So the absence of uncertainty, the absence of challenge, was the pain?
ThePrimeagen
(00:12:14)
Yeah.
Lex Fridman
(00:12:15)
That’s pretty profound, Prime.
ThePrimeagen
(00:12:18)
I’m more than just good looks. I want you to know that.
Lex Fridman
(00:12:21)
It’s a low bar. What do you identify as? I’m enjoying asking the general question.
ThePrimeagen
(00:12:27)
38, male.
Lex Fridman
(00:12:29)
Male.
ThePrimeagen
(00:12:29)
Husband of beautiful wife.

Types of programming

Lex Fridman
(00:12:30)
Okay. You stream about all kinds of programming, but what kind of programmer are you? Are you full-stack developer, web programming? And maybe can you lay out all the different kinds of programming and then place yourself in that, in terms of your identity. Sexual identity as well.
ThePrimeagen
(00:12:49)
Yeah, we can put it all in there.
Lex Fridman
(00:12:51)
Okay.
ThePrimeagen
(00:12:52)
Plus, obviously those two are very, very tightly coupled.
Lex Fridman
(00:12:55)
I have seen you on the border of sexually aroused by certain languages. I think you got real excited about OCaml, or…
ThePrimeagen
(00:13:01)
OCaml. Let’s go. Thank you Dillon Mulroy [inaudible 00:13:08]
Lex Fridman
(00:13:08)
Okay, wow.
ThePrimeagen
(00:13:09)
Yeah.
Lex Fridman
(00:13:09)
I did not expect that. That escalated quickly. Anyway, what do you identify as?
ThePrimeagen
(00:13:13)
Okay, so first let’s do the previous or the in-between question first, which is the different archetypes. I think that’s a really interesting question, because if you go on Twitter or you’re new, your thoughts are probably that there is just web programming, and maybe there’s some other stuff, yeah, like game programming, but you’d be like, game programming in JavaScript and on the web. There’s this very myopic view of the programming world, and I bet if you ask a lot of people these days what is the most popular form of programming, they’d probably say web. If you said what contains the most amount of repos, how many percentage of repos on GitHub are web-based, they probably say 90% or some huge number. But the reality is that there’s an entire embedded robotics world.

(00:13:56)
You know, you’re familiar with the ML side of things. There’s networking, there’s going to be just performance, operating systems, compilers. There’s just huge amounts of variation of all these different types of programming verticals that you can be. And so we often talk about programming in perspective of web, or something that’s pretty narrow, and I think that’s just a social construct of Twitter more than anything else, that actually I don’t believe it’s that representative of the entire programming world out there. And I think a lot of programming’s really, really fun. There’s some really great stuff. Building your own language is just a very fun experience to do. Every programmer should just do that once, just to have a completely different perspective on how things work in life.

(00:14:36)
But as far as what do I do, I’ve always looked at myself as a tools engineer. So at my time, at my jobs, typically I would start off on the UI, and then they’d be like, “Okay, well hey, we need a library for this thing.” So then I’d be the one writing the library. So in 2012, 2013, I was writing a UI library for the web that can behave just like an iPad, so you can pinch and zoom on it, but it’s still a web page. Because we didn’t have any of that stuff back then. It was a canvas, had to do all the matrices operations and all that stuff to…
Lex Fridman
(00:15:04)
Nice.
ThePrimeagen
(00:15:05)
You know, it felt like you’re on an iPad, but it actually wasn’t on an iPad. And this was iPad 2 by the way, so this is a long time ago. And so every single time I got into a job it’s like, “Okay, hey, we need to do a library. Hey, can you work on a build system?” So back then there was no Grunt, there was no gulp, there was no any of those things, so I had to hand roll my own JavaScript build system. And so I always fell into these positions of building tools for developers to be successful. And I’ve always really enjoyed that region. So as I went on to say Netflix, spent 10 years there, I’d say the majority of my 10 years were building things for developers to use that they could be successful at their job.

(00:15:45)
And so I’ve always really enjoyed that aspect, because your shareholders and the people that use your program understand programming and they’re going to say, “Hey, I need this.” And typically the thing that they need, they actually want. Whereas with people, people want stuff, but what they actually need versus what they actually want often are this weird separation. That’s like the old Henry Ford quote, “I just want a faster horse,” and he’s like, “No, what you actually want is a car.” And so you have to play this game of trying to really figure it out, whereas developers, it’s like, “I know you know what I’m doing. I know what you want. Let’s figure it out together.”
Lex Fridman
(00:16:20)
Actually that gives you a really nice big-picture view of programming in general. So I love the idea of just starting at the interface, like you need to pinch and all that kind of stuff, and then figure out the entire thing that requires to make that happen, including maybe the side quest tooling, how to make it more productive and efficient, all that kind of stuff. So the entirety of the thing. That’s really cool.
ThePrimeagen
(00:16:43)
Yeah.
Lex Fridman
(00:16:43)
Okay, so that would be full stack? By that general definition of full stack, meaning…
ThePrimeagen
(00:16:49)
Perhaps, yeah.
Lex Fridman
(00:16:50)
Versus systems starting at the bottom and trying to optimize a certain kind of specific thing without seeing the big picture of what the resulting interface would look like.
ThePrimeagen
(00:16:50)
Yeah.
Lex Fridman
(00:17:02)
And a lot of people in web programming, they never go beyond the front end of how a thing looks. They kind of always assume there’ll be somebody, some grunt in the shadows, in the darkness of the basement, that will implement the back end.
ThePrimeagen
(00:17:17)
Some Gilfoyle out there will be doing the back end.
Lex Fridman
(00:17:19)
Gilfoyle.
ThePrimeagen
(00:17:19)
Yeah, I like to call myself a generalist, just to give some ideas. At one point at Netflix I built the WebSocket connection. So for TVs, how WebSocket works is code I just wrote. And so I built the framing thing, and before that I was doing stuff with memory, and before that I built a UI for a tool. Right? I can just do the thing. You just tell me the thing to do and I’ll just go do the thing. I don’t try to get super good at one specific activity. I don’t want to be a Kubernetes engineer, who’s the world’s greatest employer, but if I had to go learn Kubernetes, I’d go learn it and learn how to deploy some things, and then hopefully move on to the next thing, if that makes sense.
Lex Fridman
(00:17:58)
I posted about the fact that I’m talking to you on Reddit, and there’s a lot of wonderful questions. Somebody mentioned that I should ask you about DevOps. Can you explain what DevOps is? Is it a kind of special ops of programmers, is it SEAL Team Six of developers? What’s DevOps? Can you define… Are you a DevOps engineer?
ThePrimeagen
(00:18:14)
Well, people keep telling me DevOps isn’t real. There’s actually, you want platform engineers, cloud engineers, infra engineers. I just often think the easiest way, if we’re doing just some basic nomenclature, it’s just DevOps are the people that make sure that when you launch a service and all of that, it doesn’t just disappear. Right? It’s all the backbone of being able to operate something at scale.

(00:18:37)
Really, if you think about it, if you’re just writing a mom-and-pa website, people that do PHP that are doing WordPress and all that, they’re going to build something, they’re going to hand it off to, I don’t know, Linode, DigitalOcean, some company. They don’t really need a really complicated build, deployment, all this. It’s just someone with a simple website so they can sell their goods. And so they don’t really need that. And so that’s kind of how I think of a DevOps, is when things need to scale, that’s the person you hire.
Lex Fridman
(00:19:04)
Yeah, those people are actually amazing.
ThePrimeagen
(00:19:06)
Yeah.
Lex Fridman
(00:19:07)
The time I spent at Google, it’s like oh yeah yeah, there’s all these fancy machine learning people, but the folks that are running the infrastructure, basically that make sure that shit doesn’t go down, they’re like wizards, and they’re essential.
ThePrimeagen
(00:19:23)
It’s a very incredible vertical of job. And obviously I’m using a very broad term to describe, I’m sure, a bunch… You know, because making sure stuff doesn’t go down, you could also say that’s an SRE, right? Site reliability engineer. Whatever, the ones that wear the bomber jackets at Google. And so when we say DevOps, I think people get very particular about terms specifically in this category. They’re like, “Well actually, you’re mentioning infrastructure engineer versus site reliability engineer.” It’s just like, “Okay, yes, I hear you,” but generally when someone thinks DevOps, they think somebody that manages the servers and their life cycles and the reliability. There’s DevOps. Is it real? I’m not sure.
Lex Fridman
(00:19:58)
Okay.
ThePrimeagen
(00:19:59)
Did Vercel kill DevOps?
Lex Fridman
(00:20:02)
Question mark?
ThePrimeagen
(00:20:03)
Question mark.
Lex Fridman
(00:20:04)
Yeah. Wow, you’re almost a journalist. That’s a headline. Let’s go back to the beginning.

Life story

ThePrimeagen
(00:20:10)
All right.
Lex Fridman
(00:20:11)
Baby Prime. So you mentioned Netflix.
ThePrimeagen
(00:20:15)
Oh, I worked at Netflix by the way.
Lex Fridman
(00:20:17)
For people who don’t know who ThePrimeagen is, he mentions the fact that he has been very successful and has worked at Netflix in basically every other sentence.
ThePrimeagen
(00:20:29)
Correct. Almost as much as I mention Neovim.
Lex Fridman
(00:20:33)
Oh, great. Tell me more about Neovim. No, please don’t. So, baby Prime. At the very beginning. You’ve had one hell of a life, and I think it’s inspiring to a lot of people. You’ve gone through a lot of painful low points, including meth addiction, loss, and like you mentioned, you’ve come out of that to become a successful programmer and a person that inspires a huge number of people to get into programming, and just to find success in life. So maybe… I would love it if you laid out just your whole life journey from the beginning.
ThePrimeagen
(00:21:11)
So I guess if we’re going to start with this whole journey, I think it’s probably best to start to when I was about four or five years old. That was the first time I was ever exposed to pornography, and it’s kind of just earwormed me for a large portion of my life. And so I don’t think there was a day that didn’t go by from when I was a very young lad all the way up until I was twenty-some years old where I didn’t think about porn on the daily basis. And so it was just every single day, even that young. And so it was just a very mind-consuming, time-consuming, thought consuming thing that plagued me, starting at a very young age.

(00:21:47)
When I was seven years old, my dad died. That was a really tough period of life. I still think about this time that I went over to China, and there’s some rules that we were given, and one of the rules was just like, “Hey, don’t talk about God, and if you do, use the word ‘Dad’ instead.” And I was just like, “Okay, Dad!” It was the first time I said that word in 17 years or some long time. It was so weird to say that phrase. And I was just like, “Oh, that was just the strangest thing I’ve ever said in my entire lifetime.” It just felt so weird.
Lex Fridman
(00:22:23)
Yeah. Yeah.
ThePrimeagen
(00:22:24)
So, kind of rewind. As I got older, obviously was very good at computers, good at accessing porn, of course, played video games on the Internet. Fun fun kind of side quest story. I think the guy’s name is Lord Toc on Twitch. I can’t quite remember his name, but he built this game called Grail, G-R-A-A-L, and Graal Online. And when I was a young lad it was just like Zelda, except for it also had a level editor and it had a C-like language, and that’s how I discovered how to program, is I looked at these symbols and figured out what they meant, and then I was able to make things happen in the game. And that’s my introduction into programming. So thank you that guy, whatever your Twitch name was. But all right, so keep on going.

(00:23:08)
As I got older, I was super bad socially. I was not a very great social person. High school is brutal, got made fun of a lot, really I wouldn’t say had a great time during high school. Definitely felt very out of place or offset or maybe misplaced, if you will. I’m not sure what the right word is. And so of course at that point, I just always wanted to be accepted, to fit in and all that. I did forget to say one side story. After my dad died, my older brother, he started getting into drugs, and along with that he exposed me to pot, so at eight years old I was smoking some marijuana for a while there, until maybe 11 or 12, and took a break, and then again did a lot of that as I got a little bit older, but…

(00:23:55)
So I got a lot of these exposures fairly young. 16, 15 through 18, lot of drinking and all that. When I graduated, or as I was graduating high school, I had such sadness, if you will. I was very sad about how everything went, tried to commit suicide, obviously it was a very poor attempt and I’m still here today. I’m very happy about that aspect. I’m glad that I didn’t follow through with anything, had to go to the hospital and all that. And when I was done, I just still remember coming out of the hospital, and at that moment it’s kind of like something broke in you. Have you ever read the book Wheel of Time? It’s 14,000 pages or something like that, but right around page 12,000, Rand has to intentionally kill a girl, the main character. And that’s the moment he breaks, and he gets into like Hard Rand. [inaudible 00:24:46] Rand, if you will, for those that know Wheel of Time will appreciate all that. For those that don’t, very confusing, and I understand. Not the Amazon movie show, not that Wheel of Time.

(00:24:57)
So now that we go back onto it, at that point it’s just like something kind of broke in me, and I just didn’t care anymore. So all the social awkwardness, if you will, all that, just died away with me, but also so did everything else. And so I started using a bunch of drugs. LSD, mushrooms, meth. Did a bunch of meth, did a bunch of that stuff, and then went off to college and continued to do a bunch of stuff. I took too much acid to where for quite a few years, I had little squigglies on the side of my eyes whenever I’d walk by high contrast objects. And so it’s just that whole period of life was just kind of marked by just poor decisions.

(00:25:39)
And then sometime when I was about 19 years old, somewhere in that range, I just had this one evening where I felt the very dramatic and real presence of God. And I kind of had this choice, like Frodo, on a razor, where it’s like if I go either way, I’m going to fall off, and I need to change my life. You get to make the choice now. Do you want to do that or not? And so I remember going, “Okay, I do want to change my life. I don’t like this experience. I don’t like what I’m living. I am still very sad, I still feel very desperate. I still feel all those things. I’m just pretending to be this other person.” And then I just went to sleep that night. Nothing changed in my life. Everything was still the way it was. I woke up the next day, the same person, and I was just like, “Oh, that’s just such a strange, weird experience.”

(00:26:29)
And I just went about my day. And then I remember, I think that evening, I looked at porn, and all of a sudden I just had a conscious… just this deep, profound shame. And I was like, “I’ve never felt shame in my life. I have no idea what’s happening now.” And then all of a sudden when I smoked pot, I just felt deep shame. And when I hurt somebody or did something wrong, all of a sudden… It’s just like I got a conscious from that evening. That’s what my gift was, if you will. And just at that point, I didn’t even have a choice. I had to change my life, because for whatever reason, I’ve been changed in a moment.

(00:27:05)
And so from there I started actually trying in school. I always joke around that I got 2.14 in high school. I had a teacher hand write me a note saying I was the worst student she’s ever had. All that kind of stuff. I was not a really great student. And then in that moment it’s just like, “Okay, now life’s changed,” and I start trying to learn, and I try to become a good student. And it turns out it’s really hard. I was really bad. I still got Cs. I went and took pre-calculus and failed pre-calculus, and I’m like, “Oh my gosh, I used to be the smart math guy, and now I’m the idiot failing.” And so I’m just questioning myself and all that, and I spent hours upon hours in a studying, math learning center, and then just at some point years into this journey, I’m like a year and a half into this journey, at this point, something clicks, and I go from being the worst person to just immediately becoming the best.

(00:27:58)
Everything after that is just, I don’t know what happened. All of a sudden I was the best person at math. I started going into my computer science classes. I just really got everything. Everything, at just years after trying, just all of a sudden became easier. And I’m not sure if it happened over the course of weeks or when the easier started, but it was just first predicated by just a huge amount of difficulty. And then this is where I started really desiring and loving the process of learning, was when things started getting easier after all those years.

(00:28:27)
Because I just was motivated by this desire to do something, not thinking it was going to get any easier, and then all of a sudden it just started getting easier, and it was great. And that’s really where I guess I started having the biggest parts of my life change at that point. I started really, really, really wanting to never look at porn again, because every single time just such shame, and I really wanted to stop. And that was by far the hardest addiction to quit. Smoking cigarettes was also a really hard addiction to quit, shockingly hard addiction to quit, but porn by far was just the worst of them all. And then I think about 22, I was finally done with all kind of addictions, if you will, and then for a year I just worked in all that, and I think right around, maybe it was 21 and three quarters, somewhere in that range, I’m not really sure where I stopped all the addictions part, but… Or at least the outwardly addictions. And then at some point, six months later, a year later, met my beautiful wife. Things just started falling more and more into place. I loved more and more work. I loved programming. I started programming 12 hours a day. I watched the Social Network movie, and after that, I was just like, “I’m doing a startup.” And so that night I started my first startup, and I was just like, so… It was in PHP by way.
Lex Fridman
(00:29:41)
Nice.
ThePrimeagen
(00:29:41)
PHP, yeah, 5.2 or something like that. It was great. Great times. And I was just so motivated to do that, and I would just program for… Sometimes I’d program for 24, 36 hours straight, and just nonstop, that’s all I wanted to do at all points. I think my wife got a little sick of me. She would be like, “Can you drop me off at school?” And I’d be like, “No, I’m programming.” I was not a very nice… You know, I didn’t think through things that well.
Lex Fridman
(00:30:04)
Yeah yeah. Yeah.
ThePrimeagen
(00:30:05)
I was just so into it and I just did it nonstop, and that’s kind of how I became me, is that story, if that makes sense.

Hardship

Lex Fridman
(00:30:13)
Let’s try to reverse engineer some of the pain and some of the triumph. You made it sound easy at times. Let’s try to understand it better, maybe when you were seven years old. What do you think about the pain you’ve experienced there, losing your dad? What do you think? What kind of impact did it have on you? What kind of memories do you have at that time?
ThePrimeagen
(00:30:33)
The best way I can put it is that I just never knew what a dad was. I was young enough that I could maybe repress or just even have the capability of remembering things long-term. Because I know most people don’t remember a lot from when they’re young, and so I’m not exactly sure. I probably was at one of the best possible ages, if I’m going to lose a dad, to lose a dad. You know? If you’re going to lose one, if you’re 11 or 12, it’s a terrible age. That’s what my brother was, and he fell into drug addiction and never got back out. So I just have more of a fuzziness and just kind of a longing. I just wish I had a dad.
Lex Fridman
(00:31:10)
What impact did that have on your evolution, on your life, having that longing?
ThePrimeagen
(00:31:16)
I think that’s why I was so bad socially, in the sense that I was looking for approval, right? I needed approval. I think a lot of people desire that approval or that loving figure, and I just didn’t have that. So I think I just looked for it in everything else, right? If I were to psychoanalyze my actions. During the time, it’s not like I was actively thinking that, but yeah, I just always wanted something to fill in whatever that was I felt.

High school

Lex Fridman
(00:31:44)
I think a lot of people listening to this will resonate with your experience in high school. Being the outsider, being picked on, struggling through a lot of different complexities at home. What advice would you give to them?
ThePrimeagen
(00:31:58)
The worst part about high school is that you’re surrounded by a bunch of people your age and it feels eternal.
Lex Fridman
(00:32:05)
Yeah.
ThePrimeagen
(00:32:05)
You don’t think… The people that are around you, you feel like are the people that will be there for the rest of your life. At least that’s what I thought. And I didn’t really even realize this until many years later, that they are going to be some of the least consequential people in your life.
Lex Fridman
(00:32:21)
Yeah.
ThePrimeagen
(00:32:21)
Which is very shocking to think about, especially if you’re in it right now.
Lex Fridman
(00:32:26)
Yeah.
ThePrimeagen
(00:32:27)
Right? Right now they are everything that your experience is, your whole reality. And then one day it all stops, and then real life starts to begin.
Lex Fridman
(00:32:36)
Yeah.
ThePrimeagen
(00:32:37)
That’s such a shocking thing, and if I could just tell myself that, maybe I would have been a bunch of different person.
Lex Fridman
(00:32:42)
That’s so beautifully put. I mean, it is like a trial run. You know at the beginning of video games, there’s a little tutorial? That’s what that is.
ThePrimeagen
(00:32:42)
Yeah.
Lex Fridman
(00:32:51)
And actually that should be a chance to try shit out, to take risks, because real life will begin with, there is more consequences after that.
ThePrimeagen
(00:32:51)
Yeah.
Lex Fridman
(00:33:00)
Because real life will begin where there is more consequences after that.
ThePrimeagen
(00:33:03)
Yeah.
Lex Fridman
(00:33:03)
Here you can, if you like a girl, ask her out. Try, try shit. If you get picked on, hit that guy back. Try shit out.
ThePrimeagen
(00:33:11)
I’m not going to condone punching another person.
Lex Fridman
(00:33:13)
I will. Beat the shit out of him, and take some jiu-jitsu and learn how to take him down. And then that girl that rejected you will be like, “Hmm, maybe I’ll give that guy a second chance.” Be a bad motherfucker. It’s a chance to try stuff out. This is a very motivational speech for kicking ass.
ThePrimeagen
(00:33:31)
It is true. I mean, there is something very true about that, that I think especially… I mean, I have no idea what the girls experience of high school would be like, but as a guy, there’s definitely a lot of like physical requirements in high school. There’s a lot of physical measurement, at least where I grew up. I think that might not be true in all high schools, but if they’re filled with boys, it’s probably true.

(00:33:51)
And so it’s just like, yeah, it probably does help to do those things, to go to BJJ, to do any of these activities. Because even if you don’t ever kick someone’s ass, just having some level of confidence in yourself is probably a very valuable thing. But just remembering that this is such a short, tiny moment in your life is just like a huge help.
Lex Fridman
(00:34:12)
I mean, the way you phrased it is exactly right. That’s what it feels like. That these are the people that will be with you for the rest of your life and this is the whole world. And so that means that there’ll be just tremendous amount of impact if somebody picks on you or if you fall somewhere low in the hierarchy and the status hierarchy of this high school, that means you’ll be low in the status hierarchy of the world and you’re fucked for the rest of your life.

(00:34:39)
And that carries a tremendous amount of weight. It’s just why psychologically it’s extremely difficult to be… I think it’s understated often by parents, by society, how difficult it is to be a high schooler, how difficult psychologically it is, how it actually makes sense that some people would suffer from depression and be on the verge of suicide; is very, very difficult.
ThePrimeagen
(00:35:01)
Yeah, I think it’s even… People always say, “Back in my day,” blah blah blah. I think it’s genuinely harder today than it’s ever been in the sense that when I was a kid, there was a qualification to people. Meaning, this is a cool guy, this is not a cool guy.

(00:35:15)
Today, there’s a quantification of people. You have 32,514 people following you, you have 12. The people can visually… They can inspect your exact social value on whatever platform you’re on. And that has to be just so much harder.

(00:35:31)
And I can imagine that there’s a lot of just so much weight to put on that, that it’s just… it feels probably way worse and way more damning to be uncool because you have an exact number of how uncool you are.
Lex Fridman
(00:35:45)
Yeah. The challenge there. And the task, the quest is to remember that just because your social circle on social media and in high school thinks you’re uncool, it actually might mean you are cool. And you need to find that cool and grow it and let it flourish so that when real life begins, you can fucking come out of the gate firing on all cylinders because-
ThePrimeagen
(00:36:15)
That’s a great way to put it.
Lex Fridman
(00:36:16)
I think if anything, high school is really bad at picking out the cool people. Whatever the system, the hierarchy that forms, it’s such a basic bitch hierarchy. You’re good at very generic shit. That’s how you rise.
ThePrimeagen
(00:36:33)
Your parents bought you an expensive car.
Lex Fridman
(00:36:35)
Expensive car, right?
ThePrimeagen
(00:36:36)
Just-
Lex Fridman
(00:36:36)
Materialistic shit. Yeah, exactly.
ThePrimeagen
(00:36:38)
It’s a greedy search. See, they didn’t have a proper search, so they’re just hitting that local optima.
Lex Fridman
(00:36:42)
But the… I mean, even the objective function for that greedy search is just a really shitty one, where those people that win the game of high school are very often not going to be the people that win the much more exciting, beautiful game of life. So do epic shit and try stuff out.

(00:37:01)
The weirdos are the ones that are going to succeed, the weirdos in high school. Probably because they also get bullied and they get to be tormented more psychologically and get to explore their own mind and think through what it means to be a human being more.

(00:37:16)
Because if you’re winning in high school, you’re not being challenged, you’re not self-reflecting, you’re not trying shit out. So there is some degree to being tormented as long as it doesn’t break you. The porn addiction, that’s another powerful one that I think will probably resonate with a lot of people. And it’s interesting that you say that’s one of the hardest addictions to overcome.

Porn addiction

ThePrimeagen
(00:37:43)
Let me say it this way, some addictions have a much bigger societal look and porn is just not one of them, which makes it super hard. None of your friends are going to cheer you on. If you go on Twitter and say, “I quit porn,” they’re going to be like, “Well, that’s good for you but not everybody…”

(00:37:57)
No one makes that argument with meth, right? No one’s going to be like, “Well, not everyone has to quit meth, okay. It’s actually a fine industry and people who are the ones producing it, they’re good also, right?” No one’s going to make that kind of argument.

(00:38:08)
Whereas with porn, you’re going to have a whole thing and friends are going to think you’re dumb for doing it or whatever. It’s like you have… It’s a much more difficult one in just like that. So it feels accepted.
Lex Fridman
(00:38:21)
And I think it’s also an addiction you can practice, participate in privately and hide it from the world. There’s certain addictions that are harder to hide from the world for prolonged periods of time.

(00:38:31)
And porn addiction is probably one you can just have for many years and then it can deepen. That’s probably a serious issue. Boy, am I glad I grew up before the internet because porn is so accessible, so easy to go deep into that addiction. I mean, what can you speak about what impact it had on your life? Maybe some of the low points, but also how to overcome it?
ThePrimeagen
(00:38:56)
I’d say as far as impact goes is that you will have such a long and broken look at women. By the very, like I can… Again, I’m only speaking from a male’s perspective, that porn in its just most basic thing is that you use another person for your own desire or your own want. It’s not something that is deeply needed. There’s no need for porn. It’s purely a want-based activity or a lust, however you want, whatever word you can fill in there.

(00:39:28)
And it is purely an objectifying activity. Someone else is on display for your own enjoyment. And so I think you carry this around. I do think that the women that I dated during high school or the women after high school and college, I looked at them as a means to an end. And I think porn greatly kind of shifted that perspective in my head that I did not give the value that was desired to another person. It really devalues humanity just in general, is my perspective of it. And then it makes people into commodities. And I don’t think people are commodities. I think everyone has value.

(00:40:02)
And so during that, for me that’s like the great effect of porn, is that it’s just consumerism gone wild or materialism maybe, you could ask or argue, gone wild. And it’s extremely hard to quit, just like you said, because I can look at porn and then I can go out to lunch.

(00:40:21)
No one’s going to know. No one’s going to have any ideas. It’s a very private, it can be very short session. It doesn’t have to be something that takes… You can’t take acid then go out to lunch, right? Your whole day is going to be a very different day. And so it’s very quick, easy, accessible.

(00:40:38)
And then obviously there’s all the science and statistics, like men make worse decisions for some period of time after looking or being exposed to sexualized images. There’s the whole dopamine effect that’s just like you constantly need more and more dopamine. That’s why people typically don’t just watch five minutes of porn and call it a day. There’s like the hundred tab joke that’s always made on the internet. It’s because it’s just this constant dopamine cycle you’re constantly doing.

(00:41:01)
And all that stuff is great to say. And I’m sure statistics and science and all that stuff is really great arguments for some amount of people. But for me it just comes down to, is it really a good thing to do? Is it really actually something we want, is to value people in such a profane or just disregarding way? I just really think it’s just bad for the soul. Even if all the stats said it was great for you, I still say it’s actually bad.
Lex Fridman
(00:41:29)
Yeah. You have to look at the long-term big picture, psychological impact it has on your relationships with human beings in general. That’s my, more generally than just porn, my problem with the quote, unquote, “sort of manosphere”, is I think sleeping with a bunch of women is great, wonderful. But the problem is making that the primary objective of your life, similar with porn, is you devalue one of the most awesome things, which is intimacy. That’s true for deep friendship, that’s true for relationships.

(00:42:06)
And I think porn does that in its purest, darkest form, which is: the thing that matters is the sex, not the deep connection with another human being. And I think, again, going back to high school and the manosphere, the objective function if it’s to get laid, which helps with status and confidence and all…

(00:42:28)
All that is wonderful, I think. Again, can be an addiction. But the thing that’s even more awesome for a lot of people is a deep friendship or deep intimacy with a romantic partner. That’s also fucking awesome, and both of those are great.
ThePrimeagen
(00:42:45)
It’s objectively better to have… I would say that there’s no universe that exists or there should be no argument possible that exists that a guy who has meaningless sex has a better or a more meaningful life than, say, me and my wife who’ve been together for 15 years.

(00:42:58)
We have a very… I can depend on her in all circumstances. Whereas if you live that other life, it sure could be… It could feel great, but there’s no meaning to it. There’s no actual real value to it.
Lex Fridman
(00:43:10)
That’s absolutely correct. I do think that getting laid can have a tremendous positive impact on the confidence of a young man. I think just there’s a certain number of sexual partners from which you can collect a lot of data and it can free you about, like not to be so nervous about the opposite sex, not to be so nervous about human interaction.

(00:43:34)
And that will allow you to see the world more clearly and to actually find that one partner with whom you can be deeply intimate with. Sometimes the nervousness around this societally constructed value in getting laid can cloud your judgment.

(00:43:54)
And if you just release that by getting laid a bunch of times, then you could see the world clearly that getting laid is not nearly as important as you said, as finding the right human, including I should put in that pile, not just a romantic partner, but friendships, deep lasting friendships.
ThePrimeagen
(00:44:14)
Well, I mean I think you’re right that our society puts a lot of emphasis on getting laid. And I’m sure that’s true among any group of males throughout any point in history. I’m sure that’s a very common joke that’s never actually never stopped at any point. So I’m sure that exists but… And there’s probably some truth to the sense that after you’ve… Who was it? Jim Carrey. “I hope that everyone can get rich so they realize that money solves none of your problems.” The realization that this thing that society told you is hyper important is actually not the important part. It is a very important…

(00:44:45)
It’s a great sign that your relationship is healthy. Like if me and my wife were to have no sex at all for months on end, something’s gone wrong, which means what… we are no longer on the same plane. But it’s not also a good identifier. Just because you’re having a lot of sex, it doesn’t mean you’re having a good relationship.

(00:45:03)
And so it’s like a unique… I forget the right term here, but it’s a unique way at looking at the problems. And our society puts so much emphasis. And maybe that’s why porn was so hard to quit, but my guess is it’s just all the dopamine effect that it is.

(00:45:21)
But for me, the most important part and the thing that actually has real reward is having that… having just my wife. I do not look at… I desperately try not to look at any other woman. I’m hopefully not going to get caught… Mark Zuckerberged at the White House like that.

(00:45:37)
I don’t look at porn. My wife has complete confidence in me that there is not going to be a situation in which she has to question me in any kind of sense. And that builds a much more deeply, I would argue, a very deep relationship because the trust is that much bigger. I think the deepness of the relationship is probably proportional to the trust you have in each other. It’s very hard to have a deep relationship with no trust.
Lex Fridman
(00:45:59)
Yeah. And a probably a prerequisite, maybe a component of trust is vulnerability to where you take the leap of being vulnerable with another human being. And that vulnerability when reciprocated builds this really strong trust and it’s a beautiful thing. Yeah.

(00:46:20)
I personally just… Given my position, that’s even more challenging, being vulnerable with the world and there’s a bunch of people out there that want to hurt you for it, but I think it’s worthwhile anyway to be vulnerable.
ThePrimeagen
(00:46:36)
It’s always worth it. The risk is always worth it in some sense. Obviously, everyone has a different life they have to filter through their actions with, right? Because the person that has no, say, social following or anything, their risk reward profile could just be local impact, which could be just as damning or harming to them.

(00:46:54)
And so it’s always worth the risk though, in my personal opinion, because finding my wife has been obviously the most impactful or changing thing in my life. Or second most, I’d argue that one night with God would probably be the most impactful thing that led to everything else, but then the wife would be the next most impactful. I mean, I’m cleaning up after myself and stuff now. Changed man. I’m a changed man.

God

Lex Fridman
(00:47:17)
Can we try to reverse engineer that moment of you finding God. What is it at 19? Because it feels like that was a big leap for you to escape the pain, to escape the addiction or the beginning of that journey. What do you think happened there?
ThePrimeagen
(00:47:37)
I think it just felt like I just… There was no line that I wasn’t willing to cross. Everything was fine and just like… It just all of a sudden, just in that moment, it’s just like I had I guess some sort of deep fear and understanding I am going down a path. Is this really the path you want to go down?

(00:47:58)
And I don’t know what the result of that path would be or anything like that. I don’t tend to speculate on things I don’t understand. I just know that in that moment I had the option and I just chose… I didn’t want it anymore. Right?

(00:48:13)
It’s kind of mixed in this whole thing where it’s just like I had no value. I wrapped up all my meaning or value in having sex or getting laid, I had… All that stuff, all the things we just talked about, that was where all my worth was. And that is just such a terrible place to have your worth.

(00:48:28)
And it was just all came to a point. And I can’t tell you the day of the week, I can’t tell you anything other than it was nighttime and I was in South Hedges in Montana State University, go Bobcats, that’s about… Yeah, that’s the sign that we do at football games. Don’t worry about it. But that’s all I can really tell you because that night was no more or less special than some other night. It’s just the specialness was I got at least a chance to make a choice.
Lex Fridman
(00:48:58)
Because you find in that advice that you can give to others who are probably… There’s probably just an endless amount of people that are struggling with porn addiction now, young people. What advice could you give to them? How to overcome it?
ThePrimeagen
(00:49:15)
For me to overcome it, I had to realize that I was taking something away from my future wife. Some people would be like, “Oh, well, once you get a girlfriend then you can stop.” And it’s just like, “No, because you never stopped the problem.” You don’t stop a problem by replacing it. And so I didn’t have a girlfriend, I didn’t have all that. I just realized that I was truly taking away from something from my future wife. And I didn’t even know my current wife at that time. She was not in the picture. I’m not even sure if she was at Montana State University at that point.

(00:49:43)
And so it’s just that’s… Once I made that realization, I think it went from my head to my heart, which they say is the greatest distance in the universe. I finally got it. And that’s really where things change.

(00:49:59)
The ability to say like what’s going to help you change and all that, I don’t know if there’s… I don’t think there’s silver bullets, right? If someone could offer you a drug… I forget who says this phrase, but there’s this really interesting phrase that goes something like, he was a very depressed man and he was struggling with suicide and he writes about this in this memoir.

(00:50:18)
And he goes to these doctors and the doctors effectively say, “Well, here’s antidepressants, it’s going to help you.” And he says that, “Well, the problem was is that scientists told me that I could just touch my brain and make myself happy, and that’s it. They could reach in, they could configure some stuff and I’ll be happy.” He’s like, “For me, it was a lot like going out into a field and being able to take a drug to see the rain. I could look out, see the rain, it would fall down, it’d be silvery, it’d be beautiful, but all the crop would still die because there’s not actually any rain. I had to discover how to be happy myself.”

(00:50:48)
And so for me, it’s like the reason why I looked at porn is because I was unhappy. I was trying to find meaning. I was trying to find value in something, right? Something that was supposed to finally give me this ultimate satisfaction. And it just does not, no matter how hard, and no matter how much you think it will, there is no escapade, there is no pornography that will ever give you that satisfaction you’re looking for. That’s the reason why it’s addicting. And that’s my call to why you shouldn’t do it, but how to get out of it, I only got out of it by realizing.
Lex Fridman
(00:51:20)
I think that’s really brilliantly described. You knew that this thing you’re doing is preventing you from finding your future wife and future wife could mean more even broadly, this path to a flourishing, to a beautiful life.

(00:51:43)
I think there’s a lot of choices we make that are just preventing us from opening the door to whatever future. I think what’s really nice to do is to imagine, just like we said with high school, that there are a bunch of trajectories in life where you’ll be truly happy and you need to construct your life in a way where you have the chance to travel down those paths. And there’s a bunch of addictions, there’s a bunch of choices that prevent us from traveling down those paths.

(00:52:12)
So just believe that you’re going to have an awesome life and remove from your life the things that are preventing you from walking down that path, which is essentially what you did. It’s a leap of faith that if you let go of porn, that a better life is waiting for you on the other end.
ThePrimeagen
(00:52:32)
Yeah. I definitely can’t say how long it will take, a better life. But for me, there’s no way in the universe I could have had the relationship that I have without first making those steps because I couldn’t value my wife in the way that was proper for who she was. I would have valued her through the index or the lens that I currently was looking through.
Lex Fridman
(00:52:58)
Got to ask. So I’ve never done meth. I’ve never done meth.
ThePrimeagen
(00:53:03)
That was a great segue by the way.
Lex Fridman
(00:53:05)
Oh, man. I don’t know what the fuck I’m doing, honestly with this interviewing thing. But yeah, meth and LSD… I did ayahuasca. I did shrooms a bunch of times. And this topic, I should say that there’s a lot of, on Twitter and in the tech community in general, people speaking negatively about ayahuasca and some positively. I think it’s such a roll of the dice.

(00:53:34)
I had incredible experiences, but I don’t think I want to recommend it to anyone. It’s a risk, it’s a serious risk. It really is a roll of the dice that you could meet your demons and they could destroy you or you can meet your demons and let go of them. Or you could have experiences like I did, which is never… Apparently I don’t have demons. I’m pretty sure they’re somewhere in the basement, but I’ve never met them on drugs.

(00:53:58)
I’m always really happy. I’m happy drunk. I’m super happy on ayahuasca, just full of love. I don’t understand, I don’t understand where the demons are, but that’s my biochemistry, whatever that is.

(00:54:10)
And for some others, one trip could be amazing and the next one could just completely destroy you and wreck your life. So I don’t know what the recommendation from that is, maybe avoid it, but then all of us die and life…

(00:54:25)
I tend to lean into adventure but drugs is… If you fuck with the biochemistry of your brain, you can really destroy yourself in a way that it’s going to torment you. So I would generally recommend that people avoid drugs altogether, probably, unless you’re a crazy motherfucker. Hunter S. Thompson.
ThePrimeagen
(00:54:53)
What an intro to this topic.
Lex Fridman
(00:54:54)
I’m sorry. What’s meth like?
ThePrimeagen
(00:54:57)
That’s a great intro. I like… You are very correct in the sense that there is, at least when it comes to hallucinogens, there is a wild variance to what you’re going to experience. And there is no guarantee, there’s no… Just because you buy the product, it doesn’t mean you’re going to have a good time, right? There’s a lot of…

(00:55:14)
Personally, I find that stuff to be very… I believe in the spiritual realm, right? I believe demons and angels exist. I believe God exists. And that whole realm is like… I don’t know what it opens you up to, but it’s much, much different experience. Now, some people will be like, “Oh, it’s just a bunch of chemicals in your brain. They all get mixed up. LSD just takes all of your pathways and they all go… They all get kind of scrambled up in your brain.”

(00:55:37)
And it’s just like, “Yeah, the experiences are profound.” I had some really bizarre, very cool, very awful… I’ve had all the experiences in them all. I can just tell you that I personally always say the same thing. It’s like, choices that I made I can never take back. I would never take that away from myself because I don’t know if I would be who I am today without all those experiences going up to it.

(00:55:58)
But if you have not had that experience, I’m on your team, or at least partially on your team, maybe more severely, I don’t think you need those experiences. I don’t think they’re going to… You don’t have to put yourself through that to make a good decisions or to realize that people have value, right? You don’t have to do that.

(00:56:16)
So as far as what is meth like? Meth is like… If you’ve ever done cocaine, cocaine starts off with like a 15-minute dance party. Just… It’s just so intense. It’s so great. And then it just followed up with like a five hour… just feeling wiggly, right? I don’t know how else to describe it. Meth is like that except for I didn’t get as much dance party or any dance party, but instead I just got that part for like 12 hours.
Lex Fridman
(00:56:40)
Yeah.
ThePrimeagen
(00:56:41)
So did a lot of skateboarding, did a lot of running around, staying all night.
Lex Fridman
(00:56:46)
Would you say it’s a pleasant feeling or is it more like an escape from the loneliness of life? Is it pleasant or negative in the actual moment? Not the consequences but in the moment.
ThePrimeagen
(00:57:01)
So I mean, this is just a very interesting kind of area, which is that not… Universally, you can’t say that. Often you’ll find that there’s kind of these two groups of drug addicts. There’s those that like the opioids and those that like the uppers. They typically don’t like… There’s very few people in the drug world that do both. They’re really just like find their side and they go for it.

(00:57:25)
So is meth a thing that everybody’s going to enjoy? Well, categorically, as you can see, and just how people experience drug addiction, no. But for me it’s just I had a really… It kind of feeds into the ADHD nature of this… Because you know you’re kind of high energy, you’re like always in the moment. So it’s just like you’re in the moment, but it’s just like, “Oh, I’m in the moment!” Everything’s just so intense!

(00:57:48)
You just want to really be in the moment. And so it’s just experiencing that constantly. And so was that great? Well, some people… My wife always tells me this, being nervous or… I forget, the anxiety of a situation can also be the same thing as like thrill. I forget the exact way. She’s probably super disappointed that I messed this up.

(00:58:10)
But it’s like you could perceive those two experiences in very different lights. Some people get in front of a crowd and it’s thrilling. Some people get in front of it and it’s just the worst experience of their lifetime. They would actually literally rather die, which is a crazy thing to think about than stand up and speak.

(00:58:25)
And so for me, meth was that thrilling side, but at the same time, it still didn’t quite give me that thing I wanted, whatever I was looking for. I’d use it to help try to get that thing I want, but it was never giving me that thing I wanted.
Lex Fridman
(00:58:43)
Yeah. For me, I’ve had all really wonderful experiences. Do not recommend them. But like with shroom-
ThePrimeagen
(00:58:50)
That’s a YouTube policy by the way that you have to say, “By the way, do whatever you do, do not do a illegal activity.”
Lex Fridman
(00:58:54)
I-
ThePrimeagen
(00:58:54)
But I had great experiences, but whatever you do, don’t do it.
Lex Fridman
(00:58:57)
Mr. ThePrimeagen, I have no master. I don’t have YouTube or whatever. I’ll say whatever the fuck I want. I’m just-
ThePrimeagen
(00:59:06)
But seriously, you do.
Lex Fridman
(00:59:10)
… kind… No. No, I don’t give a shit about YouTube or anybody, honestly. I’m just careful about the words I say because just because I had positive experiences, I don’t want young people listening to this think they should try the experience. I think the much more powerful message is that life is awesome even without that. That’s something I definitely experiment with on the alcohol side.

(00:59:34)
So for me, I’m an introvert. I’m afraid of the world. Social interaction fills me with anxiety. Alcohol is definitely a thing that helps with that sometimes, but I think honestly it’s not even the alcohol, it’s having to do something while a person is talking to me. I could just drink a liquid. “Yeah. Mm-hmm.” There’s a social thing. With a beer, it’s like… “Yeah. Uh-huh. Yeah, we’re having fun.” And I think it’s…

(01:00:02)
For me, it works the same as… If the liquid actually looks like alcohol, it does the same purpose often because alcohol… If you have a whiskey or a beer looking thing, it kind of sends a signal that we should be having fun. So we’re socializing, right? We’re fucking getting crazy. And then that mean…

(01:00:24)
You don’t actually need the alcohol. You can get fucking crazy without the alcohol substance, but there is some kind of social signaling that happens when you have a drink in your hand. So I’ve been to get-togethers where I’m not drinking, but just doing a fake drink situation and I can also have fun. So I’ve been…

(01:00:47)
But that said, traveling across the world, there are times when you get to be able to down a bottle of vodka. That’s very essential for my line of work, but that’s almost like a cultural experience versus a necessary component of a successful social interaction and one that brings you happiness.

(01:01:05)
So not drinking… I think you can have fun and not drink too. So all of this… Man, I’m so careful saying drugs have had a good effect on my life because I think for most people, no, for majority of people, they will in the long term have a negative effect. So I think if you were to choose one or the other, just no drugs and no drinking means one day you can be the President of the United States kids. And I should say… Oh, man.
ThePrimeagen
(01:01:42)
That is-
Lex Fridman
(01:01:42)
That means Diet Coke-
ThePrimeagen
(01:01:43)
… his funniest line.
Lex Fridman
(01:01:44)
Diet Coke is great.
ThePrimeagen
(01:01:45)
That’s his funniest line, which is, “You would hate me if I drank.” Which I just like… To me, that tickles me to no end. Just like, “Oh my gosh, that is such a funny line.”
Lex Fridman
(01:01:52)
Self-awareness and humor is wonderful there, but yeah.
ThePrimeagen
(01:01:55)
But I am on your team. All of the reasons why I used drugs and all that, it’s some level of escapism. I’m sure that’s like… would be the archetype or the box I’d put that into or the pursuit of trying to feel something that cannot come from them.

(01:02:09)
It’s like trying to find meaning in your job. You can find satisfaction in what you do. That is a very good thing. You can find satisfaction and be happy with what you’ve created. You can be thrilled by the experience, but you cannot find… I doubt you can find purpose. Maybe some people in specific jobs.

(01:02:26)
This obviously have very broad strokes, I’m painting with. Like if you’re an EMT and you save someone’s life, maybe there can be purpose in that whole experience, right? So I’m not saying all things, but as programming goes, most programmers, you cannot just simply find your purpose.

(01:02:39)
And same with drugs, you cannot find that thing you’re looking for, but they are a very great distraction. And then at some point that distraction comes with a heavy cost. I think Dr. Faust would probably know the best about the heavy cost, but it’s just you’re making one trade for another and at some point the bill comes due and that bill can be very, very large.

Perseverance

Lex Fridman
(01:03:00)
The other moment you mentioned that I think is really inspiring is that you failed pre-calculus. You really struggled in school. You realize that school is really hard and then eventually you’re able to sort of persevere and, I don’t know, break through that wall of struggle. Can you, by way of advice, figure out what happened and what kind of advice you can give to people who are struggling?
ThePrimeagen
(01:03:25)
Yeah. I’ll paint it in a more clear picture, or a very fast speed run of it is that I took pre-calculus, failed. I took pre-calculus again, failed, took pre-calculus again and got a C. So I took it three times. Then I took Calc over the summer, so Calc 1. In that one at the end, the final… The final was a two-hour final. I finished it in 30 minutes and that was the highest score in all of the school. And I proceeded to be the highest score in all calculus and Diffy Q.

(01:03:54)
I was the only person out of 400 people to finish the Diffy Q final. And I got the highest grade. And so I was like… I got really good. So I somehow went from really bad to really good. And my only… The thing that I did is that I had to win. It was not a option. It was not like, “Oh, this would be really great.” It’s like, “I will not graduate, I will not finish my stuff if I cannot do this.”

(01:04:16)
And so every single day I got up, I went to my however many hour class it was. Right after that, I went straight to the math learning center, did those problems. When I got home, I just got the book and it had the odd answers in the back. And I would try to walk through the problems over and over and over and over again until I absolutely got it. And it just became this thing where it’s just I…

(01:04:37)
Just simple rote memory took over and the ability to just effectively have the times table, but for calculus, all stuck in my head. Inverse trig substitution, trig substitution, doing Taylor and MaClaurin series. All those things, just over and over and over and over again. Eventually they became easy. They became very easy. It’s just that I had to cram it in there.

(01:04:57)
And some people, you hear these stories, whether they barely show up to class and they get As, I’ve never been that person. I’ve always been the person that has to sit down, read through everything, and I’m bad at abstract concepts. I like the concrete into the abstract, not the abstract into the concrete. Very bad at talking about things theoretically, then trying to apply them. But if I can do it once literally, then it’s really easy for me to go into the abstract.

(01:05:20)
And so it’s just like… For me, it’s just I had… There’s no substitute for the hours. So if I were to give advice, it’s just that you have to have time in the saddle. Hour after hour will make you slowly better. And at first, it’s crushing. It’s defeating and it’s not fun because you are bad at it.

(01:05:40)
But then at some point you’re just not bad at it if you can just do it long enough, and you’ll start getting okay at it. And then at some point you might even get good at it. And when you get good at something, it feels amazing.

(01:05:50)
There’s like an exploratory thing. If you’ve ever played a musical instrument, you stop having to think about all the little teeny things you have to do to be able to play something correctly. And you start thinking about how you can explore that space.
ThePrimeagen
(01:06:00)
…play something correctly and you start thinking about how you can explore that space. It’s like it’s a completely different problem. And same with programming, programming has an identical kind of feel to it. It’s just like you’ll cross that barrier and it becomes magical as opposed to a chore.
Lex Fridman
(01:06:15)
Yeah. Once you cross that barrier, somehow other things become easier. But then if you want to have a truly successful life, then you find the next barrier. Yeah, I’ve always been the same. Everything’s come really hard.
ThePrimeagen
(01:06:27)
Yeah, I’ve had no free lunches. Everything’s just been a lot of pain and struggle.
Lex Fridman
(01:06:35)
I think somebody said that on this topic that you think work smarter not harder is a phrase that you dislike. Somebody on Reddit told me this.
ThePrimeagen
(01:06:47)
Yeah. I don’t just dislike it. I hate that phrase.
Lex Fridman
(01:06:49)
Okay. Tell me about your hatred. How do you feel?
ThePrimeagen
(01:06:54)
The reason why I dislike that is that there is kind of a hidden suggestion there, which is that you already know what smarter is, so just do that. That actually things should be easy. You should just not have to try that hard. You should just do the quick, easy, obvious path and boom, it’s done. It’s like I’ve never experienced that in anything I’ve done. Everything is actually really hard and most of the time I don’t even know what I’m doing, so therefore I don’t even know what smart looks like. And so for me, the only way I can learn how to work smart is by working very, very hard and knowing that there’s no shortcuts. And then when I finally figure out what smart is, when I work smart and work hard, it is that much better.
Lex Fridman
(01:07:39)
I think there’s a deep profound truth to that.
ThePrimeagen
(01:07:42)
There’s a lot of these phrases that just drive me nuts in our society,
Lex Fridman
(01:07:45)
But that one is… Sorry, that one is really accepted if you can just linger on it because it really bothers me as well. So one, which is a really nice thing you said, the presumption there is things should be easy and you’re a failure if you don’t see the easy path. That’s kind of the implied thing.
ThePrimeagen
(01:08:02)
Just work smart, daug, why are you putting in all those hours?
Lex Fridman
(01:08:05)
And so it makes a lot of people that struggle feel like they’re a failure because I don’t see it. And then the choice they have, well, I’ll just be lazy and then maybe the profound truth will come to me somehow. And yeah, I don’t think I’ve ever, and I don’t think I’ve met great engineers that find the smart way without the extremely hard work. The annoying thing about those great engineers is then looking back, they forget the hard work because they remember all the joy they now are experiencing from all the efficient, smart work they figured out how to do. They forget. So when they give advice they give the stupid advice of, well, just do it like the easy way

(01:08:53)
And here’s the easy way. But no, you have to put in the hours. Musical instrument is a beautiful example of guitar and piano. I’ve put in, I don’t know how many thousands of hours. And now when I’m explaining stuff jiu-jitsu as well, I sound like one of those people just relax in jiu-jitsu. By the way, just relax is a really wonderful thing for physical endeavors like piano and so on. But to learn how to relax your hand, how to relax your mind, your body and use whatever the biomechanics of your body to apply the correct kind of leverage and the timing and all that, that takes thousands of hours of learning. Just to learn how to relax takes a lot of really hard work. In jiu-jitsu that takes many months of getting your ass beat over and over until you ride the bus home crying.

(01:09:55)
Your ego completely shattered and destroyed. And then a little element is figured out late that night or next morning. And from the depression, there’s this little plant that grows this flower of insight. And you use that insight to then get your ass kicked again all next month and year. And then you grow and grow and grow. And from that you discover how beautifully simple jiu-jitsu is or Judo is, just speaking for myself, or piano or guitar. And then yes, the profound truth or the mastery of a skill feels simple when you finally arrive to it, but the path for most people is going to be a hard one.
ThePrimeagen
(01:10:46)
I think I should make an addendum to the phrase, I think the phrase should be work hard, get smart.
Lex Fridman
(01:10:51)
Nice. That’s a t-shirt.
ThePrimeagen
(01:10:53)
That’s what it should be.
Lex Fridman
(01:10:54)
Yeah, agreed. Okay, that was a tangent of a tangent.
ThePrimeagen
(01:10:57)
Can I say one more cultural phrase that I absolutely hate?
Lex Fridman
(01:10:59)
Yes.
ThePrimeagen
(01:11:01)
The journey is better than the destination. Everyone’s heard this. Just take one second to apply what that means. That means forever starting from now, you are only going towards a place that’s worse. That literally is what it means. Enjoy the journey, celebrate the destination, that should be what it would be but no. People say these phrases, they’re everywhere. There’s these very shallow phrases that have no logical bounds to them. You’re just like, why would the journey ever be better than the destination? I think this might even be a C.S. Lewis quote is that C.S. Lewis was like, nope, this is terrible. The journey is not in fact better than the destination.
Lex Fridman
(01:11:43)
I love the demotivational posters. Progress, moving forward is better than moving backwards even if you’re still going nowhere. There’s a lot-
ThePrimeagen
(01:11:54)
I feel that one so much being in California for a few years, that is painful.
Lex Fridman
(01:12:00)
Positivity, if it doesn’t break you today don’t worry, it will try again tomorrow. It’s just a lot of really great posters.
ThePrimeagen
(01:12:07)
I didn’t even know this was a thing.
Lex Fridman
(01:12:09)
This is a thing.
ThePrimeagen
(01:12:10)
Oh my gosh, I want that.
Lex Fridman
(01:12:11)
Yeah.
ThePrimeagen
(01:12:12)
Hey. Hi, this is ThePrimeagen. One thing that I forgot to mention in this podcast, which feels just so foolish to me for forgetting, is just what a big role my mom played in my life. She had to work 18 hours a day after my dad died. She really made her house be able to survive. I always looked up to her and I always thought her amazing. And she really was the reason why when I decided to get my butt kicked back in gear, she’s just someone who I looked to as an internal inspiration for me to continuing, to keep on going because I really wanted to make her proud. And all those years of just high energy effort, I really wanted to make sure that she knew that I was just so dang appreciative for it. So hey, I just wanted to say thank you. Love you, mom.

Netflix

Lex Fridman
(01:12:55)
For people who don’t know, you worked in Netflix.
ThePrimeagen
(01:12:59)
By the way.
Lex Fridman
(01:13:00)
By the way. Now, how did you go from there, from the hardship that we mentioned, from the struggle, from the addictions and so on to a place where you were working at this incredible engineering company and building cool shit there? So tell the Netflix story.
ThePrimeagen
(01:13:21)
Yeah, so I kind of alluded to it earlier that I wanted to do my own startup so for, I forget how long it was, one or two years or two and a half years, built a startup. PHP, jQuery, everyone’s favorite language is all put together. You can solve math stuff with jQuery. So I just was totally into just non-stop doing that. This is the height of Stack Overflow. I was asking really dumb questions on Stack Overflow like what is more pythonic? And then you get a bunch of up votes and try to steal a bunch of karma away, all the fun stuff to do. Good times. And I was just so into it breathing and I just breathe it in, breathe it out, and that’s what I do all day every day. And so it’s just like non-stop building of a startup. Ultimately that startup failed and so I had to go get a real job.
Lex Fridman
(01:14:06)
Can you say what the startup was?
ThePrimeagen
(01:14:08)
It is so wild thinking about it in the past. Before I tell you what it is, I want to tell one quick thing about my dad. My dad in the early ’90s, like ’91, ’92, was building kind of a phone card company where you’d be able to pre-purchase long-distance minutes. Now, if you remember the ’90s at about what ’97, ’98, ’99, 10-10-220, all those different things down the center, all those companies where you can pre-purchase long-distance minutes kind of came out and were very, very big. And so my dad was six years early to that notion and ultimately his startup failed. But he was just really early to something that would catch on really, really big specifically in the telecommunication space.

(01:14:51)
Me as I grew up and did my own startup, I did a startup where was text message marketing. This was in 2010 where you could receive, say texts about various deals, all that kind of stuff. And of course, 10 years later now you don’t stop receiving texts and text message marketing is all the rage. And so I also, much like my father had a startup in the telemarketing space in which was just like a half decade too early.
Lex Fridman
(01:15:15)
So is it fair to say you’re almost always ahead of your time, that you’re a visionary of sorts?
ThePrimeagen
(01:15:20)
No, in fact, I am not ahead of my time. I just got, some would say I got unlucky on that situation. But it seems so obvious to me at that time when I was doing it, 80% of phones were dumb phones. Most people had flip phones. When I went and sold Via Text, is what the name was of that specific product. And we had the short code via text too, so it was pretty clever, six digits. When I went out and sold it, I only had a flip phone during that time. I didn’t even have a smartphone because they were kind of untenable for a lot of people. So it’s kind of just wild times to think about.

(01:15:56)
But then after that, obviously had to get a real job. We were living in an apartment right next to campus, Bozeman, Montana. And the guy below us must’ve been on some amount of drugs. He threatened to kill us several times, would just scream and just lose his marbles all the time. Very unhinged man, angry downstairs man is what we call him. One time my wife dropped a battery, double A. Okay, so we’re not talking about a B battery or D battery. We’re just talking about a double A, drop it pa, land on the ground, “I’m going to kill you.” Like crazy, absolutely unhinged behavior down there. So I had to go get a real job, we needed to move out of there. We were going to start our life.

(01:16:34)
And so I worked at a small place [inaudible 01:16:38], which I kind of talked about the boredom there. Got to go to a place called WebFilings where I’m working just tons and tons of hours. During all that time I’m still trying to figure out startups. Did one where you could pre-wish your friend’s birthday messages, and then it would automatically send it via Facebook beforehand. We called it Greet Feed. It was pretty clever. Nonetheless, I say all that story because everything that I was doing was exploring, building, finishing things, working, learning about corporate life, learning how to communicate in corporate life, being able to be successful at a job, learning about a bunch of technologies that we’re about.

(01:17:13)
And one of the big technologies during that day, specifically 2013 was RXJS, if you remember that one, RxJS, that’s a link from C# kind of ported over to JavaScript.
Lex Fridman
(01:17:25)
And for people who don’t know, I guess C-sharp, what is its closest neighbor? Java. Is Java-
ThePrimeagen
(01:17:30)
They obviously just took Java and ripped it off at one point, but now it’s such a dynamic, interesting language that it seems like it could be a really cool bounds of practical versus not practical. It’s just I’m not really into wearing pleated pants and programming at a Microsoft house.
Lex Fridman
(01:17:42)
Is pleated pants a requirement?
ThePrimeagen
(01:17:46)
I think so.
Lex Fridman
(01:17:47)
Okay, we’ll get back to this.
ThePrimeagen
(01:17:49)
Can we just get back-
Lex Fridman
(01:17:49)
All right.
ThePrimeagen
(01:17:49)
Triggering me here.
Lex Fridman
(01:17:49)
WebFilings.
ThePrimeagen
(01:17:54)
So anyways, WebFilings was that’s where I had to do all the matrices stuff and build systems and just kind of all that. And it really pushed me because they also wanted me to do 60 hours a week. It was not very healthy work-life balance. It was very hard work. And kind of that really hard work going to cutting edge stuff, really understanding the world, really made it so that I was able to just be able to talk about stuff very commandingly because we had to build really complex state machines for the UI for what we’re building. And so when I went and started getting a LinkedIn and all that, inevitably just due to the fact that I’ve touched all these technologies and I had some sort of paper trail saying, I’ve touched these technologies or Microsoft. Dang it, Lex
Lex Fridman
(01:18:36)
Pleated pants.
ThePrimeagen
(01:18:36)
Pleated pants reached out. No, Netflix reached out and said, “Hey, I see you’ve done RxJS. We do a lot of it. You want to come and interview with us?” And I was always told that you should never reject kind of a handwritten personal invitation to interview. This was way before bots and even the bots were pretty obvious to tell they were bots. This was a manager at Netflix, Jeff Wagner, first manager ever. And he just wrote a really nice note and just like, “Hey, I see you’re doing a lot of these things. We really need help with JavaScript. I would love for you to come interview. We’re even using a lot of RxJS if you’re interested in that.” And so I was like, all right, I can come and I’ll interview. And lo and behold, interview went on and I called my wife I think halfway through the interview and I was just like defeated, absolutely crushed because I said…

(01:19:28)
And she might remember this but I said, ” We now have to make a decision. Are we actually going to move to California or not?” Because I already knew I had the job at that point. I was just knocking them out of the park. I was doing a great job on that. And so I just knew for a fact, I’m getting a job at Netflix. There’s this thing that people always get so freaked out about when it comes to interviews and all that. And I luckily somehow avoided this. I don’t get test anxiety, I don’t get any of that because when I go into these situations, my only goal is to show the things I already know. And so it’s like I walked into this situation, I’ve been preparing for this 80 hours a week for the last five years. So just walk in and I’m just showing the things I know.

(01:20:08)
And it was perfectly fitting for Netflix at that time period in the 2013, early JavaScript days on television. And so it was just awesome, just worked out perfectly. Got hired there.
Lex Fridman
(01:20:19)
So we’re in California with Netflix. This is San Francisco.
ThePrimeagen
(01:20:22)
Los Gatos. So if you’re familiar, so classic symbol people do which is this is San Francisco, Oakland, San Jose, Los Gatos is just kind of a little bit south of San Jose, same mega contiguous city.
Lex Fridman
(01:20:39)
Yellowstone is in Montana, Yellowstone, the show. So is it basically like that, Kevin Costner riding on a horse? Were you riding on a horse to campus?
ThePrimeagen
(01:20:48)
No, but I love those stereotypes. Actually to be completely fair, when I was 15 years old, I was driving around on what is now a very busy populated street shooting gophers out the window of our car with a 22. So it’s like Montana was a different place at one point than it is today. And there’s plenty of parts of Montana that’s still very rural, still kind of more of that old world. So yeah, a little bit you can get whatever you want from Montana. As far as culturally goes, I’m not really sure the best way to put the difference between California and Montana. It’s just different expectations. One thing I can really appreciate about California, or at least when I say California, I mean the Silicon Valley Because obviously LA and the Silicon Valley, very different attitudes, very different mindsets. You can’t really compare one to the other.

(01:21:38)
One thing I can say that’s really positive about the valley is that everybody is operating on this idea of trying to build or create something, and there’s an energy to it that’s very exciting. You meet somebody and they have a startup and they’re working on the startup. And it’s very exciting. And there’s a lot of negative aspects to that, and we can all agree that our entire life being commercialized has probably not been that great. But the kind of the experience of being there and everyone’s excited to build something, it’s a really cool experience.
Lex Fridman
(01:22:06)
Yeah, it’s really great. The excitement, the energy.
ThePrimeagen
(01:22:09)
Yeah, Montana doesn’t have that.
Lex Fridman
(01:22:11)
I have a romantic admiration for the shows like Yellowstone being out on nature. It’s beautiful. I like riding horses. Somebody also said Reddit is full of wisdom about you. Some of it could be fake news, but something about horses and this kind of thing. You like horses, you like riding horses?
ThePrimeagen
(01:22:28)
We have horses on… Our neighbor had much more hilly land and one of the horses broke its leg, so they had to put it down. And so we just said, “Hey, we are on much flatter land. You can just have your horses in our property.” And so we just have horses that run around on our-
Lex Fridman
(01:22:44)
What about milking cows? Somebody asked about cattle and cow.
ThePrimeagen
(01:22:50)
So I’ve only had open cows. If you don’t know, cow means girl, open means that, hey, they’ve tried to get the cow pregnant. The cow did not get pregnant first try. And so they’re calling that gene, they’re getting rid of that gene. The open cow is going to now go out to pasture for the year and then get turned into delicious T-bone steaks and various things. And so we would house open cows on our property. So no, there’s no milking of open cows.
Lex Fridman
(01:23:16)
Okay.
ThePrimeagen
(01:23:16)
They’d be very upset if you tried to milk an open cow because they’re not milking cows. You have to get that cow pregnant. And then once you get it pregnant, you have to kind of put it into this permanent state of milking and all that. And it’s a little bit more complicated than say what we did, which was just cows on eating grass and I didn’t have to touch them.
Lex Fridman
(01:23:35)
Okay, well, that’s wonderful.
ThePrimeagen
(01:23:36)
Reddit is not a great place for wisdom about me. They’re going to give you the craziest answers.
Lex Fridman
(01:23:41)
We’ll return to Reddit time and time again, my friend. So yeah, you took the leap into Netflix. So what was that like?
ThePrimeagen
(01:23:52)
This is one of those things where when you talk about it, people love to trivialize this because it’s like, oh, you’re taking a leap of faith by going into a fang company. And in 2013, sounds super risky. My wife was 36 weeks pregnant. We had to travel to a place where we knew not a soul. We were about to have our first kid. We didn’t even have a doctor. If you don’t know what having a baby does, you kind of want a relationship with a doctor. There’s a whole thing that goes on there. So it was a really hard and great experience. So I went to a job in which their culture deck… So during this time, this is where Netflix still had kind of that old generation X feel to it. Their culture deck was hire fast, fire fast. It was very in your face about like, “Hey, this is how we operate. You don’t meet the standards, we kick you out.”

(01:24:39)
So it’s like I’m leaving a place where it’s more secure to go to a place I don’t know anybody, to a job that’s bold in its claims about firing everybody with a wife that’s just about to have a baby. And I’m from Montana and every Montanan’s born with a natural dislike of California. So there’s all these things kind of flowing into it where it’s just going to be like, wow, this is a very intense experience. And it was hard for sure. It wasn’t just some easy simple experience that we were just like, oh, well, I work now at Fang. We had to kind of work through that. Having a kid was very difficult. Our first kid was very difficult. Not having any family around to ever help you took a much larger toll on my wife than me, for sure.

Groovy

Lex Fridman
(01:25:23)
What was the technical learning curve for you? You showed up in your plaid pants, dressed up?
ThePrimeagen
(01:25:29)
Yeah.
Lex Fridman
(01:25:30)
What did you have to learn about the Stack? Because Netflix, I imagine is this incredible infrastructure. It has to deliver just a huge amount of data. I’m just blown away by Netflix but also YouTube. These companies that have to deliver, serve a huge amount of bits.
ThePrimeagen
(01:25:55)
Netflix has it easiest. Out of all the companies Netflix by… Even though you could say maybe we beat YouTube in view hours, I’m not sure if we do, but let’s just pretend Netflix has five x more view hours than YouTube. Whatever it is, Netflix has a fundamentally easier problem than all other companies. And let’s get back to that. I’m going to first tell you about the Stack, but I’ll tell you why it has a fundamentally easier problem. So when I first got there, they gave me my PlayStation three. My boss said, “Go learn some code. Come back to me in a couple of days and tell me what you’ve learned. And then I’m going to start giving you bugs to fix.”
Lex Fridman
(01:26:27)
Wait, PlayStation three, what are you talking about?
ThePrimeagen
(01:26:30)
Well, I was on the TV team. I had to go plug in a PlayStation and start launching programs onto the PlayStation three and figure out how to work Netflix on a television device.
Lex Fridman
(01:26:38)
Oh, so you have different kinds of devices. Why PlayStation three, is other different-
ThePrimeagen
(01:26:43)
It’s just 2013. That’s what you have.
Lex Fridman
(01:26:44)
Any devices that plug into the TV? Okay, cool.
ThePrimeagen
(01:26:46)
Yeah, not as many TVs had Netflix, let alone what they called their Darwin app, which is their new application. So if you bought a VIZIO earlier that year, you’d get their older one there. It’s called Plus UI. You get their older version. And so not many had the newer version. We no longer supported Plus or we never actively developed on Plus, we only did stuff on Darwin. And so I had to learn that whole stack, the backend or the middle end, the middle layer between the actual backend and the front end was written in Groovy. And as I went around, Groovy is if you’re not familiar with Jenkins, then you’ve probably never interacted with Groovy. But Groovy is a JVM language. It’s a very interesting language, but here’s how it got started at Netflix.
Lex Fridman
(01:27:30)
Oh, it’s Apache. Apache Groovy is a powerful object-oriented programming language that runs on the Java virtual machine released in 2007. It has evolved to become a versatile language that combines both static and dynamic typing capabilities.
ThePrimeagen
(01:27:43)
All right, so the AI is kind of lying to you. Groovy is not a powerful great language. That statement makes it seem way cooler than it actually is. You’ll meet one out of 100 people that have touched Groovy that said, “Oh yeah, Groovy’s great.” The other 99 will be like, “Heavens forbid, you ever have to touch that language.” So when I got there, nobody, not a single soul at Netflix, those 40-some engineers had any idea how Groovy pretty much worked. Somehow people just hacked together these scripts and put them all on there and it worked. And this was before there was a Groovy RX port. We wrote our own version called WX. It was a nightmare, observables, all these things. I remember one time they told me that, “Oh yeah, with RX it’s really easy. You just say what you need to do. It maps out and boom, boom, everything will run and all that.” And I was like, “Oh wow, really?”

(01:28:33)
So all I did was go like observable.sleep one because I just wanted to see it sleep and then do the next thing. And it turns out when a thread sleeps itself, no thread can wake it up. And I just turned off all of staging because I ran it like 10 times. Like, oh, it’s not responding. Oh, it’s not responding. Oh, now it’s not even coming back. Broke all of staging for everybody. So no developer could work for the rest of the afternoon because I locked up all the instances because it turns out no, it was in fact not multi-threaded. Every assumption we’ve been told is a lie. No one had any idea what they were doing. It was a wild time. And so I just simply naturally gravitated towards that because I’m good at print off debugging. I’m good at doing those things.

(01:29:12)
So I was like, yeah, I’ll just figure this out here. I will do this. So I had the rewrite how we do the data structure on the front end for the TV from what is called a LoLoMo, list of list of movies into LoLoRoMo, which is a list of list of recommendation objects for movie. Why would we need to do that? Think about this. You have two lists, one has Live Free Die Hard, Bruce Willis because you love Bruce Willis. The other one has Live Free Die Hard because you want tough men doing tough jobs. Well, during those days we’d only have one way we could show evidence why you wanted it. So we couldn’t say, “Oh, because you liked this other movie.” You’d go to that one and say the same thing. So we had to add one level of indirection where we could decorate the video with the recommendation information.
Lex Fridman
(01:29:52)
Okay. So you can abstract away into the space of recommendation versus the space of movies directly.
ThePrimeagen
(01:29:56)
Yeah. So you can’t hang it off the video because obviously then it would be the same for everything that shows that same video.
Lex Fridman
(01:30:01)
That’s amazing.
ThePrimeagen
(01:30:02)
I had to do all this and I wrote it in Groovy and I just did it-
Lex Fridman
(01:30:02)
Such a funny name.
ThePrimeagen
(01:30:06)
And people were like, “How did you write this in Groovy?” And I was just like, “Well, I read the language reference for a day and then programmed it.” Well, what do you mean? It was a very radical language, shall we say. And so I just simply became the person that knew these things, so they just give me more and more jobs with that. And so that’s kind of how I excelled, being the person that was willing to do the thing that no one else was.

Printf() debugging

Lex Fridman
(01:30:28)
Yeah. Can you actually speak to the print off debugging. You walk into a system and there’s a lot of systems in the world like this. Twitter was like this, when Elon acquired Twitter and the rolls in and there’s this old junky code base that’s just like a giant mess, and you have to basically do print off debugging. What’s the process of going into a code base and figuring out what the fuck? Well, how does this work? What are the flaws? What are the assumptions? You have to reverse engineer what all these other engineers did in the past and the mess across the space of months and years, and you have to figure out how all that works in order to make improvements.
ThePrimeagen
(01:31:06)
I’ve always just been good at print off debugging because one of my first kind of side quest jobs that I got was writing robots for the government when I was still at school. And so I’d kind of do this contractually for so many hours a week. And my boss, Hunter Lloyd, great professor by the way, he just said, “Hey, here’s your computer, here’s the robot, here’s how you plug it in. Here’s how you run the code. Can you write the flash driver, the ethernet driver. Can you write the planetary pancake motor? Here’s some manuals, I’m missing some. Just figure it out, I’ll be back.” So that was government work for me. So I was like, okay, I’ll figure all these things out. And I figured them all out and the only way to really get anything out of the machine was to print. And so it’s like I had to become really good at printing my way through problems. And so that kind of became this skill I guess I adopted is that I can just kind of print off debug my way through a lot of these problems.

(01:31:56)
Obviously I’m not a game developer, probably a different world probably should use… I think John Carmack was on here and talked to how great the debugger is, different world. Because when I was at Netflix, there’s machines that exist somewhere on AWS, I’m not logged into them. I don’t even know how to log into them. I’m not even sure if I have credentials to log into them. They run once somewhere and I have to figure out what happened and why it’s happening. So it’s like I’m going to become… This is what I’ve trained for. I’m a print off debugging champion. So it’s just like I could just run through these things really quickly and figure out why they’re happening the way they’re happening.
Lex Fridman
(01:32:26)
You’re a special human. I think that’s an incredible skill set to have to be able to drop in into any code base, drop into any situation, and do print off debugging. Meaning you’re in a dark room and you’re feeling around that room to try to figure out what the room is.
ThePrimeagen
(01:32:40)
Well, I had the code so it’s like I can kind of blueprint what’s happening. I don’t understand the services or anything, but you can start guessing pretty quick as to what’s going wrong.
Lex Fridman
(01:32:48)
Right. But then the print side of that helps you confirm your intuitions, test your intuitions and build up more and more information. And then you start to accumulate this bigger picture from that, what the edge cases are that break the system and not. I think that just that kind of situation is intimidating for a lot of engineers. They break down at that point. I think it really is a powerful thing to be able to come into a code base, that’s generally a skillset of very few of us start from scratch. And actually this is the fundamental problem of web development and in general where they’re like, I don’t know what’s going on. I’m going to write my own thing from scratch. As opposed to actually doing print off debugging on the space of languages, on the space of problems, because there’s a lot of wisdom and solved problems already in this code base.

(01:33:50)
It’s a much more important skillset to understand, to learn from the mistakes and the wisdom of the past, of the ancestors that came before and build on them as opposed to throw it all out and start from scratch. This is something obviously you see a lot with a JavaScript framework that comes out and you won every single day.
ThePrimeagen
(01:34:12)
I have a very great story about that, that this is what I think has shaped me the most about my perspective of other devs. There’s this dev and he always just wrote things in just what I thought was such a bizarre and weird way, and this had to do with Falcor. So our data fetching library for Netflix, This would run on mobile. So I had to write in Objective-C. It had to run on television and it had to also run on web. So it ran on everything. And me and one other person were responsible for this thing working. And the request side where we’d have to de-dupe the information that we already have, the requests that were pending and the new data.

(01:34:45)
So I had to figure all that out based on what someone’s requesting, and then just only optimally request the stuff that we don’t have. He wrote in such a goofy way and I’m thinking, man, this guy is just… What a goofball. So I delete it all and I start writing and I’m like, look at how much nicer this is. It’s looking so good. I’m like, Ooh, there’s that one edge case. Okay, I can see why he wrote it this one way. That’s not a big deal though. The rest of my code’s really great. By the end of it, I’m like, I literally almost line for line just reproduced what he already wrote. It’s slightly different towards my style, but I just wrote the same code. I’m like, I’m an idiot.

(01:35:20)
I am the idiot in this situation because it was already a solved problem. I just didn’t take the time to learn what he did. Instead, I relearned what he did by rewriting the entire thing.
Lex Fridman
(01:35:29)
I think that’s a skill set that is extremely important for people to learn. I see that in myself. That’s a constant struggle for myself when facing a code base, for example. But this applies generally in life, where somebody did a lot of work to do a thing, you should invest a huge amount of time and get really good at figuring out what they did, why they did it. Do a lot of print off debugging to understand what they did. It’s a much more efficient way to understand a problem deeply than to start from scratch. Even though there’s a constant temptation to start from scratch, because starting from scratch is fun. You do get the puzzle solved and all that kind of stuff. It’s just not going to be the right thing to do. Usually pain is the right thing to do, and it is for most people painful to understand other people’s code bases.
ThePrimeagen
(01:36:21)
I highly recommend starting from scratch if you want to understand a concept. You don’t know how an HTTP server works, create a TCP socket, learn how to parse HTTP. It’ll become very easy and you’ll go, this is the reason why whenever I get a request, I have to await the text. I now understand why the text is for whatever reason not there. I get it. I now understand it. And so you kind of gain these new perspectives just by simply parsing something out.

Falcor

Lex Fridman
(01:36:50)
All right. Back to the wisdom of Reddit. Apparently there are memes and legends about your programming arc in Netflix. This Falcor system you mentioned, somebody, I think it was Teej, how do you pronounce his name by the way?
ThePrimeagen
(01:36:50)
Teej.
Lex Fridman
(01:37:05)
Okay, Teej.
ThePrimeagen
(01:37:06)
TJ would be his name, but we call it Teej or Telescopic Johnson.
Lex Fridman
(01:37:10)
Oh wow, so many names. DDoS, distributed denial of service attacks, you apparently were able to accomplish the simplified version of that of just DoS. That’s a legend. So you basically broke down the system somehow.
ThePrimeagen
(01:37:25)
Yeah.
Lex Fridman
(01:37:26)
Can you tell the story of that?
ThePrimeagen
(01:37:28)
I’d be glad to. So there’s this Falcor business, and I did discover the bug before anybody else and I did report it to security and it was so bad. It actually got its own name, Repulsive Grizzly Attack, and they even give examples of how to do it. Effectively what it means is that there is a request that targets both memory and CPU and will destroy. There you go. Look, how Netflix… The next one down was the article that was actually written. I don’t get mentioned, which is a little bit upsetting considering I was the one that discovered it and told everybody how bad it was.

(01:38:01)
Anyways, and had the right to fix for it or the first fix. So this is how it works, is that you can do something pretty similar, I believe with GraphQL as well. It has the same kind of danger. Any of these kinds of RPC request as much or as little of the data as you would like frameworks, are vulnerable to this kind of attack. So with Falcor, what you do is you give it an array. That’s an array is called a path, and that’s the path to the data. But sometimes you don’t want to have to write out, I want movie, I want row zero or list zero or row zero column zero title. I want row zero column zero description. You don’t want to have to write out all that.

(01:38:42)
So instead you could just be like, I want rows zero through 10, columns zero through 10, titles and descriptions. So you can write in a very compact, nice little format and it’ll give you all that data. It’ll go to the server. The server will fill that all in and give it to you. Oh, dang it, list three, it only had three videos in it. So what happens when I try-
ThePrimeagen
(01:39:00)
Three, it only had three videos in it. So what happens when I try to re-request the data? Well, I need a way to be able to tell my system that you’d have requested the data and there’s nothing there. So call this like a boxed value. So it’s going to be like type, something, value, there’s nothing there. We’ve already requested it and there’s nothing there. It’s like a sentinel value, if you will, a boxed value. And we have this little special flag weed pass called materialize. Meaning that when you ask for a path, we will make sure we fill it out so we don’t accidentally erase anything. And at the very end we’ll say okay, the thing does, the request you’ve made has already been made and there’s nothing there. Well, what happens if I request rows zero through 10,000 columns through 10,000, one more item through 10,000 and then a whole bunch of properties and then ask it to materialize?

(01:39:49)
Well, I’m about to go create billions of objects in the JVM, and what happens to the machine? It stops running. And then if we try to JSON… Even if it could create a mall, we then ask it that JSON serialize, it’s not going to do it. It’s impossible. And so that was the attack vector, is a simple wild loop would’ve taken down and held down Netflix for a very long time. Because one request would kill one machine on AWS. And so that means it would just turn it all off. And this was on the website? This was on TV, this was on mobile. This was profound. And here’s the worst part, it was in production for years so we couldn’t even roll it back. There was no like, “Oh crap, let’s just roll back to two weeks ago and we’ll fix forward and figure out.” No, it’s like we could roll back to 2011. That’s our option is 2011 and that’s it. So we had to figure out a way forward and all that. And so it was like… The amount of problems that would’ve happened if someone would’ve discovered this is unstatable.
Lex Fridman
(01:40:55)
Just to be clear, the infrastructure that’s serving the videos would shut down.
ThePrimeagen
(01:41:00)
Yeah, the UI, you couldn’t perform any actions in the UI. You surprisingly could still stream video but you would never be able to get to a video to stream. Because every action you would take would be completely shut down. And so it wasn’t a DDoS because you didn’t need a bunch of computers to try to overwhelm the system by making a bunch of requests, one request, one machine. If we had 50 machines serving the millions of requests, it’d only take 50 requests to shut down the entire UI.
Lex Fridman
(01:41:23)
Isn’t it possible to do DoS or DDoS on basically any software system Like defending against all the closing all those attack factors is probably really difficult. If you take any sufficiently complicated software system, there’s probably so many ways to overwhelm it.
ThePrimeagen
(01:41:42)
Yeah, this is why people use CloudFlare. I think d HH said it best, which is like we have our website and we have a strong bodyguard on the outside. So CloudFlare has a bunch of utilities all built in. Because obviously this is why everyone hates all these Bluetooth devices that connect to the internet because they just turn into attack vectors where people use those to DoS or DDoS other sites. And so you don’t need something sophisticated, you just need a bunch of requests to come in and you can take down websites. And so that’s why these fronts are really good at discovering where these problems are.

(01:42:13)
But DoS is a bit different, because it doesn’t have to be overwhelming by using resources with a whole bunch of requests. It really just means simply that there’s a denial of service attack. One of them could be there’s a RegEx attack that existed where CloudFlare actually did it to itself and shut itself down, which is there’s a RegEx expansion attack where given the right RegEx, if you know someone’s running a specific RegEx, you can actually provide input that is maximally bad and that thing goes to super processing. It takes 10 seconds to process a single request, then you only need to make hundreds of requests and you shut down the whole service. It’s not like you need some giant machinery to make one trillion requests. You only need just some small amount to completely destroy a service. And so there’s… The web is an extremely difficult place to do it correct.
Lex Fridman
(01:42:59)
This is super fascinating. I do also wonder how many ultra competent, what is it, black hat hackers there are, versus the good guys versus the bad guys. How many bad guys there are and what is the average… What is the distribution of skillset on the bad guy side that are constantly trying to attack?
ThePrimeagen
(01:43:23)
I assume there’s probably a huge number of just really simple ones, script kitties, right? Just people trying to just do things. And then there’s a huge amount of social engineering that just goes in where hacking’s done, not with a computer but just by one of the classic ones. Kevin Mitnick had this one in his book which was you’d call up somebody pretending to be like, “Charlene, we’re doing some auditing and I think your pin’s out of date on file. Is it 2323, still?” And they’re like, “No, it’s 4747.” You’re like, “Oh, thanks Sharon.” Boom. You just hacked him. Right? The classic people love correcting bad information.

(01:43:57)
This is like a standard. So there’s all these ways people hack. And so my assumption is that there are really great white hat hackers, there’s really great black hat hackers. But the vulnerability space, the harp, the thing is that discovering a vulnerability and you don’t let anyone know, the white hat hacker still has to make that same discovery. And that’s where I think the real thing is that black hat hacking in some sense has a fundamentally easier job or at least a job in which they can take advantage of for much longer periods of time. One’s the process of discovering who’s breaking the system. The other one’s trying to figure out how to break the system. And it seems like most software is held together by toothpicks and glue and there is a lot of dangers in every piece.
Lex Fridman
(01:44:40)
And also the social engineering aspect, that’s a real attack vector. I think that’s the attack vector that will do in the longterm the most damage in the world. Especially as AI tooling becomes easier and easier to convince people at scale, sort of do that email Grandma. I think that’s a really serious attack vector, like human psychology and all that. I assume whenever there’s a girl that approaches me, it’s some kind of social engineering project, some attack vector, some intelligence agency. In fact, I’m pretty sure-
ThePrimeagen
(01:45:12)
We’re back to A Beautiful Mind, aren’t we?
Lex Fridman
(01:45:14)
Beautiful mind? Yeah. I have a whiteboard upstairs that I calculate everything, everybody’s trajectory and move.
ThePrimeagen
(01:45:20)
You’re not wrong though with the attack vector, especially in the day of AI. One thing that I don’t think a lot of people are talking about as we integrate more and more AI is that prompt injection is an extremely hard thing to defend against because it’s not really clear how you defend against it. If it’s just a, at the end of the day word calculator make word come out. If you can figure out the proper word calculator input, it might just break its bounds and start doing something it’s not supposed to do. And there’s a whole future word.

(01:45:49)
There’s all these products that are going to be vulnerable to things they never thought about. It’s one thing where you forget an edge case while you’re programming. Now you have to guess what people might be able to think of making something that has access to a system be able to do. Right. And you don’t have a way to reason about it. Its reasoning came from Reddit, and other words that it’s read and how to put things together. This is a very… It’s a massive space that’s going to be happening. It’s why I’m personally thinking don’t give too many powers yet. We don’t know the attacks that are about to happen.

Breaking production

Lex Fridman
(01:46:21)
Yeah, the more power we give to software systems, the more damage they can do. That certainly is the case. But the more awesome they could do, and that’s the knife’s edge that we all walk along as a human civilization together, hand in hand. Will we flourish or destroy ourselves? Question mark. Folks on Reddit, the good folks on Reddit, demanded that I ask you about the time you broke production. Is this related to Falco? Did you break production? Is this fake news?
ThePrimeagen
(01:46:48)
I’ve broke production quite a few times. I’ve broken productions for so many stupid reasons. One time I broke production because I came up in the PHP and PHP. Static means static for the lifetime of the PHP and PHP was the lifetime of every request, right? That’s why PHP was so inefficient was that every request was its own instance, and therefore static memory was for the lifetime. I guess I never put that together. And so I had some objects that I made static because I was like, oh I just need this for the lifetime of the request. And lo and behold, those weren’t lifetime. A whole bunch of bad data got all over the place. People were showing up saying they were from all these different countries and everything was all wrong because I just… “Whoopsie-daisies.”

(01:47:25)
I just made a whole conundrum with that. So that was one time I did it. Another time is I took down, if you were on the homepage on the website waiting for Lady Gaga’s video to come out and you were watching the countdown go down, if it reached zero, the billboard would freeze and it wouldn’t work. If you refreshed it, it would work. But the reveal, the big reveal, I screwed that up and my boss got real upset and so did other people in Hollywood got upset about that one. That was like a, “My bad. Sorry, Jeff Wagner, again.” I remember that one. I remember that one specifically. One time I released a bug where again on the billboard, if you pressed add to my list, I accidentally programmed in an infinite loop, and your whole webpage would just freeze.
Lex Fridman
(01:48:12)
Are some of these bugs difficult to discover until you started-
ThePrimeagen
(01:48:14)
That one seems really easy looking back at it.
Lex Fridman
(01:48:17)
In for a loop? Yeah.
ThePrimeagen
(01:48:18)
And we actually, during those days we had manual QA that are supposed to go through everything. So I didn’t feel as bad because my manual QA counterpart also missed it. We all missed it. But it was just so simple. Just press that button, boom. It just completely freezes the website.
Lex Fridman
(01:48:33)
Polluting the code with global variables that are holding values, SPHP I think allows you to do, that’s a tricky one to discover, because you rely on it, then there could be somebody else assigns a value to it.
ThePrimeagen
(01:48:47)
Yeah, it’s a data race everywhere. And I just didn’t understand… In my head static was like, “Oh, this is for the life.” I was just so locked into the PHP world at that time that I just made just such a, looking back on it’s so obvious. But during the time, it’s hard.

Pieter Levels

Lex Fridman
(01:49:04)
So in general, pushing to production, I talked to Peter levels about this. He, obvious he’s operating as mostly a solo developer, but he often on the website said thousands, not hundreds of thousands of people use. He often ships to production, pushes to production, meaning just no testing, just like push to fix. What are the pros and cons of that approach in general to you? What do you think?
ThePrimeagen
(01:49:31)
It’s obviously much easier the smaller your organization is. I think no one would argue that sentiment. If it’s just you working on a singular project, it is obviously much easier for you to push directly to production because you are the only one working. You know all the ins and outs and if something were to break, you would discover it. So to me that makes sense. I think the way he operates is perfect for what he does. You couldn’t take what he does and move it to say Microsoft or Netflix or Google because that would obviously… It would just be a disaster, just due to the amount of people all pushing to production. And so I personally love that. I think that you have to gauge both the application you’re building and its complexity and what you’re pushing, and how many people are working on it.

(01:50:15)
I think those all go into how you can do that. Because not all applications are created equal either. That application I was making was zooming and scrolling where we had all of our own everything. It was a very deep heavy logic app, and that was regardless of what was happening on the website, most of the code was library code. And that becomes way harder if you don’t have a good test suite and stuff to run before you push it out. Because when you squeeze that ball, different things come popping out in different areas. And that’s a very harder problem than say if you’re doing more of a heavy visual one because a heavy visual one, you’re affecting just this one area’s visual stuff and you can test it and that’s normally the end of it.

(01:50:57)
Whereas, you know… So it depends on the coupling and everything. So I love his approach by the way. I have such mad respect for anyone that operates that way because I think is a great way, it just is so good because it breaks this notion that tech Twitter has that oh well you have to use all these expensive services, you need to use all these things because if you don’t use all this stuff, “If you’re not using the latest version of React, if you’re not using the latest version of this, you’re simply not going to make it as a startup. It’s impossible.” And it’s just like, “No, no, that’s not software.” Most of software isn’t the new stuff. Most of software is old crappy software that someone has to maintain, and it actually is really, really great and has lots of really hard problems. And if you look at it differently, it’s actually fantastic.
Lex Fridman
(01:51:38)
For people who don’t know his tech stack, in terms of web development is PHP, jQuery and SQL.
ThePrimeagen
(01:51:45)
Yeah, all great stuff. I’m just surprised he still uses jQuery just given the fact that at this point on the modern web, everything is, you have document query selector and ad event listener click, right? It pretty much has everything you already need. It had DOM content load, all the reasons I used jQuery back in the day was adding a click on a button was hard. You had to deal with IE7, IE8, IE9. Those are hard differences. Whereas now, it’s just so easy. I’m just surprised it’s even that.
Lex Fridman
(01:52:13)
That’s definitely a trade-off. I still use the exact same stack, PHP, jQuery and different flavors of SQL. But the question there is you keep using jQuery because you can get the job done really fast and there’s no significant performance hit that you detect. So like why swish to something else? But it’s always probably as we’ll talk about good to explore and to learn.
ThePrimeagen
(01:52:39)
Not all tools are great at solving all problems. And so what you think is really the problem is you run into this trade off, which is you have some tool belt that you’re very adept with, you know all the ins and outs. There’s no unknown unknowns, but there’s no surprises in this. You know what you’re building, you know what you’re getting into. You will go through and you’ll be able to solve the problem. But if you ever use a different language or a different experience, you can find that some things are able to represent states way easier, in a way more efficient way. And you can solve problems really efficiently in some versus the other. And so it’s like if you don’t take the time to explore as well, you could be missing out on something that makes you twice as good on this one specific problem like subset. And so I value being able to look at all problems. And so I don’t want to get stuck on one thing though. I see why people do, which is for the efficiency sake.

Netflix, Twitch, and YouTube infrastructure

Lex Fridman
(01:53:34)
Let’s just return to the infrastructure of the platform of Netflix and, speak more generally, Netflix, Twitch, YouTube. Anytime I use any of these services, I’m just blown away by the infrastructure it takes to deliver this service. YouTube and Twitch are unique, versus Netflix where the creators can roll in themselves and upload stuff. So on the consumption side, YouTube has over 100 billion views a day, over one billion hours watch time. But on the creator side, one million hours of videos are uploaded every day. One million hours. It’s like you have to service both and you have to deliver everything… It’s just incredible to me. Can you maybe speak to your own intuition, just zooming out on it, what it takes to deliver that kind of infrastructure?
ThePrimeagen
(01:54:25)
For me, the thing that I find vastly complicated and I can’t imagine the engineering hours, is how do you even create an edge in that situation? And what I mean by an edge, when people say this phrase, if you’re unexperienced, an edge is where you deliver data. You want that edge to be as close to the customer as possible because that’s where the data lives. And then the communication between the customer and what you’re doing is really, really small. Obviously the speed of light adds up, the amount of hops adds up, the amount of services that you have to remotely call adds up.

(01:54:54)
They all add up and they all add inefficiencies to the system. So something like YouTube, they want to be able to serve that data as quick as possible, but their data changes constantly and relevance is almost directly tied with the newness of the item. So it’s like how do you even cache these things out? How are you doing this? So they must have such an incredible caching network that I can’t even… I can’t even fathom what it takes to do that. That just to me is just so impressive. A million view hours in how many different resolutions with how much data? What is a million view hours? Is it 4K million view hours, along with 1080p, along with 720p, along with 1440p? That number is an insane number.
Lex Fridman
(01:55:33)
Actually, it is brilliant what you said, which is for YouTube often the new thing is extremely important to show to everybody. And so, you can’t rely on caching or trivial kind of caching.
ThePrimeagen
(01:55:34)
Yeah.
Lex Fridman
(01:55:48)
You have to deliver the new thing as quickly as possible. Yeah, it’s incredible. So there’s the entire system, the recommendation system that knows each individual human watching YouTube and it has to integrate into that the new thing, while also caching this incredible cluster of possible videos that you’re potentially interested in. And integrate into that ads in the case of YouTube and so on.
ThePrimeagen
(01:56:19)
It’s a really tough problem because you have to think what is the cash hit rate on this? Because the problem now actually comes down to space, space actually becomes a real problem. How many hundreds of petabytes do they have that they have for like, “Okay, what do we cache and where do we cache this?” The number, I think in the terms of gigabytes or maybe megabytes, they have to think in probably versions of bytes I don’t even know the name for right? It’s such a different problem and that’s why I said Netflix. Netflix has a much easier job when it comes to caching. So if you’ve never looked it up, it’s called OCA and that we know what videos we’re releasing, we know what videos are hot in specific areas. It’s a very limited set. We’re not going to all of a sudden get, “Oopsies, we got a million new view hours.” We don’t even have to worry about that as a problem. So it’s like, “Okay, we know Stranger Things season five’s about to drop, we’re going to pre-cash Stranger Things season five in every single OCA across the world because that thing’s about to get hammered.” And so it’s like it’s able to do such a different decision-making than what you have to do with something like YouTube. And then Twitch is even more wild because now you’re actually ingesting video and trying to make it go out all at the exact same time for all video and you have to transform that video from whatever format and whatever the bit rate is into something that’s more efficient in the system like that. Hats off to Twitch engineering, because that’s some serious work.
Lex Fridman
(01:57:44)
And here’s some asshole, Lex, coming out and tweeting about YouTube features. Listen…
ThePrimeagen
(01:57:53)
You’re not wrong on the features you asked for, though.
Lex Fridman
(01:57:56)
I think this is an engineering problem of how do you allow fast iteration and addition of features that shouldn’t have to be integrated or impact the whole code base. So at the edges of the code base improve on certain features, without having to consult the mothership of the code.
ThePrimeagen
(01:58:19)
It’s the large team, right? That’s the fundamental problem. When you get into YouTube size, there is the team/organization that deals with data warehousing. There’s the team slash organization that deals with delivery. There’s a team slash organization that’s like the middle layer, how you even… They’re going to be like the little micro-surfaces to talk to these places. Then you have this front-end engine. So for a small feature, you have to get middle team, you have to get backend team, you have to get all these things. Quick example, Netflix. Are you familiar with the dystopian, Black Mirror?
Lex Fridman
(01:58:19)
Yeah.
ThePrimeagen
(01:58:51)
Okay. Season one, episode one. Do you know season one, episode one? Everyone who watches Black Mirror typically knows this episode.
Lex Fridman
(01:58:57)
Okay, yeah. I don’t remember what it is.
ThePrimeagen
(01:58:58)
Forgive my language, but they call it the pig-fucker episode.
Lex Fridman
(01:59:01)
Oh yeah, of course.
ThePrimeagen
(01:59:02)
Once you’ve seen the episode, you will then know this episode. Well, when Netflix adopted it, I got pulled into a room, there’s like a VP, a product designer, a VP, and they said, Hey, we’re about to release our own version of Black Mirror, season three, I think at that time. We need episode one, season one to not be the first thing people see. So let’s just reverse the season order. That required me… I had like 20 engineers I had to gather together to be able to have this happen. And that’s just the problem of big companies is that eventually every little thing has to become its own team. And so even small… There’s no such thing as a small feature.
Lex Fridman
(01:59:42)
Reversing the order of the dropdown that selects the seasons is a meeting with a bunch of VPs and engineers. That’s really interesting.
ThePrimeagen
(01:59:50)
Yeah.
Lex Fridman
(01:59:50)
There’s got to be a way to accelerate that. The natural scaling of a company and the bureaucracy that grows, yes, slows that down. But just having seen Elon work a lot, his teams are able to still keep it very fast, even as the company grows. There’s got to be a process to doing that, especially for the pig-fucker episode. I don’t know where that’s in the priority list, but for important things like that, you should be able to do that quickly. I don’t know. Can you speak to how would you do that?
ThePrimeagen
(02:00:25)
Well, I can tell first how it was done. Remember… So at a place like Netflix, there would be… I think that at that point it’s called a product called Dexter. I can’t remember. There’s our actual movie metadata warehouse that’s going to be highly integrated with Hollywood, where that side is able to manage all that. So I’m like, “Hey, you need the ability to mark things that need to be reversed because we’re going to run into this a bunch.” And we did. We ran into quite a few topical shows that all need to be reversed and all that. And so it’s like, “We need to be able to reverse episode numbers, season numbers. We need to be able to hide season or episode numbers.”

(02:00:58)
In the case of the Chelsea Handler Show, it was like a daily show, so it’s like you don don’t need episode numbers, you just need the latest one. And so there’s this whole problem that exists. And so it’s like, “Okay, you need to work on that for your UI over there, then you need to be able to store that data. Then we need to be able to go to the people that can actually get the video data out of that and provide it to our service layer. I need to go talk to them and convince them they need to be able to give me the new methods and everything to do that. Then I need to be able to go write the methods to get it down, and then I need to go to the UI and make that accessible. Now I need to go to the website people, I need to go to the mobile people, I need to go to the TV people.”
Lex Fridman
(02:01:31)
Yeah.
ThePrimeagen
(02:01:32)
And so it’s like, you could see this thing like snowballing. And for us, the big thing that Netflix did that was so well is after I met with these people that were high level, I was the captain. “I’m the captain now.” So I went to all these teams and said, “Hey manager, I need an engineer. We need to get this done within the next couple of months because we got Black Mirror coming out.” So she would go, “Okay, here you go.” “The map team, I need someone to help me with being able to get data out of the lomo for this.” And so it’s like, “All right, you’re working with this engineering.”

(02:01:59)
I’d go to the BMS team, “Okay, I need this engineer.” I’d go to the billboard team, “I need this engineer.” I go to all these little places to get all these little pieces of data. And then I was the captain, I was like, “You’re working on this, you’re doing this, you’re doing this, you’re doing this, I’m doing this. Let’s go.” It’s like that worked. And we were able to go pretty fast for a big company. And the fact that it required 20 engineers to do such a simple task, we were able to do it in, gosh, I’d say about three weeks worth of effort. But that was still… I thought that was amazing, comparatively, to how many people move.
Lex Fridman
(02:02:29)
Well, because you have the freedom of the agency to do it. You said the captain of the ship. That’s really powerful. For big companies, that’s a risk. Because you can fuck it up. You might not see the bigger context legally the bigger context of the impact on the industry or all the contracts that are made, all that. So it’s a risk, it’s a risk, but it’s a risk you have to keep taking. And then when you fuck up, you fix, and then maybe pay the cost legally for that, whatever. But the long term, that risk pays off because you’re going to keep creating a better and better product, evolving where the industry is going, constantly innovating ahead of where the industry is going and so on. Yeah.
ThePrimeagen
(02:03:08)
And not only that, I think one thing that is just so important is that yes, the product will get better, but the people that you hire and the people that you keep around are better because they’re the ones that show maturity. They’re the ones that can just… You give them something and they can rally the troops and make something happen. That’s a very great group of people to hire. And so you also naturally select out great engineers that aren’t just simply good at coding, they’re good at coding and they’re good at explaining and they’re good at convincing. And you have to create a very lean audience that can move fast.
Lex Fridman
(02:03:40)
And I think for great engineers, having to wait for like, “Okay, let’s schedule a meeting for next Wednesday with the VPs and…” That destroys their soul. And they either don’t want to contribute anymore or they leave the companies or they just tune out and take the golden handcuffs and just buy a nice house and focus on a family.
ThePrimeagen
(02:04:04)
Yeah. I feel like I would die under that… Honestly, that is my death sentence is where it’s just that there’s no reason to try, there’s no reason to do anything. I’m just going to go in there, effectively zombie through my day and call it… I don’t want to live like that. I want to feel like I’m trying to do something.
Lex Fridman
(02:04:22)
I should also mention on top of that, so you’ve brilliantly laid out how incredible the challenge that Netflix has to solve. On top of that with YouTube, the metadata thing, because users are able to upload video and there’s an API where they can upload automatically and change all this kind of stuff automatically. Every one of those things is an attack vector, as we mentioned. That’s something they have to consider seriously on the engineering side. And on the legal side, they can get into trouble all kinds of ways. So they have to consider all of that, which is fascinating.
ThePrimeagen
(02:04:59)
The legal side is obvious, but it’s not really like… I would never have initially thought someone would, say, upload images that you’re not allowed to own or have. But that guarantee you that happens. Then you have the whole kid side, right? I think about when you mark something as kid-friendly, how many times have they snuck porn into a Taylor Swift video or whatever? That was like a few years back, there was that whole Taylor Swift or whatever. I forget what it was, I thought it was Taylor Swift. But there’d be these mock videos that’d come up and then, boom. It’s like, that is such an awful problem and I’m so happy that is not a problem. I have to try to figure out.

ThePrimeagen origin story

Lex Fridman
(02:05:32)
Okay. So yes, YouTube and Twitch and Netflix are doing an incredible job. You eventually chose, the madman you are, to leave Netflix and to start on a new journey of being a wolf pack of one, start streaming. What was that? What was the story of that?
ThePrimeagen
(02:05:52)
So I was streaming for almost seven years now. It started actually at Netflix. We did a charity, Extra Life, shout out to Extra Life for starting my streaming career, effectively is just you stream and whatever money you raise, it goes to Kids with Cancer research. They are a great charity in the sense that they take no overhead and they raise their own donations for their website and everything. And so it’s a very great, straightforward charity. Really love what they’ve done. It was super cool because I live in South Dakota now, but I actually could choose a hospital directly where the money goes to. So there’s a direct impact from A to B. So it’s a pretty cool organization. And so my friend, Guy Cirino, Nice Try Guy is what I like to call him, he was probably the single greatest engineer I’ve ever met in my lifetime.

(02:06:37)
And he was just like, “Hey, come do this. We’re going to all do this.” So I played Fortnite. So before I did that, I was like, oh, I better learn how to stream first. I better get affiliated so I can take subscriptions. And then if anyone gives me a subscription, I’ll also pay that forward. So June 2018 or something like that, I start, I start streaming and I start streaming some Fortnite. End up getting affiliated, end up doing the whole extra life thing. I end up really enjoying it. I’m like, “This is a lot of fun.” I’m playing Fortnite at that point. Okay, so mind you, I’m a Fortnite streamer at that point, and I start really enjoying it. I keep doing it and then one day I decide I’m going to do some programming. Because I really love Vim and I think I’m fast at Vim, and maybe people think programming is cool. Because there was no really programming section at that point.

(02:07:22)
And I did it. I had like 30 people show up, which was just… And it felt like incredible numbers at that point. So I was like, “Oh my gosh, there’s like 30 people watching me program.” So it just kept on going and it kept on happening and it just kept on growing. And I did it for year after year. I would do my job, I would come home, I’d eat dinner with the kiddos, I would read them Lord of the Rings and the Hobbit during that time, I’d read to them for a half an hour, then I’d set that down. And then three nights a week I would program until like 2:00 in the morning or play video games until 2:00 in the morning streaming and building up this whole side thing. And I did this for a long, long time, and then eventually it just kept working out so well and I started making YouTube videos. And then that started getting better. And it was just a long, long grind until April of last year.

(02:08:08)
I went to the Streamer Awards and I got to announce the programming category and Pirate SoftwareOne. It was awesome. It was a great time. And during that time he gave me a challenge coin and just said, “You just got to go for it. Just go full time.” And so I just sat there and my wife can attest to it. It was like an emotional turmoil thing and it just took a lot of, it was pretty awful because Netflix is very safe option. It was both very fun. It was challenging. I liked a lot of the people worked with, it was overall a really great thing. I had a really great boss, really appreciated him. I still have text him now and then he’s really great guy. So it’s just like I’m leaving all these things for something that’s unsure. And the reality is that streaming and all these things, people love you one day, they could hate you the next day. There’s all this stuff that goes into being on the public side.

(02:09:04)
And I had Netflix as the backing, so it’s like if public hated me the next day I’d be like, deuces, I’m out. I don’t care now it’s like, now I’m going to do this as a job. And so there’s a whole huge turmoil to this whole thing that went through it. And eventually I just said, okay, I’m going to make this. It resonated with me when I first made the decision to join Netflix. I’m getting older. There’s not a lot of chances to do something unusual. Those chances go down constantly. As you get older, this might be the last crazy thing I get to do. Let’s just try it. So in April I went full-time and I guess I haven’t looked back. I’m only not even a year into doing this as a full-time gig. And it’s just been a lot of fun. And the biggest thing is just being able to really explore and do these things on stream where people really enjoy watching and engaging. It’s been a great, hard, fun, amazing, difficult experience.
Lex Fridman
(02:10:02)
It’s a really inspiring leap. It’s a really hard one to take for many reasons, like you outlined, but also the loneliness of it, I think it’s a pretty lonely pursuit.
ThePrimeagen
(02:10:14)
It is.
Lex Fridman
(02:10:15)
Just you and the camera and the audience and the ups and downs of that. And there’s not really a team.
ThePrimeagen
(02:10:21)
I do have one lucky thing I’d say that, my editor, Flip, shout out Flip.
Lex Fridman
(02:10:25)
Flip, shout out.
ThePrimeagen
(02:10:26)
He said it would mean the world to him if I said, “Shout out, Flip.”
Lex Fridman
(02:10:28)
I love you, Flip. We all love you.
ThePrimeagen
(02:10:31)
Oh, man. He had, as he would say, he had nothing going for him. He had a really hard growing up. A lot of rough life decisions have gone into his life and he’s crawling back out of it. And he just said, Hey, I’ll edit full-time for you. So I just said, “All right, like 50/50, whatever I make on YouTube, you get. We’re going to do this together.” And we did that for years, making $0 a month pretty much. And so it’s just like that was an incredible jump and now we get to work together. So I do get that one team aspect that I think is really nice. But it’s not like it was at Netflix where I could hear about stuff people are building, I don’t have a team, I don’t have product or cycles, I don’t have a manager that I have to try to make happy. It’s just like… It is very lonely. And I don’t think a lot of people realize how lonely it actually can be.
Lex Fridman
(02:11:19)
Yeah. So combine that loneliness with, in my case, I don’t know how many people attack you.
ThePrimeagen
(02:11:24)
I have a shockingly low amount of attack rate, I feel like.
Lex Fridman
(02:11:28)
Yeah, people generally… It’s sometimes fun sort of teasing, that kind of thing, but it’s mostly just really… You give so much love to the world and inspire so many people, even when you’re making fun of stuff, yeah. But with me taking the loneliness of it combined with just really intense attacks, it’s tough. It can be rough. Psychologically, really a tough journey. You miss working with a team, just from even a software engineering side, where you can share code or talk over code?
ThePrimeagen
(02:12:01)
Yeah.
Lex Fridman
(02:12:02)
Yeah, the collab…
Lex Fridman
(02:12:00)
Or talk over code or the collaborative aspect of it.
ThePrimeagen
(02:12:05)
Yeah, multiple things there. One, hey, we love you Lex, so don’t let the things get you down.
Lex Fridman
(02:12:12)
Thank you. Thank you. I love you too.
ThePrimeagen
(02:12:15)
Thank you. Hey, little bonding moment here going on. But one thing I really miss-
Lex Fridman
(02:12:20)
Not in a sexual way, just to be clear.
ThePrimeagen
(02:12:21)
The tension is a little tense.
Lex Fridman
(02:12:23)
I’m getting uncomfortable. Yeah. Anyway, team.
ThePrimeagen
(02:12:27)
It’s just the one thing I really miss is just, even when I hated how people did it, just seeing how other people solved things, it’s really amazing just the raw creative power so many people have, and just being like, oh, wow, I would’ve never done it this way. Crazy, right? Wow, this is awesome. And then you kind of internally process this and you’re like, oh, I now have a new little tool in my tool belt. Because at some point it’s really hard to find a mentor when you’re first, young and you’re just starting out programming. I mean, anyone with a couple years of experience will be not just a little bit better than you, but infinitely better than you. It feels like crazy how much better people are. And so you have to get mentors and you learn from people. And then as you get better, that amount of availability gets really small. And so it’s something that I really do miss is that forced hard problem solving together.
Lex Fridman
(02:13:19)
I think there’s also a skill to mining the wisdom from other people. I generally try to approach even junior people, young folks. It’s just mentally, at least for me, it works as a hack to assume they’re the smartest person in the world, way smarter than me. And so I take every single word they say as potential wisdom, and that helps me sort of mine for potential wisdom there. Because it’s so easy once you get older to judge, to be like, yeah, okay, okay. I’ve been through that. I remember feeling like that. I remember thinking that. That’s incorrect, whatever. But just kind of assume that I don’t know what the fuck I’m doing, and the other person is this sage. And in that kind of interaction, I think you could actually learn a lot. And my favorite interactions is when we both think that way. So from there, I think that’s a catalyst for a great collaboration and interaction.
ThePrimeagen
(02:14:19)
It just also makes everything much nicer. It really stinks to work with someone that’s combative and negative. I don’t mind combativeness if it’s like I’m trying to figure out what’s best to do right now, versus combativeness just because you’re a negative person and things have to be this one particular way, because if they’re not this one particular way, it’s the end of the world. And that’s actually really hard for me to work with.
Lex Fridman
(02:14:43)
What’s the origin story of ThePrimeagen name?
ThePrimeagen
(02:14:47)
The origin story of ThePrimeagen name was, are you familiar with a video game called Turok? Nintendo 64. Turok had Turok I and then Turok II. Turok II was a brutally hard game. This is back when first-person shooters, they would only give you a certain amount of health, and you had to go discover health and get that health. And you had to beat the whole game without effectively dying. That’s the first version right there. That’s like Turok I and Turok II.
Lex Fridman
(02:15:16)
Turok is a renowned first-person shooter video game series featuring dinosaurs, action, and sci-fi elements. The franchise has evolved significantly since its inception in 1997.
ThePrimeagen
(02:15:27)
There you go. So in 1998, there, you can see it right there.
Lex Fridman
(02:15:30)
Turok II, Seed of Evil followed in 1998 featuring larger levels, more challenging puzzles, and deadlier enemies.
ThePrimeagen
(02:15:37)
The notable difficulty, it was very, very, very difficult. So I spent, when I got it, it came in a black cartridge, not like your standard gray Nintendo 64, the black cartridge. Badass game. And I got it and I put it in and I played, and I played every day for 10 hours a day, for a month straight. And I beat it. And it was such an incredible, great experience. And the last leader of Turok II is called the Primeagen. And so when I was a kid, when you’re in fifth grade, that’s super cool, named after the bad guy. And so for a long time on any internet thing, like Grail Online that I mentioned earlier, the name was ThePrimeagen. It was great. And then I became an adult eventually, and it’s just like, okay, I’m an adult. My name’s Michael Paulson underscore. And that’s what I was on the internet for a long time was that. And I remember it was like 2017, 2018, somewhere in there.

(02:16:36)
I remember just how bad the tech world had kind of become. It was just like this super pretentious place, tons of dick measuring, just everything that just was the worst. Ken Wheeler got canceled over playing the Circle game. It was just like, it is so hard to describe to people that weren’t there, but it was just the worst place to be. Tech was extremely unfun. It was extremely awful. Everything was just so, it wasn’t academic because it was research. It was like we’re building the most sophisticated things, and this is for the smart people and everyone else is the dumb people. Don’t worry. We’ll design for you, dummy. We’ll show you how to make the perfect architecture. And I remember changing my Twitter handle because I got so upset and just went back to my video game name. I was like, I want things to be fun.

(02:17:24)
I want this to stop. And so when I started streaming tech, my goal became to destroy whatever that tech mentality was, because it includes nobody. Everyone thinks that they’re the smart people and they design for the dummies. And it’s just like, no, I want tech to be this place where people feel like they can be creative, and excited, and actually build something. And if you’re new, it’s okay to be dumb and ask dumb questions. Learn from your dumbness. No one’s expecting you to be smart. Pick whatever you want. Actually do something and have fun and build your crazy ideas. Oh, you’re going to reinvent the wheel, reinvent the wheel, understand what you’re doing, learn it really good, and interact and stuff. And it was just so different than what was out there. And the name… Arnold Schwarzenegger talks about this thing where, when he first started acting, his name was the thing that people hated.

(02:18:14)
As he once said, you have a strange voice, you have a strange body, and your name, your name’s unpronounceable. No one’s going to Schneitzinfinitzel, no one’s going to remember that. And he said, but now the name is the strong part. And for me, I’ve always felt akin to that, though my name’s not nearly as cool, nor am I as popular as Arnold, nor am I as tough or good-looking or successful. But nonetheless, it’s just the name represented this counterculture movement within myself, in which I just hated what was there and I wanted to defeat it. And so this has been the thing. And now people remember me so well because of how weird my name is. And so it’s just like for whatever reason, it became its own thing. And so that’s the… Now I would never change it, and back then I would never change it because it was my rage against the machine moment, if you will.
Lex Fridman
(02:19:01)
Yeah, I love that as a symbol of rage against the machine and the rage being fun.
ThePrimeagen
(02:19:07)
Yeah. I just want people to be creative and have fun again. It’s okay.
Lex Fridman
(02:19:11)
What about the mustache? It’s an epic mustache. It’s an epic stash. It has a life of its own. Is there an origin story or did you guys discover each other at some point? Or did it emerge from the darkness of the struggle that is your life, or where does it come from?
ThePrimeagen
(02:19:29)
Well, the original mustache is that it was no-shave November back before it became Movember. It was no-shave November back in the day. And after no-shave November, you had all this hair. And so what’s the natural thing you got to do? You got to sport a mustache for a day. So whenever I’d forget to not shave for a long time, and then I’d let it start growing out really big, I just go, oh, this is kind of funny. I’ll have a mustache. So one day when I was streaming, it’s just one of those times I just didn’t shave, and then I started just letting it go, and then I got kind of a beard, and then I just had a mustache. When I did it, people were just like, yeah, it’s mustache time. And I was just like, heck yeah, it feels like a lifestyle decision.

(02:20:07)
This is the fun times. And so all of a sudden it was just exciting to have a mustache. And I shaved it off and I was like, oh, okay. But then part of me is like there’s this weird energy that comes from just having a mustache. So I was like, I’m going back. Told my wife, forgive her. She was very not as thrilled about my decisions to have a mustache long-term, but I just decided to have it back and it was the right thing. It’s always been the energy that I had was the mustache. It was always been there. It just never was visible until later on it feels like.
Lex Fridman
(02:20:40)
Yeah, we’re chatting offline how one of the components of a successful relationship is sacrifice and your wife was willing to take the sacrifice of allowing you to have a mustache.
ThePrimeagen
(02:20:48)
I clearly was not willing to sacrifice not having one.

Learning programming languages

Lex Fridman
(02:20:53)
You do this incredible thing where you tried a bunch of different programming languages when you stream. You go all out on certain programming languages like Rust and then go and then try to pick a new one, but also are experimenting constantly. So maybe one question I could ask is about learning. What’s your approach to learning a new programming language, and maybe what’s your advice on learning a new programming language when you begin that journey?
ThePrimeagen
(02:21:26)
So I’ve kind of done a bunch of different ways to go through this learning process, and I’ve tried a lot of different ones. Something that is obviously successful is just start building something. Just put your hands on the keyboard, especially if you already know how to program. You’re like, okay, I’m now using Zig. How do I do a main function so I can just run the program? Okay, now I know how to build. Okay, how do I do an if statement? What does it look like? Okay, how do I declare my own functions? How do I do modules, right? You just kind of Google your way through it, if you will, to get to the end product and build something. It’s a great way to do things because I find that repetition, rote learning is obviously the best way to do this. You have to kind of go over it a bunch and you can definitely get out and build a lot of stuff with that. I like that initial kind of get used to things.

(02:22:12)
But on top of it, I find that, by doing that, you also fall into traps. You kind of Google and you try to solve a problem in the language based on all of your previous experience. And so you don’t have what makes that language special. You have what all the other languages make special. And so you end up not really being able to use it very effectively, but you can certainly kind of learn it and get kind of good at it. And so the second approach I’ve been doing lately, and this has been inspired by the creator of Ghosty, Mitchell Hashimoto, is to just start by reading the language reference, the whole thing. And so lately I’ve been just kind of going through and just reading the entire manual for these languages. Like Zig, I’m almost done with that one. It’s like eight to 10 hours of just sitting down reading, and I’ll whip out my computer and kind of practice a couple of the things from the actual docs, and that way I can learn all the things.

(02:23:01)
So then when I start building again, I’ll remember, okay, I know there’s a thing over here, let me go reread about it because now I have it indexed in my brain somewhere that will remember. And so I don’t think there’s a right or wrong way. I mean at the end of the day, the right way is always that you have to build something eventually. You cannot just read about it. You have to put your hands on the keyboard, you have to build something out. And then once you do that, that’s where you really discover what makes it painful or what makes it great. And if you don’t have the breadth of what the language offers, you just may make it painful by simply being bad at it.
Lex Fridman
(02:23:30)
Where exactly are you reading this-
ThePrimeagen
(02:23:35)
Language reference.
Lex Fridman
(02:23:38)
The language reference.
ThePrimeagen
(02:23:38)
So it just goes through every feature top to bottom, right?
Lex Fridman
(02:23:38)
That’s a lot. Yeah.
ThePrimeagen
(02:23:39)
Every way it’s described, all the different things. I think Zig’s is, it’s a decent size, but it’s not just simply read the words. You want to internalize each concept as well. So it takes a long time. So I’m a slow reader.
Lex Fridman
(02:23:51)
So you’re building, in AI terms, like a background model. I don’t think you can just start building once you’re done reading because you probably forgot how to do a for loop. You kind of forget the specifics. You just are building up the design choices, the set of features available, what are the strengths and weaknesses, all that kind of stuff. And then you start building. That’s really interesting. Probably not the thing you would recommend to a junior developer, somebody who’s just starting out at first.
ThePrimeagen
(02:24:24)
If you don’t know what an if statement is, that’s not a good way to learn. To me, the best way to learn that is really hands on the keyboard and building extremely simple things, and slowly growing in complexity. Because understanding what a class and methods and instances versus the blueprint, which is the class versus functions versus modules versus all that stuff. That just takes time to learn. And so that’s a completely different style of learning.
Lex Fridman
(02:24:46)
I wonder because for me, learning right now, AI is a huge help, but I already have a lot of experience. I wonder, if you’re starting from scratch, whether that’s a good idea. But I still think it’s probably a really good idea, but basically generate some code using AI and figure out what it’s doing by playing with different parts. Maybe can you comment on that aspect, like the use of AI as part of the learning process?
ThePrimeagen
(02:25:15)
This is where I have both the hopeful and the doomer take at the exact same time, and it’s the same thing with Google or Stack Overflow, it’s all the same kind of take, which is it’s just making things more democratized in some sense. I get to ask questions in probably the most personal possible way with my own voice, in my own words, and it’s able to produce out answers and hopefully help guide me. Now, regardless of just say the errors and the incorrectness of it, ultimately just using it as a learning tool and being able to just formulate and read answers in your own voice, I think is super powerful. And I think it’s super amazing. But the part that I think is going to be really difficult is that we don’t value remembering things anymore as a society. Since the internet came about, I can just look that up. I can just look that up. You don’t need to memorize your times tables.

(02:26:12)
You can just use a calculator. You can just do all that. I remember I just was sitting on the airplane and I watched someone do the world’s most simple addition and subtraction like 10 times on their phone. And why are you not just… You should already know, you should be able to do these things. And I realized that we kind of offload our brains, right? Oh, I don’t need to know these things because I can look them up. And that’s not a bad answer in some sense. I can understand that. I don’t need them to remember every last thing, but then it also makes me realize that you kind of develop this learned helplessness, that a new error comes up. I’ll just ask the AI.

(02:26:46)
AI says, oh, okay, I got to fix this line. I fix the line. You didn’t actually learn anything. You kind of just used it as a quick means to get something out and move on. And so you sacrifice knowledge for speed, which is a great thing in some… We have to make those trade-offs all the time in engineering. Sometimes you have to move fast at the sacrifice of knowledge, and I’m totally on board for that, but I worry that what we’ll create is an entire generation of incompetent programmers who can do some amount of things well, but anything that is unique, bespoke, or requires some extra like little elbow grease, might become very difficult. It might cause a whole chasm where juniors remain juniors forever.

(02:27:26)
And I don’t want to see that. I want to see people grow. I want to see people actually be able to take this as a craftsmanship thing. And so that’s both my hope and my worry is that AI think can do both really. If you could ask whatever question you want and you don’t have to rely on, say, a book to give you that exact answer. And if the book just said it wrong and you can’t understand it, it’s just like, sorry, you don’t get to learn what this is. Like recursion for me, I spent way too much time until someone gave me the right problem to understand recursion. You could imagine AI could have solved that for me way faster because it could have gave me the right problem and walked me through much better. But what happened if I just always have recursion solved by them and not actually learn it myself?
Lex Fridman
(02:28:03)
So if I ask AI to generate code to do a certain thing, actually a large percentage of time, most of what AI generates is going to be correct for me, but some percent of time it’s not, fundamentally not. And for me to recognize the difference between those two, I think it takes a lot of experience. I think to learn that skill of knowing, no, no, no, a different new out of the box solution is needed here than the one you’re providing. You’re missing the point. That’s a skill, and how do you learn that? You learn that by building from scratch. So both are probably really necessary.

(02:28:44)
But I think as a first step of learning how to program, it’s pretty nice to generate a function, to generate for loops and all that kind of stuff, and then just fuck with the different lines and modify them to try to adjust the behavior of the program, and from the way the behavior of the program adjusts or bugs are created, you learn about the syntax of the language, the behavior of the language, all that kind of stuff. So I think it’s a super powerful way to learn. But yeah, you need to also write from scratch.
ThePrimeagen
(02:29:17)
At some point you have to take off the training wheels, because I think what you’re really spotting is the difference between reading and writing code. I can read a lot of languages very well. I can see what’s happening. I can understand it, but I would not be very good at writing it. I can understand a lot of things about C++ and I can read it, but I’m just not that because I just don’t done it in so long. I can’t remember where all the semicolons, and colons and you do public and private, and how should you do naming conventions? All those things kind of add all together, and then you’re just like, oh, I’m really bad at writing it, though I can read it. And so there’s a skill gap chasm that exists between those two.

Best programming languages in 2025

Lex Fridman
(02:29:56)
All right. Well, let me talk about the various languages. The cheesy, ridiculous question of what’s the best programming language? Let’s say, what’s the best programming language that everybody should learn? Maybe let’s go with the top five. I’m going to pull up the Stack Overflow developer survey, because I think we have… You don’t like them?
ThePrimeagen
(02:30:21)
You got to remember, because I mean, you’re a data guy. You know about biases and data. What does Stack Overflow naturally bias towards?
Lex Fridman
(02:30:28)
Well, they have the different slices of professional developers, junior developers, they have different slices. Okay, what is the bias?
ThePrimeagen
(02:30:36)
I hear you, but who fills out a Stack Overflow survey? Someone who participates on Stack Overflow. Who’s participating on Stack Overflow? Largely very, very new people, and that one guy that loves answering questions. And so I’m not sure if Stack Overflow is a great place to get data. It could be a very biased set of data.
Lex Fridman
(02:30:52)
Is it really only new people?
ThePrimeagen
(02:30:55)
I mean that’s who’s using Stack Overflow.
Lex Fridman
(02:30:58)
All right. Most popular technologies. On this…
ThePrimeagen
(02:31:02)
JavaScript, HTML, Python, SQL.
Lex Fridman
(02:31:05)
SQL is the more general kind of… I’m sure they’re not doing the individual sort of flavors of SQL. By the way, pronounce SQL versus SQL?
ThePrimeagen
(02:31:15)
It’s squeal.
Lex Fridman
(02:31:16)
Squeal? You squeal?
ThePrimeagen
(02:31:19)
Squeal, I think is the correct way.
Lex Fridman
(02:31:19)
Squeal.
ThePrimeagen
(02:31:20)
I did SQL because I didn’t know the audience. I don’t know if they can handle the truth, which is its squeal.
Lex Fridman
(02:31:25)
The squeal of joy, squeal…
ThePrimeagen
(02:31:27)
Squeal light, my squeal, Postgres squeal.
Lex Fridman
(02:31:31)
By the way, I had a lot of joy from earlier saying pigfucker, for some reason.
ThePrimeagen
(02:31:34)
It’s such a [inaudible 02:31:36]. I mean, can you believe that was a real conversation that I had?
Lex Fridman
(02:31:38)
Yeah, that was. TypeScript, BAS, Java, C Sharp, C++, C-PHP.
ThePrimeagen
(02:31:45)
It largely kind of aligns with the world you’d expect, but Assembly, why is Assembly more popular than Ruby? Who’s writing just Assembly by… No one writes Assembly by hand other than maybe that one guy that’s developing TLS 1.3 and hand rolling a cryptography algorithm to be the fastest possible algorithm.
Lex Fridman
(02:32:02)
Yeah. Assembly is a weird one. Maybe people write it maybe in school, but even in school now for a operating systems course or something like that, or system engineering. I don’t know if they write Assembly anymore. I don’t think so anyway.
ThePrimeagen
(02:32:18)
And Swift and Ruby being less popular than Assembly seems ridiculous. But nonetheless, okay, so you get my ideas behind that, but as far as top five languages go, that’s probably too broad because you could just name so many. I think you should probably archetype it by what do you want to do? So if you want to get into game development, perhaps C Sharp, C++ could be good choices. Or JavaScript and doing Canvas games, I could see that also working. But you’re limited by doing JavaScript obviously, because you can’t do as much because the language is just not fast enough to do as much. So it’s like a good thing to remember. If you’re going to be doing backend stuff, if you want a job, if you’re looking for a job, maybe C Sharp slash Java, or JavaScript, or Go would be great choices. If you’re looking to do embedded, you probably want to do C, C++, like that would probably be a good choice. And so I think you have to first determine what do you really want to get out.
Lex Fridman
(02:33:14)
If you’re just curious about programming, which I talked to a lot of people who are, yeah, you can consider jobs, but basically their question is, okay, what’s the first language I should learn, and maybe what are the several languages I should explore?
ThePrimeagen
(02:33:28)
Can I say something that’s going to make a lot of people angry?
Lex Fridman
(02:33:30)
Yeah, sure.
ThePrimeagen
(02:33:30)
I think the first language people should learn if they have no idea about anything is JavaScript.
Lex Fridman
(02:33:35)
Yeah. Why would that make people angry?
ThePrimeagen
(02:33:37)
Oh, because people just, first off, I’m not supposed to say anything nice about JavaScript.
Lex Fridman
(02:33:41)
Yeah, usually that’s the meme, that you hate JavaScript, right?
ThePrimeagen
(02:33:44)
Yeah. No, JavaScript’s a beautiful language, and it has a lot of things that are very great for it, and one of them is that you can express anything with very little effort. And so someone that’s new, I think it’s really great to be able to draw a box and move a box. That’s great. You get to see it visually. I think that’s one thing that’s really great about JavaScript is that you can do that. Then you can go, okay, I want to learn about the backend. I want to make a request now.

(02:34:08)
You can write a quick backend in it. Now you’re starting to get familiar with programming a little bit. I can save this to a database. I can bring it down. I can put it on a screen and I can animate it all around, and I can even put it on a canvas and render it in 2D or 3D. So it’s like there’s so much variety of what you can do with JavaScript. It’s a great way to get introduced into programming. But then at some point you have to go, okay, I now need to learn more about this whole thing.
Lex Fridman
(02:34:31)
I mean, yeah, just like you said, you can make games, you can do front end, backend for web development.
ThePrimeagen
(02:34:37)
You can even do embedded. They actually have… Like there’s Wes Bos is building his Roomba or something and programming it with JavaScript and React, which is just the world’s worst language to choose for embed, but you can still do it.

Python

Lex Fridman
(02:34:51)
Also, we mentioned sort of in terms of applications, anything that relates to data or machine learning, Python is the sort of the leader there, so that’s a great one.
ThePrimeagen
(02:35:01)
It seems like Python, CUDA stuff and C++ would be a dynamite in that, because a lot of these Python libraries are assumed you’re just smuggling in C++ underneath the hood or C.
Lex Fridman
(02:35:11)
Okay, so JavaScript. I’ll say Python.
ThePrimeagen
(02:35:14)
Python’s a great one too. You can get quite far with it, but you can’t write the front end. What happened if you love the front end? What happened if you really just want to design things and you just didn’t know that?
Lex Fridman
(02:35:24)
Well, it’s okay. So for that, JavaScript.
ThePrimeagen
(02:35:26)
But Python’s a good choice because you can’t do the ML stuff in JavaScript nearly as easy.

HTML & CSS

Lex Fridman
(02:35:30)
Do we count HTML and CSS as programming languages?
ThePrimeagen
(02:35:33)
I think there’s some technical definition that it is. If you use this certain amalgamation of CSS plus HTML, it actually has, it can be a Turing complete language. But I mean for practical purposes, no. HTML is not a language. For me, yes, the Turing test is a good one, but for those that are just not wanting to be as academic, if I can’t write a function and an if statement, I don’t feel like that’s… If I can’t loop, if, and function, I don’t feel like that’s a good, that’s a programming language.
Lex Fridman
(02:36:03)
Although modern HTML has a lot of features.
ThePrimeagen
(02:36:05)
It’s crazy how much it has, but it’s more of a specification than anything else. I specify it to be a pop-up. I specify it to have this kind of accessibility, this kind of look. Under these conditions look like this, transform like this, move down here.

Bash

Lex Fridman
(02:36:21)
I don’t know. I kind of like these popular programming languages in this list. I like JavaScript.
ThePrimeagen
(02:36:25)
You like Bash?
Lex Fridman
(02:36:26)
Oh, yeah. I like Bash a lot. Yeah. Why?
ThePrimeagen
(02:36:28)
Okay. Bash is kind of one of those ones where it’s like, do you really like it? I like it up until I need an array.
Lex Fridman
(02:36:34)
Oh, as a programming language, just no, but I like the command line.
ThePrimeagen
(02:36:35)
Okay.
Lex Fridman
(02:36:39)
Do you like Bash? No, nobody likes Bash. Do you mean-
ThePrimeagen
(02:36:39)
Someone is so offended right now.
Lex Fridman
(02:36:45)
It means do you use it a lot? Yes. I mean, it’s good to learn, right? It is good to-
ThePrimeagen
(02:36:55)
It is.
Lex Fridman
(02:36:55)
… Be comfortable in the command line because it’s a bit of a superpower. It’s like, I think I follow on Twitter, FFMPEG.

FFmpeg

ThePrimeagen
(02:37:02)
Great account.
Lex Fridman
(02:37:05)
There’s certain Twitter accounts that are just legit. And I think FFMPEG, they have all these sort of parameters that you can add on the command line, that it’s like one of those cryptic languages that only very few wizards understand. But once you begin to slowly understand, and I’m only at the very sort of beginning stage of that journey to mastery, the powers you gain at every step, it grows exponentially, it feels like. I mean, FFMPEG is just this incredible, what would you call a library system? There’s just the people behind them must be just brilliant masterminds because they have to work with all these codecs, with all these containers, with all the mysteries of the media codec universe they’re masters of. And they understand compression, which is another super fascinating technical set of problems that, I don’t know, just FFMPEG just fills me with joy that it exists. But you need kind of Bash type comfort, command line comfort, to work with it to really unlock its power. Yeah.
ThePrimeagen
(02:38:12)
I think FFMPEG is probably one of the most consequential libraries of our day, and the Twitter account is so unhinged. It is the most amazing thing to see because I think FFMPEG does not get the love it deserves. Every single application, OBS, probably FFMPEG underneath the hood. Everything, FFMPEG underneath the hood, and yet they do not get the love they deserve. I just love it. I just think they’re the best.
Lex Fridman
(02:38:39)
Yeah, I would say JavaScript, HTML, CSS, Python, SQL, I mean that is SQL Squeal is a programming language. It’s an incredibly sophisticated programming language. Yeah?
ThePrimeagen
(02:38:52)
SQL is interesting. I believe you can classify it as a programming language. It does have, if. You have case statements and it’s pretty crazy what you can do with it.
Lex Fridman
(02:39:00)
You could do functions, you can do all that stuff. You shouldn’t.
ThePrimeagen
(02:39:02)
Yeah, for stored procedures, that’s how you make your life hell. I will say that all the top languages there, none of them are strict static typed languages. And so even TypeScript, I don’t like this any. And so for people that are learning, doing something that’s much more strict would be great. Something like Go, Rust, I mean even C Sharp, C++. Anything that kind of changes your perspective of types I think is really helpful to kind of go through. They’re not getting nearly as much love on this most popular language list, but I think they’re very fantastic.
Lex Fridman
(02:39:38)
All right, well, if I put a gun to your head, top five languages, let’s list them out. There’s a bright-eyed 20-year-old asking you, what are the top languages, five languages to learn?
ThePrimeagen
(02:39:51)
If I were to pick five languages that I think people should learn, or at least, let’s restate it this way, I’m going to say a couple languages and you should at least explore some of them. I think you should explore a loosey language, so Python slash JavaScript, where there is truly only one type, which is a boxed value, which is a multivariate, different types underneath the hood.
Lex Fridman
(02:40:12)
What did you call it? A loosey language?
ThePrimeagen
(02:40:13)
A loosey-goosey language. It’s a dynamic language, and so I think it’s really good to explore one of those too. So I’d put Python or JavaScript right there. Even Lua, throw Lua in the bunch. I think you should explore a strict language, so I’d do something like Rust, Go. I think those are both really, really great.
Lex Fridman
(02:40:31)
C++?
ThePrimeagen
(02:40:32)
You can do C++. You can do some type erasure in C++. You can do it with Go as well, but for the most part it’s a great language to do that in. It can get a little wild. New C++ seems great. Everyone keeps telling me new C++ is great. It has every feature you’ve ever wanted and all the features you don’t want.
Lex Fridman
(02:40:47)
Yeah, exactly. I mean there’s smart pointers, there’s dumb pointers, there’s all kinds of pointers. There’s no memory leaks. That’s not an issue.
ThePrimeagen
(02:40:55)
Foot guns, face guns, soft beds. There’s everything in there.
Lex Fridman
(02:40:58)
Unless you like memory leaks, it has that too if you want that kind of thing. It’s great.
ThePrimeagen
(02:41:02)
Okay. How about this one? Languages that I actually want to really learn, that at least sit in my curiosity bank. There’s three languages, which is going to be Swift, Elixir, OCaml, and then I’m going to throw Odin in there, just because Ginger Bill is great. But Elixir and OCaml, I don’t have a strong functional language underneath my belt. That’s something that I just genuinely lack.
Lex Fridman
(02:41:26)
Yeah, I’ve heard incredible things about Elixir, about Odin, about OCaml. Obviously, I’m a person, as you know, who loves Lisp.
ThePrimeagen
(02:41:33)
I have never done Lisp. Lisp could be in that category too, just, or Closure I think at this point is what everyone tells you to use.
Lex Fridman
(02:41:39)
So in the case of Lisp, I don’t want to speak negatively about Lisp, but it’s important about modern community, what the community looks like. It seems like there’s an excited, maybe small, but an excited community around Elixir, Odin, and OCaml, so that helps. Because then you can post shit on Twitter that you’re like, I accomplished this. People get excited and it’s nice. It’s a good feeling.
ThePrimeagen
(02:42:01)
You can post something on Twitter and you’ll get a thousand likes if you do something cool on Elixir, which is that’s a pretty big amount of people to like a post for such a niche topic. Programming’s already a pretty small topic. Then you get into functional program. That’s a small topic in a small topic.
Lex Fridman
(02:42:17)
Yeah. I don’t get that much. If I post something about Emacs, I’ll get crickets if I post something about… If I proudly use Neovim, there’d be a lot of people like, good job.
ThePrimeagen
(02:42:27)
Because it is the best editor.
Lex Fridman
(02:42:29)
Yeah, maybe it’s just hype.
ThePrimeagen
(02:42:31)
Come back to the Civil War, Lex.
Lex Fridman
(02:42:33)
Yeah. Sometimes you have to sacrifice and go from the superior editor that is Emacs and choose Neovim just to be popular. You sacrifice integrity and values and quality for just popularity. It’s a choice you made.
ThePrimeagen
(02:42:45)
Absolutely. I love how you put it.
Lex Fridman
(02:42:49)
Okay. Anyway, what were we talking about? I like how you’re doing this in bunches. That’s great.
ThePrimeagen
(02:42:53)
Right now, my kind of side honeys that I’m exploring-
Lex Fridman
(02:42:57)
Side honey?
ThePrimeagen
(02:42:57)
Yeah, side honeys. They’re not my mainstay. Right now Go’s kind of my favorite one to build a web app in. If I’m going to build some sort of backend with a lot of complicated logic, Go’s just so convenient. But I get really frustrated with its ability to express everything that I need. If you have a list, a heterogeneous list, a list that contains two types, Go’s just really not that fun to use. And I could see, so the ones I’m exploring is Jai or J, or the language as Jonathan Blow says, and Zig. And both of them have a lot of power to them. They’re both very interesting. They definitely have foot guns in them. They’re definitely more, they don’t take it easy on you. Zig seems like it’s a really amazing language, and so does Jai. They’re both very cool.

Performance

Lex Fridman
(02:43:42)
Yeah. Actually, I saw Dave Plummer’s testing of close to 100 languages for speed, and Zig came out on top.
ThePrimeagen
(02:43:51)
That was a mistake. I mean, when I say mistake, nothing against Dave Plummer. He’s an extremely talented engineer. It’s just that Zig, C, C++, all those languages that were being tested, they’re all LLVM backends, right? That’s the one that actually turns the thing into the executable part. And if there’s a variation in speed, it just means in one language you didn’t quite express what you are supposed to correctly. There’s the language ball test that’s been bouncing around on Twitter. Zig was like sixth or seventh below I forget what language it is. I played around with the example, added the word “no alias” to the argument, which means that the piece of memory that’s coming into this function, there’s no global pointers, there’s nothing to it, and so the compiler can make these really cool optimizations. And I made it faster than the C version. So it just means that it’s just not correctly specified is all that means.
Lex Fridman
(02:44:40)
Yeah, but it’s still exciting. To me, the competition between Zig, Rust, and C++ is really interesting. Part of it’s for speed. Part of it’s how easy it is to write performant code.
ThePrimeagen
(02:44:51)
I’ll say something. That’s the reason why I think Zig is so interesting comparatively to say C or Rust. C is the ultimate language. It can do anything, have pre-process or macros. You can do quite a bit with it, but it’s also really difficult.
ThePrimeagen
(02:45:00)
So macros, you can do quite a bit with it, but it’s also really difficult and it’s also really simple and you can learn it. So it’s kind of its own unique beast. And when you get really good at C, C is a magical language and people are really great at it and people speak very highly of it. Rust is like this ultra safe language. What you can do in C, you just can’t even express in Rust. Rust is going to be the safe man that holds you at night, keeping you warm, right? It’s going to be just the greatest.

(02:45:24)
But somewhere in the middle lies Zig. Zig has optionals. If you’re not familiar with optionals, that just simply means there’s a value here or there’s not, but you first have to check that before you can use it. So it prevents that whole null pointer dereferencing segfault problem. And that’s not available in C, just by default, you have to kind of build that thing in. It is the only option in Rust, but Zig says, “Hey, if you have a pointer, you can’t express it as null unless if you market that it can be null.”

(02:45:48)
There’s ways around it, there’s other types of pointers and stuff like that that can do that. But for the most part, Zig will give you safety for the most part. So it’s like a little bit of safety, but more like C. So it kind of gives you everything you want in that region where you can express safe code and unsafe code. It’s very easy to write. It’s very pretty. Or at least the idea behind it is very pretty. The language itself is bland, but.
Lex Fridman
(02:46:12)
Wow, there’s beauty in everything, Prime.
ThePrimeagen
(02:46:14)
Yeah.

Rust

Lex Fridman
(02:46:15)
You’ve programmed in Rust a lot. What do you love about Rust? What are the strengths? What are the weaknesses? Maybe you can speak about memory management that you already mentioned, the challenge of memory management that several of these languages address, but yeah, what do you love about Rust?
ThePrimeagen
(02:46:31)
What I love about Rust, I love the ability to free the memory that you’re using is directly tied to the stack. So whenever you create something, there’s a stack variable or there’s some amount of stack memory, whether it’s a pointer off to the heap, a pointer and a length. So some amount of memory on the stack and then some memory on the heap because a string is not all on the stack, it’s some on the heap, some on the stack. And when that stack variable goes out of scope and gets cleaned up, it also cleans up what’s on the heap. So it kind of simplifies this whole idea of, whoops, I forgot to free my memory. It just does it for you.

(02:47:07)
So it’s not a garbage collector, which will do it sometime later. It’s not like C where you have to call it yourself, it’s somewhere in between. Now, there’s a lot of strategies people use, arenas and all that that make that C part much easier. I’m just not even mentioning it, but it just makes it a lot easier. But Rust does that really beautifully and it’s just like a really cool idea about it and I really like that. And the second thing that I think Rust does really, is such a good thing is that mutability of something is you have to specify it. So you don’t just create a variable and then mutate it. You have to say this is not only a variable, it’s a mutable variable. And I think that just makes code really readable and really understandable. Because anything that does not have the word mute next to it, you know for a fact it cannot change. There’s some rules around that, but you get the general idea.
Lex Fridman
(02:47:57)
Unlike most programming languages, you have to explicitly state that this is going to be changed.
ThePrimeagen
(02:47:57)
Yeah.
Lex Fridman
(02:48:03)
Yeah. That’s really interesting. I mean it’s safe, it’s trying to be, and the safety might be, create limitations. Let us consult the AI overlords. Rust is a blazing fast memory efficient systems programming language that emphasizes performance, type safety and concurrency. The language enforces memory safety without using a garbage collector, as you said, instead utilizing the unique, quote, “borrow checker” that tracks object lifetimes at compile time. This prevents common programming errors like null point or dereferencing and memory leaks and so on. Yeah.

(02:48:43)
So you’ve also spoken about metaprogramming. Which of these languages do you like for the metaprogramming? I love metaprogramming in C++, but it’s a giant mess. At least when I program C++ 17 standard, I believe, it’s just a mess, especially a mess to debug.
ThePrimeagen
(02:49:00)
Yeah, I would consider myself kind of a metaprogramming newbie. I have only solved some amount of problems with it. That’s kind of like what this year is for, is for me to really, I want to see where the ends can go in that. So I don’t have a strong opinion on this one. Zig, one thing I really like about Zig is that the metaprogramming is also the language itself. So you don’t have to, there’s not an alternative. So with Rust there’s an alternative. When you create a macro, you have to do the macro syntax. With Zig, it’s just, it is the thing, you just program it and you add the word comp time if you want it to be a compile time only.

(02:49:33)
So you can create the list of prime numbers at compile time in Zig, which is kind of an interesting, unique thing. So you have code that executes at compile time and then you can take advantage of the result of it at runtime. So neat, right? That’s how I’d look at it. But again, I haven’t used it to the point where I feel like I can super authoritatively talk about it.
Lex Fridman
(02:49:55)
You have been undecided, what language are you going for this year?
ThePrimeagen
(02:49:59)
I’m going to keep Go as my mainstay, my two side honeys, Jai and Zig. I’m going to explore and try to build out a service in them that can do a bunch of talking to, say, ChatGPT and ElevenLabs and send stuff down to client and work with web sockets. And I want to make sure that, I just want to see how do they perform in this realm. And I may be using the language incorrectly, like Jai, it’s not really been designed for the web world. I just got done writing the ability to read Twitch Chat and it required me to do Berkeley sockets.

(02:50:29)
So if you’re unfamiliar with Berkeley sockets, it’s like the old way of doing it, it’s how you do it in C. So you have to kind of go through the whole nine yards of creating your own connection. I had to create my own connection, I have to read from the socket, then I have to parse out all the IRC, right? You have to kind of build it from scratch. There’s not like a new TCP connection to this server. You have to be like, “I’m creating a socket.” You’re going to be of the IPv4 family and TCP and you’re going to do, you know, I’m going to now have to take your address and go look up your address with DNS, get that address back and then connect to a TCP. So it’s a lot more manual still. It’s a lot more raw in that area, but it’s fun.

Epic projects

Lex Fridman
(02:51:03)
What are some epic projects you’ve built on stream that jump to memory?
ThePrimeagen
(02:51:08)
My most favorite, sorry for interrupting you. So I’m really jazzed right now.
Lex Fridman
(02:51:13)
Let’s go.
ThePrimeagen
(02:51:14)
Okay. So jazzed.
Lex Fridman
(02:51:15)
Jazz hands.
ThePrimeagen
(02:51:17)
My most favorite project was the one I did last year. Someone built a Doom ASCII port. So you could play Doom with ASCII. So that means you could play it in your terminal. Very, very fun, very excite. So I made a Go program that could spawn out the Doom ASCII, then I took that Doom ASCII and I sent it to the browser so that people could play Doom ASCII in the browser, but then I made it so that Twitch chat could control that instance of Doom ASCII by piping in Twitch chat, taking the average of the movements over so much time and replaying it as if it was a controller. And I had Twitch chat beat level one by spamming it.

(02:51:55)
But the fun part was I used a bunch of fun encoding techniques. I used quad trees to be able to take smaller amounts to use run length in coding. Tried to create my own compression algorithm because if you’re sending out a bunch of ASCII stuff, it’s still pretty expensive because you have to represent color, color’s not cheap. On top of it you have to represent what does it look like? What does the ASCII look like?

(02:52:15)
Well, I realized there’s all these fun techniques you can do for compression like the shape of the ASCII you send down in a lot of these engines are actually just proportional to the lumosity of that pixel. So you’d use an eight to represent or a pound sign to represent white, but black, you’re going to want to do a period or a comma or a bar, something smaller. So it’s like I then developed all these different compression algorithms that turn a bunch of data, which would take, I forget how much it would take. It’d take gigabytes upon gigabytes to be able to send out to thousands of people to all see the same image at the same time, to all be able to interact with Doom at the same time. I turned it from gigabytes into kilobytes by just trying to figure out how to make it as small as possible and send it all out. It was super fun. Absolutely had a great time.
Lex Fridman
(02:53:00)
So you’re actually sending it to all the people in chat. So where’s that pipeline, how chat is able to control the Doom thing?
ThePrimeagen
(02:53:09)
Twitch chat. Yeah, so they would go, people would spam W and if you said W, it would hold down W for 150 milliseconds if the majority of people during that time period said W.
Lex Fridman
(02:53:21)
Nice. Okay. And how are they getting the input of where you are on screen?
ThePrimeagen
(02:53:26)
So originally I was going to send that through Twitch, but Twitch is like five seconds behind, so that’s why I piped it out to a website so everybody could see from my computer to the website and typical lag was right around 70 milliseconds. So it’s like they could mostly see what was happening in that short period of time. It was pretty exciting. So we had 1,000 people, or I had somewhere between 1,000 to 1,400 people smashing Ws and pressing F to fire and turning and we killed some zombies. We blew up the barrel at the very end of level one to kill the imp.
Lex Fridman
(02:53:57)
How are you getting the Ws from the Twitch chat? Is there an API?
ThePrimeagen
(02:54:03)
I was using IRC, so just a little TCP socket and then you just parse out IRC.
Lex Fridman
(02:54:06)
Okay. And there’s very little lag there. Okay.
ThePrimeagen
(02:54:08)
Yeah, I think it’s a couple hundred milliseconds though. It’s enough that it actually made it a little bit difficult because people would often overturn and then go forward and miss the door and then they had to go back and…
Lex Fridman
(02:54:19)
That’s awesome.
ThePrimeagen
(02:54:20)
It was awesome. So that was my favorite I think project of all time just because I never got to do a lot of encoding. Encoding’s kind of like, what do you normally do? Okay, I need to send something down. I don’t know, gzip it, server will just do it. Server just does the right thing. I don’t need to think about it. So instead it’s like I think about it, I’m going to send the right thing.
Lex Fridman
(02:54:37)
Yeah, you have to think about the compression. Yeah. And there you go. That’s some more love towards FFmpeg because they have to think about that a lot.
ThePrimeagen
(02:54:44)
Ultimately inspired by FFmpeg and their awesomeness.
Lex Fridman
(02:54:49)
So can you speak to just the chat community in general? A big part of what you do in terms of streaming is the humans that are communicating with you live. Can you talk to the different chat communities? First of all, which is the best chat community, YouTube, Twitch, or X?
ThePrimeagen
(02:55:11)
This is where I feel bad for YouTube, because I do think it’s technically the worst, but it’s not YouTube’s fault. And let me kind of explain why.
Lex Fridman
(02:55:19)
And then I will explain why you’re wrong. But go ahead. YouTube is great.
ThePrimeagen
(02:55:22)
I know you love YouTube but let explain why, is that when you go on Twitch, you go to anyone’s channel, they have this cultural human centipede thing that’s happening where as the memes flow in, all of Twitch reacts and morphs to all those memes. So every channel you go to has this same culture. There’s a lot of similar emotes and everything, so it’s very tight-knit. So when I stream, I get all the same jokes that you would pretty much see if you saw, I don’t know, Sodapoppin or some big streamer, Asmongold, whoever, [inaudible 02:55:56] software streaming. All the same memes would all flow through the exact same kind of pipe. And so it’s a very holistic kind of community.

(02:56:04)
So every time you’re making jokes, you’re making jokes that are in the ether. Twitter kind of has that too. Tech Twitter kind of has a set of jokes and so you can kind of see it. The problem with Twitter chat is that there’s just nobody there right now. Typically just to put it into perspective, I have somewhere between… somewhere between like 1,500 to 3,000 people on Twitch, somewhere between 800 to 2,000 on YouTube, and like 50 people on Twitter. So the difference is massive.

(02:56:34)
But Twitter has that same thing that’s developing where there’s memes that are constantly flowing through it. And so they’re very highly connected. YouTube just doesn’t seem to have that. They’re just a bunch of people and people go to YouTube for various reasons. I’m going to YouTube to learn. So they come in and they want to learn. So they’re not on the meme train, they’re not in this cultural zeitgeist train. They’re just like, “But why would you use this if statement when a switch statement in this one particular case?” And you’re just like, well that’s not what I’m trying to do here.
Lex Fridman
(02:57:02)
Yeah, you want to captain the meme train or you want to ride on the meme train.
ThePrimeagen
(02:57:07)
Yeah. Or you just want to be able to create a culture on your chat because your chat’s going to be some variation of that kind of zeitgeist that’s flowing through Twitch. And it kind of is very contiguous between X and Twitch. It just feels really out of sync with YouTube. And then YouTube particularly does a bad job. And some people would argue a good job because you can swim. Swim being, you can actually change what timestamp you’re at. So all of a sudden you’ll be like, oh yeah, something about driving to soccer in my minivan. And then 20 minutes later you’ll be talking about Zig and then someone’s like, “I personally use whatever to drive to soccer.” And you’re like, “What are we talking about?” So YouTube is a very disjointed chat as well because it depends on where they’re at within the video. Swim comes from Netflix, by the way, call it swim.
Lex Fridman
(02:57:53)
The term?
ThePrimeagen
(02:57:56)
Yeah, people said swim.
Lex Fridman
(02:57:57)
Oh, so you’re, okay.
ThePrimeagen
(02:57:59)
Swimming through-
Lex Fridman
(02:58:00)
Yeah. So you’re not just making up the term. Thank you. Wow.
ThePrimeagen
(02:58:02)
Yeah, but it’s probably made up and probably only 10 people said it at Netflix, so no one’s going to know it and they’re going to be like, “Yeah, right. That doesn’t happen on Netflix.”
Lex Fridman
(02:58:11)
So going back to projects, what projects on stream or in general?
ThePrimeagen
(02:58:14)
No, you need to answer why YouTube chat’s the best chat.
Lex Fridman
(02:58:18)
Well you kind of convinced me. Okay, why YouTube is the best chat. Well, I think I’m just a hater. That’s basically what it boils down to and I’m just talking shit.
ThePrimeagen
(02:58:18)
Love it.
Lex Fridman
(02:58:29)
And I’m probably just from the outside shooting in because Twitch is such a fun culture of memes. And so it’s just fun to shoot from the outside to egg the house of Twitch. And then I just sit back on my lawn chair with the small YouTube community just talking shit. No, you’re absolutely right. There’s a real sense of community that Twitch can form. But I just like the openness of YouTube. It’s just better at opening to the world. It’s more accessible, it’s easier to share. It’s just a more established platform, that’s all.
ThePrimeagen
(02:59:08)
Fully on that team.
Lex Fridman
(02:59:14)
For the open world. I can send it to people that don’t usually watch video game streaming or that kind of stuff.
ThePrimeagen
(02:59:19)
Yeah. If you send a Twitch link, they’re like, “I don’t like video games.” And you’re like, “Well actually it’s not video games.” That talk happens every single time you mention Twitch because Twitch does have a perspective about it that YouTube does not.
Lex Fridman
(02:59:32)
I was just on Joe Rogan’s podcast and I think it came up, he asked something like, “Is Twitch still a thing?” So that just gives you an example. And then Jamie said, “Yeah, yeah, it’s definitely still a thing. It’s still growing and so on.” And so yeah, there’s just a big slice of humans that don’t participate in the Twitch Twitch sphere. Yeah, I just like talking shit so yeah.
ThePrimeagen
(03:00:00)
That’s a beautiful answer.
Lex Fridman
(03:00:01)
But it’s cool that you sort of make it accessible on all these different platforms. And I have high hopes for X, but yeah, it’s feature-wise still has a lot of growing up to do.
ThePrimeagen
(03:00:11)
And just why do people use X? You typically are going there for a text-based interaction you want to look through. So I also think they just have a user expectation change that needs to happen. And that just takes a while. That’s going to take a little bit before people get to it. I think their idea of audio first is a great first step where people can listen to it and have the phone away maybe. There’s a lot of changes that have to happen before X can be successful in that.
Lex Fridman
(03:00:37)
I mean, X has this incredible comment section just like Reddit, right? So it’s like-
ThePrimeagen
(03:00:43)
You said incredible. That’s not Reddit. Comment section, correct.
Lex Fridman
(03:00:46)
Comment, yeah. Incredibly dynamic and vibrant even if it’s… Yeah. What is the technological platform? How does the interface and the technology shape the discourse? It’s fascinating because X has a different style than Reddit, different style than Facebook, different style than Instagram. It’s interesting. And all those common sections are different technologically, like how the sorting is done, how easy it is to sort of build a community around it? Because YouTube is not really a community. Every single video on YouTube has its own mini community. You’re all talking on just that one video. But you can’t jump across.
ThePrimeagen
(03:01:40)
There’s not like, “Hey Bill, hey George.” There’s no crosstalk that happens in multiple videos.
Lex Fridman
(03:01:44)
Yeah. But the community is awesome. I love community. I love the feeling of community and I guess that’s what Twitch really provides.
ThePrimeagen
(03:01:52)
YouTube also does have it though. They have an aggregate community. There’s a lot of fun comments and all that on the videos and a lot of thumbs up and then you see the fun discourse that happens and it’s like that’s the community, it’s just only a certain slice sees it.
Lex Fridman
(03:02:06)
I think that’s even more so on YouTube for live-streaming. All the same folks show up and they talk shit, they celebrate, the meme train arrives.
ThePrimeagen
(03:02:16)
Yeah.
Lex Fridman
(03:02:17)
Okay. So now, what projects shape you as a programmer? Whether the ones you streamed or offline.
ThePrimeagen
(03:02:27)
For me, I don’t know if there’s a one project I can point to, but I can point to a specific spot where I think it happens and where I think you can learn a lot from. Any small program you write will be somewhere between 1,000 to 5,000 lines of code I consider a pretty dang small project. You can correlate this to any feature within a larger system as well. A specific feature on a website could be a thousand lines, a couple thousand lines.

(03:02:51)
There’s a point in which all of your choices add up. And I typically find that right around 5 to 10,000 lines of code. The choices you’ve made either weigh you down or kind of free you up. And so it’s right in that, that I feel like I learned the most is because I love getting to that point in a project or in some small part of the code base because at that point I get to test, A, how good were my initial gut decisions about how I designed the software, but B, now I need to go back and think about how am I going to do testing across this in a more effective way? How can I scale this out to 20,000 lines of code? How can I do all these things with what I’ve got or do I need to kind of rethink it?

(03:03:28)
And I find that that’s really where the best learning happens is that everybody has probably a different number that exists, and as you go to each one of these numbers or how well or holistic you want your project to be, I think that you’ll come up with different numbers. And I think that number should just get bigger as you get more experienced. Because there’s projects that are a million lines of code, but they’re most certainly not holistic, right? Every part of the code base is some age at some capsule of time with some sort of programming style. Some is more functional, more class-based, more, God help your soul if it’s pre-processor macros in C++. There’s all these different kind of things you’ll find throughout time.

(03:04:08)
And so that’s why I try to think about it as the feature or the thing you’re working on. It’s usually about 5,000 lines is where I find that things get kind of, did I make good or bad decisions? And that’s where I do all my learning is right on that phase. I’m trying to get it to the point where I should be able to shoot from the hip and do 20,000 lines and not be upset about it.

Asserts

Lex Fridman
(03:04:27)
So first of all, just enjoying the thing you create part, yeah. About there you can sit back and see all the parts dancing together. For me, also debugging, you get to see the choices you make materialize as how easy it is to debug. I’m a big proponent, I think you’ve mentioned this in the past, I put asserts everywhere.
ThePrimeagen
(03:04:51)
No, you are the reason why I do that.
Lex Fridman
(03:04:53)
Yeah.
ThePrimeagen
(03:04:54)
You were like the first one. Keep on going, sorry.
Lex Fridman
(03:04:56)
Really? Okay. So for me, one of the joys, whether it’s try catch box, whether it’s assert, whether it’s with the testing, I get to see the payoff of all the minefield of asserts I’ve laid out before me in my kingdom by how quickly I can debug a system as it grows larger. And I can first of all discover errors before they become real bugs and also how quickly I can solve those errors. And that brings me joy. For me, a lot of the joys of programming is creating powerful systems that don’t break down, that work correctly, that work correctly in majority of the cases. And there, sort of the stress testing the system and getting all of the signals from that system that everything is working correctly is something that fills me with joy and makes sure that the system actually works. So yeah, that, I don’t know if it’s 5, 10,000 lines of code, if it’s Java or C++ it’s millions lines of code. But yeah, in Python, yeah, I would say 10,000 lines of code. That’s when you first get to see the magic. But anyway, you were saying?
ThePrimeagen
(03:06:11)
Okay, so you and John Carmack had a conversation about asserts.
Lex Fridman
(03:06:14)
Yes.
ThePrimeagen
(03:06:14)
You talked about this idea of putting asserts everywhere that effectively crash the program when you have some state in your program that should not be represented and you have made this choice actively. And so I’ve never done that before. And I know this is like an old technique and I obviously must be too young or too dumb to know that this was a thing people did. I grew up in Java and I think that’s probably why I didn’t run into this.

(03:06:38)
So I saw that and I was like, I’m curious about how to use asserts more. And then I ran into a person named Joran. He’s the CEO and creator of TigerBeetle. It’s like the world’s fastest, greatest financial database. And it was spawned out of a company that needed to do a bunch of financial transactions. And it’s written in Zig and what they do is they do deterministic simulation testing and they just use NASA’s kind of guarantee for creating really great software. So don’t use U size, specify your exact size of int you expect everywhere. All these kind of things they do to be very specific.

(03:07:10)
And one of them is that every function should contain two asserts. Whether it’s positive space like these things should happen or negative space, like this pointer should never be null. You’re programming into things that should never happen. Normally, you would just never specify that. You’d never think about that. So every single function everywhere has all these asserts and these asserts run both in production and in testing. They’re always on.

(03:07:35)
And then they take deterministic simulation testing and run like 200 years of just random data, just complete slop going through the system and seeing how far it goes. And when an assert happens, they’re like, here’s the input that caused it, here’s every last little bit that happened, and now you can identify where this went wrong. And it was so cool. So between you, John Carmack and Joran, that’s where I got like, okay, I got to really… And NASA, I’ll throw NASA a bone as well. NASA can join in on that one.

(03:08:04)
I was like, okay, I want to try this. And I did try it. I built this big reverse proxy for me trying to do some game development stuff. And I just went ham on the asserts. And then I built the whole simulation testing thing that could do everything deterministically. So even the result of requests would all come in specific orders. And I found a bunch of bugs that I just would never have found. And then I did it for a game I was making. I found some bugs where my cursor went off-screen, it would cause all these different problems because I just never tested them. And it’s super fun and it’s like a really great way to program.
Lex Fridman
(03:08:33)
Yeah, I think it’s a skillset you grow over time. It’s not just that you have to specify the preconditions, everything that has to be true, it’s also adding things that are like, you might not even think about. You have to sort of anticipate really weird things. And if you add asserts, especially in complicated functions or in complicated classes that are able to catch really weird things, that’s going to save you so many headaches and it’s going to help you learn about your own code.

(03:09:10)
This is one of the things, I think it was Jonathan Blow that either in conversation with you or was it in a presentation, he said that when he’s starting in a project, he usually doesn’t know how to implement it, how it’s going to work. And I think he was saying that he wants a programming language. This might have been a criticism of C++, I’m not sure, where he wants a programming language that makes it as painless as possible for him to not know what he’s doing, how he’s going to implement it, and to quickly get to a place where he figures it out.

(03:09:50)
I think there’s a fundamental part of programming is building stuff while not really knowing what the next thing you’re doing is. You kind of have a loose design, maybe a strict design, but really you’re solving puzzles that are not… It is a dark room in a fundamental sense. And there you have to anticipate the kind of weirdnesses that might emerge while not really knowing everything. Just this full fog, fog of war. And there that’s a real skill to anticipate the kind of issues that might arise and put a asserts on top of them.

(03:10:30)
And it’s also like spiritually, for me, been a really nice way of programming a building of living life as having very strict asserts that say, “You’re going to fix this problem if it ever arises. You can’t just look the other way.” This idea of treating warnings as errors. Make sure your code compiles without any warnings. That was a big leap for me. It’s like, but there’s so many of them and it’s not really that important. It’s like, no, no, no warnings. Make sure you treat every single problem, even fuzzy problems seriously, because that’s actually long-term is going to create code that’s much easier to work with, much more fun to work with, much more robust, resilient to all kinds of weirdnesses, all that kind of stuff.

(03:11:21)
So it’s a different way of approaching coding, probably more NASA-like versus web programming style. But yeah, it has made programming for me personally, much more fun because one of the most painful things about programming is creating when you get past 10,000, 20,000 lines of code and you have to find a bug. And that bug can take hours, it could take days to find, and that’s torture.
ThePrimeagen
(03:11:50)
Yeah. When your system gets sufficiently large, some of these bugs are just, they are very difficult. Bless anyone’s soul that’s working on million line code bases, because it does. I can’t tell you how many times I’ve spent multiple days just trying to figure out the root cause of the bug. Not even the fix. Just like why does this happen? And that’s hard.

(03:12:09)
So I love that. I just love the asserts because I’m not good at them, I can see it’s definitely a skill that I don’t put into practice constantly, which means it’s just not like a muscle memory type thing. And so it’s just one of those things I just love. It’s such a fascinating way to approach a problem, because I would’ve never thought, you know what I’m going to do? If I’m wrong, I’m going to crash this thing and I’m going to crash it right here because I should never be wrong. But instead you’re like, “Oh, actually that makes perfect sense. I should crash this thing. I’ve done something terribly wrong here. Why would this ever exist?” And then you’re like, “This is going to solve a whole class of problems.”
Lex Fridman
(03:12:42)
Yeah. And especially if it’s in production, it’s like, well a user’s going to see this crash. It’s like, yeah, well you should minimize the number of times any user ever sees the crash, not by having a nice blue screen or whatever the fuck, but actually stopping everything. And that’s going to create an incentive for you to never have that happen. You’re actually going to put in the time to make sure it never happens.
ThePrimeagen
(03:13:09)
And the nice part is with the web and all that, you can always pop up something and say, “Hey, things have gone very, very wrong or unable to recover.” You can give them a nice message and then log it off so you can see it, and then measure how often are you doing it. I understand that there’s a bit of interestingness to a web project like do you want to always crash a server? There’s a bit of a gamble if you release a bad version and you crash all your servers constantly. That’s a pain you’re going to have to accept.

ADHD

Lex Fridman
(03:13:36)
I think this is more applicable for single systems like robots and so on. You have struggled with ADHD. I think a lot of people are really inspired by the fact that you’re able to be productive and flourish while having ADHD. How’d you overcome it?
ThePrimeagen
(03:13:57)
Well, there’s a lot of things that ADHD affects and so I’ll start with some of the easiest things, because there’s directly applicable, then these kind of collateral damage applicable things that happen. So one thing that has really helped me with ADHD is maturity. I think that’s just a thing that everyone needs more of. Meaning that I found myself getting so wiggly and so out of control when I would try to sit down and read, and I just couldn’t handle it. I just felt like I’d read a page and didn’t read anything. The part of me that just went, “Oh man, gosh, I just can’t even do this.” I had to just simply quit listening to it and said, “Nope, I’m rereading this page.” I remember reading some pages in college like 18 times in a row, just like I’m going to force myself to just do this the correct way.

(03:14:43)
And so there’s an aspect of maturity that really helps, no matter what, I will do the thing I’m going to do and I’m going to do it well and maybe it takes me a lot longer and that’s okay. That’s not the point of it. It’s that I’m doing it and that’s the point. And so that’s one thing that I think just generally helps. And ADHD, no ADHD, the resilience, emotional resilience is just a really important aspect that just helps. And so I think that has been a large part that really helps me.

(03:15:11)
There’s things that I still obviously struggle with. It’s clear where I’m really bad at stuff, and just trying to think through all the different things that I’m bad at. There’s more things I’m bad at than I’m good at. And so programming obviously has something that just allows me to remain focused and it’s like a strength of mine. And so I started off where I could just do it for a little bit and then just through kind of that emotional resilience, I was able to start doing it more and more. And so now I can just do it for like 10, 12, 15 hours at a time and I absolutely love it. And so it’s become kind of like a joy. It’s like playing a musical instrument. I’m really into it.

(03:15:51)
But then if it came down to, “Hey, you need to go schedule your own dentistry and go do all these other things or make sure the kids have this type of stuff ready for the meals you need to pack throughout the week.” I’m historically very bad at that and will probably continue to be very bad at that. And so I must say that one of the reasons why I excel so much is because I also have a wife who is so good to me and she helps clear out a lot of the things in my life that cause a lot of me kind of getting snowballed into a weird spot where I’m just distracted getting nothing done. And so she’s really helped me.

(03:16:30)
So it would be foolish of me to claim that I’ve defeated the ADHD by myself, but instead I find that the places that I can really control I’ve done a very good job at, and the things that I obviously need to do much better at, my wife has helped me a whole bunch. And so I’ve kind of cheated. Maybe I found a cheat code, a loving wife. But that has been the thing that has really helped.
Lex Fridman
(03:16:50)
You said a lot of interesting things. So on the reading and for me it’s also audiobook side, I do the same thing and I’ve gotten much better at it, which is I tune out mentally and I read a page and you don’t understand anything on the page. You didn’t actually read it. And yeah, I forced myself to just reread it or re-listen to an audiobook, which is a much more common problem for me now, and forcing myself to really pay attention. Because I listen to audiobooks often when I run and it’s so easy to just tune out.

(03:17:26)
It’s a skill. I didn’t realize how much of a skill listening to an audiobook is, especially when there’s other sensory inputs when you run. So I have to force myself to really pay attention to every single word. And if I don’t, like tune out and don’t remember what I just listened to in the past 30 seconds, I force myself to re-listen to it. And sometimes that means five times until, it’s like punishing myself to like, “You’re going to listen to this boring shit over and over until you get good at that little skill of like zoom in.” And you’re like, yeah, there’s people, they’re doing stuff, there’s nature, it doesn’t matter…
Lex Fridman
(03:18:00)
Yeah, there’s people, they’re like doing stuff, there’s nature, it doesn’t matter. You’re listening to every single word and loading it in and trying to stay focused, even there’s just so many distractions all around you. Yeah.
ThePrimeagen
(03:18:10)
It’s definitely a learned skill and it takes a lot of time. And when I say, “Oh, I was able to do from here to here,” I’m speaking over the course of like five years of doing this every day. It’s not some small… There’s no… The nice part about that decision though is you can make that decision today. You can make it right now. You’re going to be like, “From here on out, I’ll never make that mistake again. I will say I’m going to read 50 pages, I will sit down and read 50 pages, and when I get distracted I’ll go back to the last place I remember and I’ll start again.” And like that’s a decision you can make. That’s a mature, non-emotional decision to make. And you can do that, it just may be really painful for the first couple years of making said decisions. And then it gets easier and then it gets easier, and then it becomes more natural to change yourself.
Lex Fridman
(03:18:53)
Yeah. And with every medium, with every platform, I think it’s like a new skill. For me, like using social media has been that, just like I end up like doom scrolling too easily on platforms. And one solution is not to look at all, which is kind of what I lean on mostly these days, but I feel like I should be able to check, just read, okay, feel a thing, learn a thing, and then put it down, versus you have this glazed look over your eye and you’re not really paying attention anymore and you’re dead inside and you feel horrible afterwards. I don’t understand.
ThePrimeagen
(03:19:33)
The horrible afterwards is real serious. I’ve definitely… I can 100% notice that I am a more anxious person the more time I spend scrolling.
Lex Fridman
(03:19:41)
Yeah, yeah.
ThePrimeagen
(03:19:42)
I can just feel it. It’s like something inside of me that’s kind of… I don’t know how to say it other than it like wants to get out but I don’t really know what that is. It’s not anger, but it’s not… It’s very anxious.
Lex Fridman
(03:19:52)
It’s like the opposite of the feeling I have when I wake up in the morning and I’m feeling good, and I look out in nature and look at the sun and just, and there’s like a bird chirping and this kind of thing. Scrolling through social media, even if it’s like super positive stuff or whatever, it’s still not the same feeling as the bird chirping. Bird chirping on Instagram is a different bird chirping than in real life, cause bird chirping on Instagram, I’ll start swiping until there’s demons of different types fighting inside my head and then different anxiety, insecurity, whatever the hell. Just the mixture of chaos versus the bird chirping in real life. That is beautiful. But again, that’s the same thing as with the audiobook. It boils down to… Man, these people that talk about meditation, I think that’s probably… they’re onto something, because that’s what it is be able to like focus calmly and deliberately on a thing, whether it’s reading or audiobook or existence. When they sort of observe the breath, you’re able to silent out everything else and remove everything else from focus. Yeah. That’s a skill. That’s a skill.
ThePrimeagen
(03:21:05)
I heard it put really beautifully, which is that we in America really have misunderstood liberty because we typically have liberty as just the freedom to do whatever you want. And the argument was that it’s not the freedom to do whatever you want, it’s the freedom to be able to do what you will. And how often is what you actually want to do, you don’t do because you get trapped doing something that you’ve convinced yourself in this quick moment you want to do? And so it’s like, “I want liberty. I want the ability to control my energy and to be able to do the thing I want to do, not to get distracted and destroyed in all the millions of distractions.” And some of us get handed a worse deck of cards, some of us get a better deck of cards, but I don’t think there’s anybody that doesn’t struggle with it in the technological age.

Productivity

Lex Fridman
(03:21:48)
Yeah, and that’s the skill. What can you say to the skill of achieving focus in programming? Do you have a process of how you sit down and try to sort of approach a problem? So, all the different, not just distractions but the challenges of starting a project, of thinking through the design, how to maintain real focus, because it’s really difficult intellectual endeavor.
ThePrimeagen
(03:22:15)
At this point I’m lucky, but when I first started I can remember that every last part of programming. I had to go look up, I had to go read, I had side quests at all time. Every step was a side quest. Why is my screen blinking when I’m trying to render this thing out? Oh, I didn’t know about double buffering. Why is this happening? How do I even write to the screen? How do I… Everything was a question. I had more questions than answers. And so I constantly had the problem of side quests, and I find that to be a very exhausting thing. But as I learned my instrument very, very well, I don’t have as many side quests. I become more and more able to just focus on the thing I want to do. And I find that to be something that is just super, super useful. So, when I say I’m kind of lucky, meaning that I’ve spent so much of my life preparing for this moment that now when I have the opportunity to do something, I can just do that thing and I don’t…

(03:23:08)
Like I can be just on an airplane and I can just program for hours. I don’t have to look up a single thing. I don’t have to do anything. I don’t even have to test the code. I can write 1,000 lines of code on an airplane and I’m very confident that it’s going to be 98% pretty dang good. And I’m very happy about that because that allows me just to be in the moment solving the problem I’m trying to solve. Then I have 100% of my brain power solving a problem. And this is why I also… It’s the same reason why I recommend learning how to type and learning your editor so well you don’t even have to think about the action because the people that have to… Even if you just look down, that’s still mental processing power you have to spend looking at a keyboard in which you already know where the key is. You do. At this point, if you’ve been typing for thousands of hours, you know where the key is, just stop looking down, you’ll learn really quickly.

(03:23:51)
And so it’s like this thing where it’s like, “I’m not going to spend all that time and all that mental effort looking up the thing. I’m going to just memorize… I’m just going to get it in me, and then I can go fast.” And it feels good. And so that’s how I’d kind of defeat that is because now I get to do something where it’s like there’s no more questions. It’s now me just expressing myself into this medium and it feels really good.
Lex Fridman
(03:24:15)
I’m sure there’s still like things that pull at you, like curiosities, distractions, like, “Ooh, I wonder how…” Anytime you have access to the internet, you’re going to get-
ThePrimeagen
(03:24:26)
Twitter’s a big one on that one. Yeah.
Lex Fridman
(03:24:28)
… you’re going to get curious about stuff, including, I guess you’re speaking about everything in the editors optimize, but, okay, you can always improve stuff. You can always find better plugins and macros and, “Oh, let me… You know what, this thing that took this pain point, I just found this tiny pain point, let me spend the next five days creating a plugin for my editor or whatever the fuck to remove that one pain point,” when you should have just kept going as opposed to taking side quests.
ThePrimeagen
(03:24:59)
So, I have a rule which is I do not edit my RC other than some kind of cataclysmic thing, like someone updates a plugin, I didn’t know they updated it and now there’s like a hard error in my editor and I have to move forward. But I have a rule where I will edit my RC, my Neovim RC or anything once a year. Something that bothers me, I’ll write it down, I’ll remember it. I’ll be like, “Okay, I want to change that,” but I will just not go back to it. Now, every now and then I’ll break that rule if I know, if like, “Oh, I want a new remap to be able to do this one command and that takes literally 13 seconds, like copy paste, do this, bop-bop-bop, done. Okay, I have this new remap, it made perfect sense in this situation, but I don’t go plug-in exploring. I don’t try to solve every problem.

(03:25:39)
I don’t want a perfect editor because that is a pursuit that will never stop. I just go, “This is good, good breakpoint, I won’t do it again.” So, last month I probably spent a hundred hours just editing every possible thing I could about how I start up my system and make… I can have a computer from zero to 60 in almost no time now everything the way I exactly want it, Neovim and everything all perfectly set up. Happy enough, I’m not going to touch that system again. Maybe I’ll touch it next year. Maybe I’ll take a year off. It’s just I’m fine with that. I’m fine with not being perfect.

Programming setup

Lex Fridman
(03:26:13)
All right, zero to 60, let’s talk about the perfect setup. What’s your perfect programming setup, keyboard operating system, how many screens, chair?
ThePrimeagen
(03:26:24)
All right, I like all these.
Lex Fridman
(03:26:27)
IDE, let’s go.
ThePrimeagen
(03:26:29)
So, keyboard, you’re using my favorite keyboard right there, the Kinesis Advantage. Saved my career. Beautiful keyboard. Concavity and thumb clusters are just so important because if you really think about it, especially if you’re using qwerty, when you’re pressing the symbols on a standard keyboard, you’re just doing this the whole time: backspace, enter, symbols. You’re just doing this, and it just screws up your wrist constantly doing this. And this when you’re constantly doing like ctrl and shift. And it just is like messing you up, so it’s just like right here. That’s so much nicer in life. So, keyboard most important, I’d say. Get that one done.
Lex Fridman
(03:27:01)
For people who don’t know, Kinesis keyboard, I think the thing that you experience the most is exactly the thing you just said now, which is the backspace is really easy to press versus what it is on normal keyboards. So, backspace in general symbolizes you’re deleting a thing, it symbolizes a mistake. Not symbolizes, it usually means a mistake. And so not only did you just make a mistake in what you were typing, you also have to take a physically painful action, annoying action to fix that mistake. And for most of us, we make a lot of mistakes, so Kinesis just makes it pleasant and fast and easy physically to correct the mistake. That’s probably for me the number one reason of Kinesis. Everything else, yeah, super plus with the macros and the positioning, the concavity like you mentioned, but their mistakes are pleasant.
ThePrimeagen
(03:27:57)
Yeah. I am on that team, so that’s why I love that. I would say that’s one of the most important things. The next thing that I find to be very, very important is that one monitor. I’m a one monitor kind of guy.
Lex Fridman
(03:28:09)
What? Really?
ThePrimeagen
(03:28:10)
So, when I program, when I do anything… Now, when I stream, I obviously have a second computer that runs the stream because I sometimes crash my computer, I have to restart or whatever. So, I do have a second screen there that I put stuff up, but most of the time you’ll notice that even when I’m streaming, you’ve been there, I have to physically switch to the streaming chat channel for me to read it, and that’s because I’m operating off of one screen. And so I have this whole style in which I like to navigate, inspired by StarCraft, is that I believe in the press one key, go where you want to be mentality. And so everything about my setup is press one key. So, when I want to go to Twitch chat, alt-two, Twitch chat. When I want to go to my browser, alt-one. That’s my browser. Alt-three, that’s where I go to my programming. That’s power finger, obviously. The big middle finger right there, just smash it down. Alt-six is going to be gimp, so my GNU image manipulation program, so if I want to draw, I go there.

(03:29:02)
When I used to have Slack, it was alt-five. If I have a spare terminal where I need to run some extra things, that’s alt-four. I had all these kind of… Everything is perfectly mapped out to single-key. And then when it comes down to using, say, Tmux, I have all my terminals into one single terminal. And now I’m able to kind of switch between there. Prefix one goes to my Vim editor. Whatever project I’m in, it’s always the first Tmux tab, if you will. I’m not sure… They call it a session, but I’m not sure how to describe it if you’re not familiar with Tmux. A tab. Second one is like my spare terminal, third one is my long-running process terminal, my fourth one is a long-running process terminal. So, I have it all set up, so every project I go to automatically spawns session one: Vim, session two: spare terminal, session three will also open it, so it’s like everything’s just ready to rock.

(03:29:49)
Everything has been optimized to where I do that. If I want to go to a project, it’s ctrl-F, and any terminal will bring up a fuzzy find list of every one of my folders on my operating system in which I can go to with just a couple keystrokes and, boom, I’m in that one now. And so it’s very oriented to find where I need to be as quickly as possible.
Lex Fridman
(03:30:06)
Via keyboard.
ThePrimeagen
(03:30:07)
Via keyboard. Then in Vim I developed a plugin called Harpoon, which is I press one button and I can pin one of the files to like a temporary buffer. I think Projectile is potentially close to this in Emacs. I can’t remember if Projectile… I think Projectile is closer to my sessionizing script. Anyways, so now I have four pinned files in which I can go to any of those pinned files with just a single keystroke. And so now it’s just like… Because every time you develop a feature, usually you have like three files you’re kind of primarily working in. And I can fuzzy find for the other files and that’s that, but usually I just have like these three power files that I’m always swapping in between. And so it’s like now everything is just, “I want go to the browser.” That’s one press. “I want to go to my workstation.” That’s one press. “I want to go to a specific folder, I need to change folders.”

(03:30:54)
Sometimes you work between two different projects, so in Tmux that’s prefix, capital L will swap between your last two. So, I have alternate projects, so I can even swap between projects in pretty much one key. So, it’s just like do-do-do, just trying to optimize it, so I don’t think as much, because I think search fatigue is a massive fail where you have to look for it. When I see people on a Mac do this and then explode all the different ones, that gives me anxiety. I’m like, “Why are you using your eyeballs to search for what you want to do?” Make it into a key press and never think about it again, ever.
Lex Fridman
(03:31:28)
You’re making me think a lot whether I can live with your system, whether it’s better because it feels better.
ThePrimeagen
(03:31:33)
It at least intellectually feels better. It may not be great for some people.
Lex Fridman
(03:31:36)
Well, there’s a few profound things you said, which is like really what the number of windows or tasks you’re switching between, whether it’s programming, the number of files you’re working on, it’s small at any one time, at any one space of 20 minutes or something like that. So, okay, that’s a profound truth. Sometimes we think like, “Oh I need the full freedom to search,” but you don’t. You usually work on a very small slice. But I guess the trade-off there… I always have three monitors, not when I’m traveling, but my happy place is three monitors. It’s like, do you really need all of them to be present there? So, you’re turning your head. Now, the monitors I have is two vertical ones, which is just better for certain kinds of content. They’re positioned vertically, so you can read. You can use your eyes to scan quickly.
ThePrimeagen
(03:32:29)
Interesting. So, I don’t even do that. I even have it so zoomed in that I probably only have like maybe 25 lines of code at any one time on my 27-inch monitor.
Lex Fridman
(03:32:37)
Yeah, I think that’s… I think I feel fundamentally constrained when I can’t see more because your eyes are just good at jumping. You could-
ThePrimeagen
(03:32:52)
Why not search? Why not press a couple of keystrokes? Ctrl-U, ctrl-D, jump up and down by a half page.
Lex Fridman
(03:32:58)
Because the ape visual system was designed to… You’re loading a lot of information. If every time you have to investigate this table, what’s on this table, you have to press a keystroke, you could develop the skillset that integrates that information but it’s really… There is an effective thing where if you have a sheet of paper like this and I’m looking at it, my eyes will be able to load in the structure of the information, the topics of the information. You just can do it faster, I think. There’s a big cost because it’s an extra monitor, but there is some stuff that’s vertical when vertically positioned. See, code is an iffy one code because code, 25 lines at a time, I think you can do a lot.

(03:33:53)
This is more for like articles and especially with visual information in them or documentation, you can just jump faster. But I’m trying to… As you were speaking so eloquently, I was like wondering, “Am I just like deceiving myself that I need that? Can I just keyboard shortcut-ify everything and just have everything on one monitor?” That’s something I should probably try because I’m a big proponent of just automating everything with the keyboard because you can just move really, really fast, and you don’t have to think. Because I also do creative stuff, whether it’s recording music or video editing. It’s hard… Some of these programs still make it super easy for you. On Windows, with AutoHotkey you can do quite a lot, but still there’s limitations on how much you can do with the keyboard. So, it really is a pain in the to have to use the mouse, but, man, you’re really making me think.
ThePrimeagen
(03:34:54)
Even the text one, the reading one, fundamentally I think I agree with you, that you can see a lot more and you can kind of look up and down, and see those two things. And probably in articles or things like that, if there’s a graph down here that’s really big that take up your whole screen plus text, I could see why that would be very beneficial to zoom out, to be able to have all that information, but for me, I can only look at like a square inch. Really, that’s all my eyes can actually focus on. So, when I’m reading, I’m right here. Then I have to structurally try to pattern match what I think the information looks like. Then I have to start reading it.

(03:35:24)
So, I’m not exactly sure if I actually get any real benefit of having a lot of stuff on screen, as opposed to I can relax my eyes so much I don’t even have to focus. The words are so big. I actually program pretty zoomed in. My text is bigger than this when I program, and so it’s just that it’s so comfortable, I don’t even have to exert any effort to read the code.
Lex Fridman
(03:35:46)
But you have to kind of train your brain to know that you can navigate spatially using keys.
ThePrimeagen
(03:35:53)
Yeah, Neovim by the way.
Lex Fridman
(03:35:56)
Oh, maybe it has everything to do with Neovim. Okay.
ThePrimeagen
(03:35:58)
All right. And then Neovim is obviously the next big one. I love Neovim. Reason being is that I think you can make all the arguments that you want about which editor is the best. I do not think you can make an argument that Vim motions aren’t superior.
Lex Fridman
(03:36:10)
Here we go. Can you explain Vim motions? What is this? So, Neovim… Vim is an old school editor. Neovim-
ThePrimeagen
(03:36:17)
It’s a modern take on an old school editor.
Lex Fridman
(03:36:19)
Yeah. And what’s ELI5? What does it take to work with Neovim?
ThePrimeagen
(03:36:29)
Oh, okay. I thought you were talking about a Vim motion there. That’s how… I know, but you know that meme that’s just like, “Hey, Jarvis, can I tell you about Vim motions?” Because they can’t fit anything else in their head because they only have Vim motions. You said EL5, explain it like I’m five, but in my head it’s like, “Okay, E is jump to the end of the word, L is the one more…” Dude, I’m so broken that I’m like, “Okay, Vim motion,” when I hear letters. Yeah, so, you can think of it like this is that Vim has a language to describe movements in text because its primary mode of operation is manipulating or editing text. So, it is a well-thought through set of movements, deleting, yanking, pasting, copying, all that kind of stuff that goes in, motions that are optimized for working with pretty much code.

(03:37:15)
A good example, say you have three lines of code you want to delete. If you’re in VS Code, take your little beautiful mouse, highlight those things, press the backspace. That’s lovely. Your hand left the keyboard. Very simple to do though. It’s very beginner friendly. I was a huge Vim hater, by the way, so I just want you to know that before we go into this. I was probably the biggest Vim hater. If there was Saul to Apostle Paul, I am like the Saul to Apostle Paul of Vim, just so you can see how big the gap was. Or you can do something that’s like… I don’t know what the VS Code shortcut is, but I’m sure there’s some keys you can press to delete the current line you’re on. Delete, delete, delete, you can just do that. In Vim, I can go DAP, delete around paragraph. All contiguous code in that thing. I’m going to delete, so D, then I can choose my motion I want to take, AP, around paragraph. Or maybe I want to DF, meaning jump up to the next character that matches the next character I’m going to press.

(03:38:07)
So, DF opening parenthesis will delete everything from your cursor up to the first opening parenthesis. So, you get to describe your motion in these little keystrokes. And as you get really good… You’ve seen people that can master Fortnite, it’s the same thing with mastering Vim motions. When you get so good, you no longer think about each individual movement, and instead you’re just like, “Get rid of the paragraph, jump here, jump this, highlight this, yank this, do this,” it becomes so fast that you can superiorly edit text at a very fast rate. And there comes a point when you know your language really well, you know the problem you’re really working on really well, where editing text and getting code out actually becomes one of the many bottlenecks.

(03:38:43)
People always talk about, “Well, most of the time I think…” Most of the time I’m not thinking, I’m programming. I know what I want to do, I want to go as fast as possible because I’ve been just doing it for so long and I’m so familiar with the general space that it becomes a huge problem for me. I cannot tell you how many times that I’ve been purely bottlenecked by the fact that I just can’t type fast enough and I just need to get it out of my head onto the text editor. And so that’s why I think Vim motions are superior in all aspects. Keep your hands on the keyboard, on the home row, and it can manipulate text in very wide and fast ways.
Lex Fridman
(03:39:14)
Oh, so, this is not just about writing text, this is about modifying text. It’s primarily about modifying text.
ThePrimeagen
(03:39:19)
Yes.
Lex Fridman
(03:39:19)
And I’m sure that most editors including Emacs, including VS Code can do all those same things, but there is something… They just don’t encourage you to discover those things. That’s like an important thing about a lot of technologies and programming languages that a lot of them can do a lot of the stuff, but it’s something about whether it’s the community or the style of the language or anything like this that encourages you to not be lazy in the beginning and learn the fast way to edit text, in this particular example, how to use the keyboard. That’s a fascinating just reality of how technology is used. You want to be encouraged to find the fast thing as quickly as possible so that long term it’s efficient and fun to use the thing.
ThePrimeagen
(03:40:09)
Yeah. It takes a long time for dividends, like a long time, but on top of that, notice I didn’t say Vim. I’m not saying, “Go use Vim,” I’m saying, “Vim motions.” Let me give you one more example. I’m a big fan. Let’s say you have a line that contains some variable, some function you’re calling something that takes in a string. And you need to do that again, so you would typically copy that line, you’d paste that line below, you’d go into the string and you’d change the string. Let’s say it’s calling some sort of configuration, you need to call it three times with three different configuring strings. In Vim, I like to do shift-V to highlight the whole line, and then Y. Some people do YY, but I don’t like to do double ones. I like to be able to do two different fingers because you can do that way faster than one finger twice.

(03:40:53)
It’s just a little optimization for me because you can’t press that as fast. So, anyways, I’m very optimized in my approach, so I yank the line, paste the line. CI double quotes will delete everything inside the first occurring string. Then I can type the string, escape, save. And so it’s like so optimized that I can just jump so fast in between that, whereas the copying and pasting line is probably the same speed, but the navigating to the string, deleting what’s currently in the string, and then… That’s such a fast motion in Vim, and I just do that all the time.
Lex Fridman
(03:41:23)
To backtrack, really dumb question, CI, what’s the difference between typing the letters and using the letters to navigate and edit? How do you switch between the two modes?
ThePrimeagen
(03:41:34)
Okay, so insert mode means that you’re just putting in text, and then normal mode means that you’re moving your cursor.
Lex Fridman
(03:41:40)
And how do you switch between the two?
ThePrimeagen
(03:41:42)
Escape. Escape goes from insert mode into normal mode. And to go into insert mode press I to take your current cursor and go to the beginning, A to go to the end of the year Cursor, capital A to go to the end of the line, capital I to go to the beginning of line, O to put a new line below and then put your cursor at the proper intended for the language, shift-O to shift your current line down, and then put a new line in. You can see, there’s a lot-
Lex Fridman
(03:42:04)
So, you’re pressing escape a lot.
ThePrimeagen
(03:42:05)
Yeah, I mapped mine. I do ctrl-C. Ctrl-C does the same thing except for in one edge case. People hate that. I got used to it just due to the fact that I was using IntelliJ, and I really hate pressing the escape key, so I just got used to pressing escapee.
Lex Fridman
(03:42:17)
That seems like an essential thing to do if you’re using Neovim to map escape to something.
ThePrimeagen
(03:42:22)
Cap lock would be your standard go-to.
Lex Fridman
(03:42:24)
Oh yeah, I map it too. Cool. I got you. I got you.
ThePrimeagen
(03:42:27)
Yeah, so then it’s just really easy to press it, and boom, boom, boom, not a big deal at all. But yeah, I think that if you’re willing to learn it, the emotions are superior, but if you’re not willing to learn it, then they’re not superior. You should just not do it. If you’re willing to endure pain, it’s good. If you’re not, it’s actually way worse. It’s 100 worse.
Lex Fridman
(03:42:45)
Right, so if you like pain, you use Neovim. Totally. I understand.
ThePrimeagen
(03:42:45)
Yeah, you’re totally on-board.
Lex Fridman
(03:42:45)
100%.
ThePrimeagen
(03:42:45)
See, now you get it.
Lex Fridman
(03:42:51)
If you like joy, you use Emacs.
ThePrimeagen
(03:42:53)
Sorry, sorry, did Emacs ever get a good text editor? I know they’re a great operating system, but I never caught up if they got a good text editor.
Lex Fridman
(03:43:00)
Operating system? I think you’ve been miseducated my friend. So, at least 30 minutes on Emacs versus Neovim is what Reddit requested. Have you actually used Emacs in order to be able to talk so much shit or no?
ThePrimeagen
(03:43:13)
I used it for a year.
Lex Fridman
(03:43:16)
You used it for a year?
ThePrimeagen
(03:43:16)
Yeah, yeah. Doom Emacs, Spacemacs and regular Emacs.
Lex Fridman
(03:43:19)
But you don’t even know Lisp, so did you really use it?
ThePrimeagen
(03:43:21)
I kind of hacked my way through kind of like, “Okay, so this is how to configure…” You can kind of get your way through and do all that.
Lex Fridman
(03:43:28)
So, you recommend to mastering Neovim and really learn the depths of it, but Emacs is okay to just kind of use before making a judgment. I think everybody…
ThePrimeagen
(03:43:38)
You got me on that one?
Lex Fridman
(03:43:39)
Yeah, no, and what’s Neovim written? It’s Lua?
ThePrimeagen
(03:43:43)
Yeah, so Lua would be the configuration language, but you have… It’s written in C, but you have Lua 4. And Lua is just a dead simple language. Anyone can program Lua.
Lex Fridman
(03:43:51)
I actually don’t know why… I think it’s because my love for Lisp that I went with Emacs. I think you just choose a path and you walk down that path. And because there’s just such a vibrant, intense battle between the two communities, you just start fighting just because everybody else is fighting. And then one day you’re an old warrior on a horse, and you’re wondering, “What was this all for?” And it’s quite sad, in all seriousness, that I haven’t to this day tried Neovim. I think because there is a learning curve. There’s a learning curve to a lot of these editors.
ThePrimeagen
(03:44:32)
Yeah. To really learn it.
Lex Fridman
(03:44:34)
To really learn it. And I think this is some of the criticism of maybe VS Code or Sublime or Atom that it’s so easy to not learn it, to just kind of halfass use it. And there is a big benefit to having editors that force you to have some learning curve, where you take the art, the science, the procedure of editing seriously. Because you spend so much time in it, you might as well learn how to use the thing.
ThePrimeagen
(03:45:05)
My big takeaway really, what I’m trying to say with all these words is that I honestly don’t actually think that… The editor obviously does not make the programmer, but I think it says a lot about your character as a programmer if you don’t know how to use your editor well. There’s something about a person who’s willing to commit their life to programming, and spending literally 50,000 hours doing an activity over the course of their lifetime, and never take the time to learn their editor through and through. It just seems strange.

(03:45:37)
You’d never see that in another world, where people would be able to build something or do something and just completely forget how these things work, and only just focus on one part of their craft. And so, to me, it’s just like it doesn’t matter how you use it, I want to see the person that just knows how to use it, and they know how to use it well. When there’s a problem, they can say why the problem exists, and then go and fix the problem. To me, that’s like, ” There you go. You’ve done it. You now know your tool, go forth and conquer with said tool.”
Lex Fridman
(03:46:04)
Especially for tools you use a lot.
ThePrimeagen
(03:46:07)
[inaudible 03:46:07].
Lex Fridman
(03:46:06)
You have to look at your whole life, your life, whatever, if you’re a developer or anything, what is the thing you do a lot?
ThePrimeagen
(03:46:16)
Meetings.
Lex Fridman
(03:46:17)
Yeah, yeah.
ThePrimeagen
(03:46:20)
Sorry. Keep going.
Lex Fridman
(03:46:21)
Ask a question like: how can this be done a lot better? Because every single day you do this for hours a day, how many hours did you spend on thinking how to do this better or whether to do it at all, in the case of meetings? People surprisingly just don’t do this enough. I see this, just to go back to jujitsu, there’s a lot of people that show up and do jujitsu or martial arts, and they do it the same way over and over and over, and they invest tremendous amount of energy. And they don’t ask like, “How do I do it differently to improve faster?” In the case of jujitsu or any kind of sport, same with practicing the piano or the guitar, they just religiously put in a lot of time and derive a lot of joy from getting better. They don’t enough ask the meta question of like, “How can I do this better?” And with editors, it’s surprisingly how often people do just that.

(03:47:20)
With typing, it’s surprising how many people do just that. Like you said, they’re pecking or looking down. It’s like the quality of life improvement you can have by learning to touch type, by just like typing without looking. It’s immeasurable. You’re bringing a lot of joy to your life because all of us are typing a lot. And the reason, by the way, I was extremely efficient with Emacs… I’m sure you know, all jokes aside, it feels like Neovim has more room for the kind of efficiency I’ve had with Emacs to be able to move really fast as you described me to edit. There is a real joy. It’s not just efficiency, it’s a freedom that you can get when you get really good with an editor. The reason I chose to go with VS Code is it felt like there’s going to be an acceleration of features to which Neovim or Emacs will not be able to catch up, in the… and I don’t mean in the next five years, I mean in the next 30 years.

(03:48:29)
And it felt like I almost wanted to take the pain of learning new editors constantly and just switching and learning that, because I was getting so comfortable in Emacs, with this Kinesis keyboard, everything, all the shortcuts, I know how to program, and it felt like this is not… Neovim will not be here in 50 years. Possibly might be, I don’t know, but it felt like you want to learn these constant different technologies. Cursor is a great example of that. I primarily am using Cursor now. I go back between VS Code and Cursor. Just the skill of using AI is a real skill, from the shortcuts to the timing to the layout of the windows to how I think about where, when and how to use the AI that doesn’t distract me, that it empowers me, not just for the fuck of it or for the fun of it, for the actual measure of productivity.

(03:49:24)
It’s a skill. And I feel like I would be stuck in a local maximum of comfort if I stayed with Emacs. And maybe the same should be true for me with Neovim. I should try it seriously. I’m sure there’s a plugin, like a copilot type of situation that you could set up with Neovim. I should possibly consider that. But Cursor is doing a lot of really fascinating stuff on the IDE side, not just sort of generate code and edit that code manually, it’s like continuously be able to rewrite code. It’s the idea of tab, tab, tab, tab, move the Cursor around, but also modify parts of code and do the diff really nicely, that whether it’s Cursor or VS Code that wins that battle out with Copilot, I don’t know. But that feels like a fundamentally different experience than the really efficient, joyful experience that you just described in your selling me on this is Neovim. That doesn’t have an AI in the picture, obviously immediately, but you can, yeah, absolutely.
ThePrimeagen
(03:50:30)
I would 100% agree that Cursor seems like such a cool product. I actually think there’s a lot of really neat things coming down with all of that. And I could change from Neovim. I don’t use Neovim because I love Neovim, I use Neovim because I love the instrument I play. And so it’s like if Cursor can meet those needs, I could see myself moving over. I don’t have some sort of obsessed attachment with it. I am curious though that every time I use AI… I think I just have skill issues. I think I’m just so riddled with skill issues when it comes to using AI, I have yet-
ThePrimeagen
(03:51:00)
I think I’m just so riddled with skill issues when it comes to using AI, I have yet to be able to use it in a way that I really love it.
Lex Fridman
(03:51:06)
We’ll talk about it, but before then-
ThePrimeagen
(03:51:08)
Oh, ball to sit on. I forgot to say that, ball to sit on. Desk needs to be properly heighted. One monitor. Eyes should be two-thirds way up the screen. I don’t like to turn my head. I prefer my hands in a pistol neutral position. And there we go.
Lex Fridman
(03:51:25)
A ball to sit on. Yoga ball.
ThePrimeagen
(03:51:26)
Yoga ball.
Lex Fridman
(03:51:27)
What’s that about?
ThePrimeagen
(03:51:28)
It just helps just maintain good posture, because when I have something to lean against, I do this.
Lex Fridman
(03:51:33)
You’re for hours sitting without… Wait, what are you doing?
ThePrimeagen
(03:51:37)
I sit on the ball, and then I bounce.
Lex Fridman
(03:51:40)
Is your back leaning on a thing?
ThePrimeagen
(03:51:41)
No.
Lex Fridman
(03:51:42)
What the fuck?
ThePrimeagen
(03:51:43)
Well, how else do you-
Lex Fridman
(03:51:46)
You’re the only person in the world sitting on a yoga ball as you program for hours. You do realize this, right?
ThePrimeagen
(03:51:52)
It feels great. The problem is whenever I get a back, I just slouch and I find myself just getting uncomfortable. And I’m like, “I’m uncomfortable.” My shoulders are getting goofed up. I’m chicken necking constantly. It’s just like-
Lex Fridman
(03:52:10)
But you’re able to keep your posture for hours on the yoga ball?
ThePrimeagen
(03:52:12)
Yeah. And so I can just do that. And then I find myself, if I slouch, I’m like, “Okay, Nope. Got to get back.
Lex Fridman
(03:52:18)
Do you have incredible back muscles or what?
ThePrimeagen
(03:52:19)
No. Well, I don’t think it takes incredible back muscles to-
Lex Fridman
(03:52:23)
Keep posture.
ThePrimeagen
(03:52:24)
… remain upright. Yeah, I think that’s a pretty basic human function. I would not consider myself a strong person.
Lex Fridman
(03:52:29)
Yeah. Basic human function. I don’t know.
ThePrimeagen
(03:52:33)
Facts and logic.
Lex Fridman
(03:52:34)
Okay, cool. With one screen. Neovim. What operating system?
ThePrimeagen
(03:52:41)
Linux, just because I want a good window manager. That’s the whole press one button, bring up Chrome. I just use i3. I’m sure I could use something better than i3. People always tell me all these window managers are really great. But I just have those three screens I switch between, so it doesn’t really… I don’t really care what I use, just long as I can press one button and go.
Lex Fridman
(03:53:02)
Yeah, I’m the same, so half and half. Half Linux, the other half Windows with Linux, meaning WSL. What’s that? Windows Subsystem for Linux.
ThePrimeagen
(03:53:12)
Weasel.
Lex Fridman
(03:53:13)
Weasel. See, no, there’s got to be a better one that’s more positive. Weasel just sounds-
ThePrimeagen
(03:53:19)
Seems right up Microsoft’s alley. That seems perfect.
Lex Fridman
(03:53:24)
People often accuse me of being a shill for somebody, sometimes dictators. If I’m a shill for anybody, it’s for Windows. There you go. I get paychecks every week from-
ThePrimeagen
(03:53:35)
Dang. Bought by Bill Gates.
Lex Fridman
(03:53:37)
Well, he’s not Microsoft anymore.
ThePrimeagen
(03:53:38)
I know.
Lex Fridman
(03:53:40)
Developers, developers, develop. No, I’m just joking. I think, man, I need to try Mac. I need to try. I’m surrounded by people with iPhones. I use Android.
ThePrimeagen
(03:53:52)
I use the Android.
Lex Fridman
(03:53:53)
Yeah. There you go. See? Oh.
ThePrimeagen
(03:53:55)
We’re losers together.
Lex Fridman
(03:53:56)
Losers on a sinking ship. Okay, just to stay on Neovim for a sec and to give love and a shout-out to your friend, Teej.
ThePrimeagen
(03:54:09)
He Streams, by the way.
Lex Fridman
(03:54:10)
He’s a streamer. And I subscribed. And I’ve been enjoying it. My allegiance is slowly shifting from you to him. The quality is far superior with him, the looks, the intelligence, the skillset, everything, just far superior. No. Okay, he-
ThePrimeagen
(03:54:29)
You know you’re making his day.
Lex Fridman
(03:54:32)
All right. He mentioned that he loves Neovim because it gives him the ability to eliminate having to do things he doesn’t like. That’s just a nice way to frame what this automation process that you described of automating a way, assigning shortcuts to things that are painful, that procedure. I wonder if you agree with that.
ThePrimeagen
(03:55:00)
Fully agree. We have very similar mentalities when it comes to usage of Neovim, why people should use it, all that kind of stuff, and how to even use it well. He definitely takes it probably to a further degree. He spends more time automating and all that. I don’t necessarily derive a lot of joy from getting the perfect setup. But a lot to learn from. He’s very, very good at what he does. He’s 30 years old, been programming for not too many years, and he is one of the most talented developers for sure. It’s very shocking to see how smart someone can be.
Lex Fridman
(03:55:33)
People should check him out at teej_dv. Teej.
ThePrimeagen
(03:55:38)
DV. His last name is DeVries. DeVries.
Lex Fridman
(03:55:42)
Oh, it’s not developer. Okay, cool.
ThePrimeagen
(03:55:42)
Yeah, yeah, it’s just TJ. That’s just his name just spelled fun.
Lex Fridman
(03:55:46)
All right, Teej. What do you love about him?
ThePrimeagen
(03:55:48)
Wow. How much did he pay you to ask these questions?
Lex Fridman
(03:55:51)
Thousands of dollars. Thousands.
ThePrimeagen
(03:55:51)
Just so many dollars.
Lex Fridman
(03:55:53)
I can’t even count that many dollars.
ThePrimeagen
(03:55:59)
Trust. Obviously trust is the biggest thing, especially in the, quote, unquote, “streaming” YouTube world, if you will. It’s very easy to find people that will want to be a part of stuff. People tend to latch onto things, and it’s very hard to find someone that you can really, really trust. And so he’s just somebody whom I can genuinely trust. He will always tell the truth. He’s all the right things for a good friend in this kind of endeavor.
Lex Fridman
(03:56:25)
As a good friend, he told me questions I could backstab you with.
ThePrimeagen
(03:56:30)
Okay, I hate him. I forgot how much I don’t trust him.
Lex Fridman
(03:56:35)
Speaking of a harpoon, you mentioned it. He said to ask you basically how many years or decades it’s going to take to transition to Harpoon 2 to actually release it, develop it, and so on. Can you describe what Harpoon is and why your seem to be incapable of finishing a single project?
ThePrimeagen
(03:56:58)
That was a lovely framed question. Harpoon 2 is actually done. This is what I did: To avoid the swirl in the thousands of questions I will inevitably get, I kept the master branch as Harpoon 1, and I’ve kept Harpoon 2 as Harpoon 2 branch. And people that don’t read the read me to say that I just use Harpoon 2 now, that’s their fault. That’s it. I really don’t like answering hundreds of questions about open source stuff. I used to love doing open source and all that, but I got my soul crushed during the Falcor years, and so I guess I’m just allergic to being a really active maintainer.

(03:57:33)
I build everything just for me. Harpoon’s just literally just built for me. I spent three months trying to figure out the most optimal navigation for files, and that’s what I came up with. Harpoon, it’s a take on alternate file. If you’re familiar with the alternate file, typically you’ll have this in all editors where you can go back to the file you were just in. And so that means you can have effectively two files you swap back and forth in. You’ve probably used it a bunch; really fast way to navigate. Pretty nice thing to do. I want alternate file, but three of them or four of them, and so that’s all Harpoon is is just being able to pin a file. And so I have one button to press to go to a file, another for another, another for another. And so I can have up to four. I just had my four power fingers. For Dvorak, what is that? That’s HTNS. If I go Ctrl H, T, N, or S, it goes to one of the four files. And that’s it. That’s all it is.

(03:58:21)
And you can technically make it so you can add in functions and be able to execute things externally. You can open up terminals, you can send requests off to servers. You can do anything you want with it, I just have it primarily designed for opening files.
Lex Fridman
(03:58:33)
Since you mentioned it, what keyboard layout do you use? You use Dvorak?
ThePrimeagen
(03:58:36)
I use Dvorak, but I used a custom version of Dvorak. The reason why I used it is in 2017, we are just having my second kid, it was Christmas and I’m having so much pain in my arm and I’m sitting there freaking out like, “Oh my gosh, is this the end of my career? Am I done programming? Is this all over?” And so I decided that I was going to create my own keyboard layout optimized to prevent the pain that I’m experiencing, so I used to Dvorak as the base and then laid out the symbols in a symmetrical, reasonable way so that it’s opening, closing, opening, closing, opening, closing. And they all are right here. I actually have to hold shift to press a number. Symbols are actually my first thing I get to press. And so it’s very optimized for a laptop keyboard layout so I can use my laptop in a very efficient, nice way. That’s how I got started on Dvorak and all that. I wouldn’t actually recommend it because I didn’t have a Kinesis at the time. I didn’t even know Kinesis existed at that time. And so when I discovered Kinesis in also 2017, that’s when I was like, “Oh, okay.”
Lex Fridman
(03:59:37)
Would you recommend Kinesis to people?
ThePrimeagen
(03:59:40)
I’m technically sponsored by Kinesis, so it’s hard for someone to believe someone that’s sponsored by it. But I did use it before I ever became sponsored. They’re the only sponsor that I reached out to and said, “I need a sponsorship from you. I’m going to use you either way. You can say no, but I really love it.” And for the first three years of using Kinesis, they gave me free Kinesises, Kenisi, as my sponsorship.
Lex Fridman
(04:00:04)
Kenisi. Yeah, I’m always torn. I tried to leave so many times.
ThePrimeagen
(04:00:08)
You can’t. It’s too good.
Lex Fridman
(04:00:10)
But, see, I have this absurd situation of traveling with it.
ThePrimeagen
(04:00:16)
I relate.
Lex Fridman
(04:00:17)
Yeah. I’m literally going to the war zone in Ukraine, I have a Kinesis keyboard, a laptop, and just a few other small things and that’s it. And it’s like is Kinesis keyboard really going to be 30% of volume that you’re bringing to a war zone? But-
ThePrimeagen
(04:00:37)
Looks like the answer is yes.
Lex Fridman
(04:00:39)
Yeah. Do you really derive that much value? I think it’s probably spiritual or psychological for me. It feels like home. There’s comfort associated with it. I try to leave.
ThePrimeagen
(04:00:50)
I love this experience. It’s like a relationship you have with the thing.
Lex Fridman
(04:00:56)
It is. But I’m trying to figure out if it’s a toxic relationship or not. I think it’s mostly love. I think it’s love. Like all relationship, there’s some push and pull complications, but-
ThePrimeagen
(04:01:06)
They say that distance makes the heart grow fonder, so maybe sometimes the Kinesis keyboard needs to stay at home and the laptop keyboard can be the one so that your heart grows even more fond and that connection grows even deeper.
Lex Fridman
(04:01:18)
I already miss it as you say it, so I don’t know. I think it’s coming along to all the trips. If it breaks down, though… I was worried that Kinesis was shut down as a company. I’m like, what’s the business model here? Who actually uses these keyboards? But apparently they’re still going strong.
ThePrimeagen
(04:01:33)
Yeah. Who uses these keyboards? As you use the keyboard “I have to take it with me everywhere.” I wonder who uses these keyboards.

Coffee

Lex Fridman
(04:01:42)
Yeah. Yep. I should mention that one of the things when I first became a fan of yours, I heard you talk about coffee and term… I still don’t, by the way, understand what you’re even talking about. I need to actually use it. But you run, amongst many things, a coffee company. Man, this smells so good. This one is dark mode, dark roast, whole coffee beans. There is seg origin, dash, dash location. Brazil.
ThePrimeagen
(04:02:14)
Yeah, there’s a bunch of stuff on there.
Lex Fridman
(04:02:16)
Stuff on there that’s very devy. Shop, server, web. Can you legit, as such, order coffee via SSH?
ThePrimeagen
(04:02:25)
As of right now, it’s the only way you can get the coffee is via SSH. Okay, can I just origin, origin story you?
Lex Fridman
(04:02:33)
Yeah, yeah. Yeah, right, I was going to do some kind of command line. Command to request or dash dash help or something or-
ThePrimeagen
(04:02:33)
Command coffee?
Lex Fridman
(04:02:33)
Command coffee.
ThePrimeagen
(04:02:44)
Okay. TJ and I, again, same Teej, Teej TV, about… By the way, very amazing designs done by David Hill. They’re very, very good. Let me give the basic ideas. It must’ve been about a year and a half ago, TJ and I were talking like, “Hey, every one of these people that have some sort of following, some sort of online presence, they’re always selling a thing,” but I got nothing to sell. I don’t really want to do merch. I’ve never really enjoyed doing merch. I just find that, I don’t know, it’s just not as much fun for me.
Lex Fridman
(04:03:15)
Don’t want to have a tequila?
ThePrimeagen
(04:03:15)
I don’t want a tequila. I want something that-
Lex Fridman
(04:03:19)
Like The Rock.
ThePrimeagen
(04:03:19)
And I also want something that I really don’t feel bad about selling. There’s a lot of people that will go on the internet and they’ll shill for a whole bunch of products like, “Oh, okay, try this, try this.” And this is why I’ve only ever really done Kinesis is because it’s like, well, I can point to something that was really bad in my life, I was very scared, and now it’s not bad anymore. It’s like, okay, that one made sense. But everything else always has been… It’s harder for me. And so we just talked for so long, and we love Neovim, so we’re just like, “Why haven’t we could do something from Neovim?” And we’re laughing about that, ordering from Neovim is just so ridiculous.

(04:03:51)
And then at some point, we’re just like, “Well, wait a second. And maybe we could do coffee. Every developer loves coffee. Maybe we could figure out this coffee business.” And so I had a good friend named Dax, THDXR. Dax, yeah, Dax. The most sassiest man alive.
Lex Fridman
(04:04:11)
Sassiest?
ThePrimeagen
(04:04:11)
Oh yeah, he has a lot of sass.
Lex Fridman
(04:04:13)
Beard?
ThePrimeagen
(04:04:14)
Yep, he has a beard. He does SST. He does a lot of stuff. Very, very talented. We’ll call him DevOps engineer. He’s more than that. But very talented guy. Him and another person named Adamdotdev, vegan, by the way, great guy. We take him to Korean barbecue all the time. He eats nothing.
Lex Fridman
(04:04:33)
That’s great.
ThePrimeagen
(04:04:34)
And Liz, she has been super important to the terminal coffee company. I think without her, we would not have been able to do what we have done. And then also David Hill, designer, he does Laravel. He designs for Laravel. Very talented designer. And so we all came together. And we were just laughing about how could we do something that’s just ridiculous? And that’s what we came up with. Yeah, there you go. You just open the website. You literally cannot order. We actually do not allow you to order.
Lex Fridman
(04:05:05)
The website is something that looks like the terminal. Use command below to order your delicious whole coffee bean. SSH terminal.shop.
ThePrimeagen
(04:05:14)
Yeah. You can only SSH into it. You have to copy that command and throw it in there. If you want to add in the little terminal shop for your known hosts, you could do that.
Lex Fridman
(04:05:22)
How do you handle payment?
ThePrimeagen
(04:05:24)
Through Stripe. And so one of the things, we’ll be adding a mobile checkout to where I’ll show a QR code in the terminal and you can just check out on your phone, but right now, you enter in your credentials, it goes to Stripe.
Lex Fridman
(04:05:34)
Via all terminal, like SSH.
ThePrimeagen
(04:05:36)
Yeah, SSH, obviously it stands for Secure Shell. It uses elliptical quantum safe algorithms to ensure that your data’s not being intercepted.
Lex Fridman
(04:05:45)
Yeah, but does he use AI?
ThePrimeagen
(04:05:48)
I’m pretty sure Dax uses AI. That-
Lex Fridman
(04:05:50)
You said quantum, so I don’t know.
ThePrimeagen
(04:05:52)
Quantum AI?
Lex Fridman
(04:05:53)
Can this-
ThePrimeagen
(04:05:54)
Fusion quantum AI?
Lex Fridman
(04:05:55)
Can this even be a company if it’s not using AI?
ThePrimeagen
(04:05:59)
We have some crypto chains with some quantum AI that’s powered by Fusion, so it’s pretty wild. Anyway, yeah, we just came together where we thought, what is the… That was from the Mike Tyson fight. It was literally that night Mike Tyson kissed the reporter and then walked out without any clothes. We did an ad for somebody.

(04:06:17)
But we decided to make a coffee shop, and then we thought instead of just making it Neovim, what if we made it from SSH? Because everybody has SSH. You have VS Code. Launch VS Code. You can order coffee from within VS Code. Because your little bottom terminal that has access to SSH, bada bing, bada boom. It’s fun. And so we really-
Lex Fridman
(04:06:39)
I love this.
ThePrimeagen
(04:06:39)
We just wanted to do something where there’s no level and there’s no world that makes me feel bad about selling this in people buying it. It’s good ethical coffee. We developed the entire supply chain and everything. It’s all packaged, it’s all boutique. It’s pretty high-end coffee. It tastes really, really good.

(04:06:57)
At this point, I don’t like drinking other coffee. I get upset about it because it’s not as good. And so it’s funny that I’ve fallen for my own stuff. I’m high on my own supply pretty hard right now. I just got done ordering 16 bags and gave it out to my family to try to convince them. But it’s just something where it’s like I didn’t sell you a software product that’s going to influence your startup that could potentially lead to disaster, I didn’t convince you to do a bunch of stuff that’s going to change your career, I just said, “Hey, here’s some coffee.” And it’s like a fun experience.
Lex Fridman
(04:07:27)
Yeah, it’s fun, everything. The humor on it is great. People should go to terminal.shop.
ThePrimeagen
(04:07:32)
SSH terminal.shop.
Lex Fridman
(04:07:34)
I’m speaking to people that don’t know what SSH is. And there, you can read the command and then figure out how to use SSH in order to… It’s a kind of documentation, right?
ThePrimeagen
(04:07:45)
Yeah.
Lex Fridman
(04:07:45)
On the website.
ThePrimeagen
(04:07:47)
If you can’t use SSH, you probably should just not worry about buying our coffee. That’s the whole-
Lex Fridman
(04:07:52)
Well, you can learn.
ThePrimeagen
(04:07:53)
You can learn. If you’re active and you’re a computer person, you’d like to launch the terminal and feel like a hacker, go for it. We even have subscriptions.
Lex Fridman
(04:07:59)
What I would love to see… This is how it came up I think on the cursor conversation, is that I would love it if an AI agent did this, like Anthropos computer use or something like that, actually took the action of ordering the coffee while it was programming.
ThePrimeagen
(04:08:17)
Yeah, like, “Hey, order me some coffee,” and it actually go off. “Give me dark roast.” Order coffee. It could actually go through the whole flow of order.
Lex Fridman
(04:08:24)
Yeah, the whole flow. But even better, if you didn’t ask it to order coffee, you asked it to do something, and as a tangent, as a side quest it did that. Which is computer use does that. They showed off that it’s able to go to I think Google for some images, take a pause, and then continue doing other stuff. Anyway, yeah, super cool idea. Love it. Speaking of which, let’s talk AI.

Programming with AI

ThePrimeagen
(04:08:50)
All right.
Lex Fridman
(04:08:50)
You’ve been both positive and negative on the role of AI in the whole programming software engineering experience. As it stands today, what do you think? What’s your general view about AI? What is it effective at? What is it not so good at?
ThePrimeagen
(04:09:06)
Okay, my general view is it comes down to something that’s pretty simple, which is that if you’re doing something in which is very predictable, AI is really nice. When you’re doing something that is just not predictable, AI is not very nice to use. If you’re using anything that’s more cutting edge, AI will not be using it, or AI won’t be very good at doing stuff with it. It’s not great at Zig because Zig is just, say, less documented. It’s really great at TypeScript.

(04:09:39)
I think there’s a lot of interesting things that are going to come down through AI that I think a lot of people aren’t really prepared for or thinking through. TJ’s the genesis of this idea, but the idea that I think there’s going to be a lot of market manipulation, if you will, through AI. Meaning, hey, you want to research, say, best woodworking tools. Someone’s going to be buying an ad spot. Someone’s going to be buying premium training data. They’re the ones that get the big boosts in the LLMs. But LLMs don’t really have to market as an advertisement because it’s not really directly an advertisement, they just had a more premium spot, per se, in the training data; a little bit extra learning to it.

(04:10:21)
It’s like there’s a lot of things about AI that I fear upcoming. A lot of it just comes down to people not learning or making the trade-off where productivity is the only thing that matters. And I don’t think productivity is the only thing that matters. If you want to build something complex and difficult, productivity is not the only thing. You actually are going to have to do deep learning and pursue it beyond the basics.

(04:10:42)
And so I see AI as this really cool thing. It feels like a magic trick. I remember the first time I used it, I got early access to GitHub Copilot. In fact, Nat Friedman saw my Twitch clip of me asking GitHub for it, and he sent me early access himself. It was awesome. And when I used it, it predicted an if statement correct and my mind was just absolutely blown because I had nothing before then, and now it’s just like first time ever. And I just remember thinking, man, this is going to change programming so much.

(04:11:12)
And then the more I used it, the more I just… For me personally, I kept introducing bugs, and I couldn’t figure out why. And what I realized is that I developed… I wasn’t copiloting well, I was autopiloting much better. And my ability to read code versus my ability to critically think and write code, they’re definitely different sets of skill levels. I don’t consider as well when I just read code as opposed to what I write code. And so I struggled there.
Lex Fridman
(04:11:38)
I do think that’s a skill set.
ThePrimeagen
(04:11:40)
Yeah. Skill issue for sure.
Lex Fridman
(04:11:42)
Skill issue. For people who are not aware, that’s a hashtag thing sometimes used mockingly in this case. There’s several layers mockingly, but also seriously, meaning the criticism is grounded in the fact that you lack the skill versus some kind of fundamental truth. Yes. I think that that’s the reason I use actually Copilot cursor a lot is for developing the skill of editing AI so I can just learn how to do that better and better. Because I think as I do that better and better, I start to utilize AI better. At this time, it is a bit of a boilerplate code thing, but you can do out of the box novel design decisions or tricky design decisions from scratch but fill out stuff using AI and then just learn the skill of modifying.

(04:12:43)
I personally just… It’s more fun to program with AI. Even when I delete a lot of the code, it’s more fun. It’s less lonely. It’s what I imagine pair programming to be. And I’ve never done it, but it just feels like that friction that you get when you’re staring at an empty thing is not there. Empty function, empty class, it’s just more fun, less lonely.

(04:13:19)
And I do think that a lot of the easier type of coding, it really helps with like interacting with APIs, basic things that I would usually have to look up to stack overflow for. It’s just really fast at that. As example, just interacting with the YouTube API. The YouTube API documentation is not very good. And you can just load it all in there and ask it to generate a set of functions that access the API, all kinds of read and write operations, and it figures it all out. Well, you do have to read. You have to read and check everything.

(04:14:04)
And you start to develop the skill of understanding where it misinterpreted the task. What is that skill? I don’t even know. You have to be empathic about what the AI, what its limitations are. A lot of the times that has to do with prompt engineering. You have to at the same time understand what the AI is aware of. What did you actually give it as data to be able to generate the code? A lot of times, we don’t realize that we’re not giving it enough information. Okay, okay, all right. You have to be empathic. Be like, okay, these are the files it’s aware of. This is the specifics of the question you asked it. You have to imagine you’re an intern that doesn’t know anything else.

(04:15:09)
Oftentimes, we want the AI to just figure out the things that’s left unspoken. But you can’t know those things, you have to specify those things. And so you have to actually be much more deliberate and rigorous in the things you specify, is to spell it out. And so I just have this sea of prompts that I have saved up, and I’m building these library of different templates for prompts and it’s a mess. And I’m sure there’s a lot of developers that have this similar kind of mess.

(04:15:38)
A lot of it has to do long-term with the tooling that’s going to improve that. One, the systems are going to get much more intelligent when you don’t need the nuance. And two, there’s going to be the tooling that allows you to specify those things and load it in correctly and give all the context that the system needs in order to make the good decisions. And maybe the system asks you follow up questions with, “Here’s things you didn’t make clear,” all that kind of stuff.

(04:16:02)
A lot of that has to do with the interface, with the actual design of the tools. Like we said with Cursor, it’s going to keep getting better and better and better. My sense is developers in general should be learning this to not be left behind, to see how that can be used as a superpower to boost their productivity, their effectiveness, their joy of programming versus be seen as a competitor to them or something like that.

(04:16:33)
But for me already, it’s been a big boost to productivity. If you measure the actual how quickly you’re able to get a thing done, it’s been a big… And measured not across minutes and hours but days also. Sometimes there’s things I have to do that are not that important that I’ll just out of procrastination will push off.
ThePrimeagen
(04:17:03)
I know that.
Lex Fridman
(04:17:03)
And AI helps me actually get it done, because that thing, the empty page, like I mentioned before, it helps me write the thing, get it done, get it tested, ship the thing. Maybe it’s just because it’s just less lonely to work with an AI. I don’t know. I don’t know if any of that made sense, but-
ThePrimeagen
(04:17:22)
It all made perfect sense. I really do like that phrase, it makes it less lonely. I think there’s something to that that’s interesting having just some level of interaction that’s not just like an LSP autocomplete, having something that’s actually a little bit more than just that where it actually is thinking through and you can see a different thought and you’re like, “Oh wow, that’s a way different approach than I would’ve taken. Hey, that’s cool. I like these kind things.”

(04:17:43)
And the thing is that I’m not a AI negative person. I can see why people really, really like it. I used Copilot from when Nat gave me the access all the way up until about six months ago. I used it for quite some time. And I really did enjoy the things I used out of it. It did the opposite for me. I felt like I was more reviewing than writing and I felt like I was more just letting things slide where I just didn’t really think too heavily about stuff. And just I wasn’t as engaged. And so I’m like, “Okay, something’s kind wrong here.” And that’s just a me personal thing. I recognize that is not how someone should approach these things. That’s not a good reason for why you should or should not use AI. I just don’t think that that’s right. I could probably correct that and figure out a better way to do it.

(04:18:32)
I’ve been meaning to have another AI round, and so I’ve been thinking about maybe I just need to spend two weeks in Cursor and just fully embrace what does it mean to be somebody like this? And what can I do with these new powers? Have they improved to the point where they’re actually good? And for me, because a lot of the decisions I make, a lot of the little functions I’m writing, it’s not because trying to write this function to solve this problem, it’s because I’m writing these functions or this set not just to solve this problem but because I know in about another 2,000 lines of code of building all these other things, I’m going to need to start doing this next activity. It’s like I’m trying to really try to chess move myself into the exact things that, as I let things go faster, I fall apart on that chess move. And again, skill issues for on my behalf. And I mean it in the truest sense of the word where it’s like I’m making a critique because I don’t use it well enough.
Lex Fridman
(04:19:23)
I don’t know if this is a general rule, this is my anecdote data. The better you’re programming, the less you want to use the AI, the more gets in the way. The good programmers-
ThePrimeagen
(04:19:32)
It’s fair enough, as far as I can tell.
Lex Fridman
(04:19:34)
The more beginner programmers are much more happy to use AI. When I use AI, it’s for basic for just… I don’t know if there’s a better term. It’s not boilerplate, but it’s pretty easy programming. And that kind of programming is much easier to do. The 10 X, not to use the meme, programmers that I know that are ultra productive and brilliant people, they hate AI. They’re like, “This is nowhere close to what’s needed.” There’s something to that. I still think they should be using AI just for the learning because it’s going to get smarter, it’s going to get better. It’s the same thing, it’s like when you super optimize Neovim or super optimize Emacs, you may not discover new things that are in the pipeline, so it’s always good to be training in that way.
ThePrimeagen
(04:20:26)
Let me ask you a question here just for my understanding. You talked about this idea that you have all these LLM prompts, all this big backlog of messy LLM prompts that you have these templates for that you can do various actions. You have these strategies of making itself explain itself and then do the right thing. As far as I can tell, that’s really built into a lot of people.

(04:20:46)
Well, then you make this phrase where you’re like, but then at some point, the interface is going to get better, and maybe it can do a lot of these things better where I won’t need that. Then my question is, well, is anyone actually falling behind for not using AI then? Because if the interface is going to change so greatly that all of your habits need to fundamentally change and it will be able to clarify and make all those statements, have I actually fallen behind at all? Or will the next gen actually just be so different from the current one that it’s like, yeah, you’re over there actually doing punch card AI right now. I’m going to come in at compiler time AI, so different that it’s like what’s a punch card?
Lex Fridman
(04:21:23)
Obviously open question. It’s a fascinating one. I personally think, yes, you’re falling behind. Not you, but if you’re not-
ThePrimeagen
(04:21:24)
It could be me, it could be me.
Lex Fridman
(04:21:33)
… not playing with it, you’re falling behind because the thing I’m doing with the prompts is you’re learning, you’re building up this intuition about how AI works. You’re understanding what is its strengths and weaknesses? Not even the current version, but the next version and so on. What does it mean to teach an AI system about the world? What kind of information does it need to make effective decisions? I think that does transfer to smarter and smarter models. You’ll need to make less rigorous and specific in details instructions over time, but you still have to have that kind of thing.

(04:22:21)
I think it’s a skill of almost empathy with an AI system because it doesn’t know… You know what it’s missing? It’s missing common sense. It’s missing long- term memory. A lot of things, when we talk to other humans, they have a basic common sense about reality, and AI systems often lack that kind of common sense. And they also don’t remember things. You have to realize there’s a constant blank slate happening. It’s almost just a skill of talking to an AI system that I’m training. And by having to write all those prompts and communicating back and forth to understand what kind of prompts work better or not, you build up that intuition. And also just raw the skill of reading somebody else’s code. Maybe for people who work on large teams, that’s a skill that’s already developed. For me, not so much, so learning how to modify the code that somebody else written is a real skill.

(04:23:23)
And also, the other thing you mentioned, which is considering another perspective on a piece of code is really nice, but it is also a skill to understand, okay, this is what you did. There’s a skill to asking a question that code that’s been generated such that you can have a conversation about the approach that was taken. I think there’s just a lot of subtle, little skills involved in a cooperative endeavor to code, kind of like there was a real skill issue between you and Teej when you guys did the video of two idiots, one keyboard. People should go watch.
Lex Fridman
(04:24:00)
You guys did the video of two idiots, one keyboard, People should go watch that video, where you guys obviously sucked at it.
ThePrimeagen
(04:24:07)
Yeah, co-using.
Lex Fridman
(04:24:08)
That was pretty cool, which you guys did, which is controlling one Neovim interface from two different keyboards.
ThePrimeagen
(04:24:15)
Yeah. And then we each get an allowance of certain characters or motions we could perform.
Lex Fridman
(04:24:19)
Yeah. And so you both had to communicate together. That’s a real skill. I’m sure you can get super efficient with that, but it just takes time to learn that kind of thing. So yeah, I think there’s some value to it, but I think there’s a learning curve.
ThePrimeagen
(04:24:37)
One thing to be pretty clear is that I actually use AI quite a bit. I just don’t use it for programming. And so one thing I’ve been trying to get it to is to be able to have a long interview or understand what Twitch Chat is saying and become Twitch chat and be able to speak as if it is Twitch chat. Try to learn how to prompt it in different ways. And so I think those things for me are just really fun.

(04:24:56)
I tried to get it to learn how to play tower defense. I made a tower defense game in Zig and then made it play tower defense, and then played a Claude 3.5 against OpenAI. Claude 3.5 would do better during the daytimes, and OpenAI did better during the nighttimes. I don’t know why, I have no idea what was going on there, but one would just start winning and the other one would start losing. It was just very strange. And so it’s just this, I’m learning to prompt well, but I’m learning to prompt in a very different axis. I just don’t find it very useful yet in programming.
Lex Fridman
(04:25:23)
In programming. And I should also say that I’m using it in every walk of life, in every context. I use that same kind of exploration about prompts and so on, I’m using and learning. I think it legit is a whole field in itself.
ThePrimeagen
(04:25:41)
Yeah.
Lex Fridman
(04:25:42)
Prompt engineering and how to interact with AI systems, I think it’s worth the investment. Can you actually speak to that? Because I saw you’re basically pulling from Twitch chat and having an LLM speak. I didn’t realize, you’re not reading the exact chat messages.
ThePrimeagen
(04:26:04)
Yeah.
Lex Fridman
(04:26:04)
You’re doing kind of some kind of summarization?
ThePrimeagen
(04:26:07)
Yeah. So I ended up making like eight queries off to OpenAI where it’s just like the first thing is like I have it have it like a default personality. “Hey, you’re Randall, the manager, you’re a software engineering manager.” Kind of explain their position, what they like, what they don’t like, and then be like, “These are the list of thoughts you have in your head and you need to talk to this person and ask them a question.”
Lex Fridman
(04:26:28)
This is amazing.
ThePrimeagen
(04:26:28)
“Give me 10 of these responses that you think are probably thoughts that you have and you want to ask.” Make it kind of give you a list and then be like, “Okay.” Then re-prompt and be like, “Hey, you’re Randall, you’re this, this, this, this, this. You have these 10 questions before you and now you need to select one of them and reword it in a way that sounds more like you, the engineering manager.” And I’m constantly trying to make it iterate on itself as opposed to just one-shotting it. And I found if I iterate too much, it loses what it was originally trying to ask if I don’t do it enough and it’s just too degenerate from Twitch chat. And so it’s like I have a lot of improvement to do with this idea-
Lex Fridman
(04:27:07)
Just to clarify, you’re feeding in Twitch chat, “You’re a manager, these are the thoughts you have in your head, pick out some of the most profound thoughts”?
ThePrimeagen
(04:27:20)
Effectively. It’s like depending on what I want it to do, I’m trying to work on a better system still for it.
Lex Fridman
(04:27:20)
Brilliant.
ThePrimeagen
(04:27:25)
And so it’s like, “How can I give voice to Twitch chat? Can I make it so that I can create adversarial characters against Twitch chat or for Twitch chat? Can I incorporate YouTube?” All that kind of stuff. And how do you describe to an LLM to role-play into its position? And so just thinking through those kinds of things. So maybe I am having some prompt skills, but it’s just not in the coding world yet.
Lex Fridman
(04:27:46)
Sure.
ThePrimeagen
(04:27:47)
One day I’ll get there.
Lex Fridman
(04:27:48)
I saw that you were playing with different voices. There was like a sexy voice?
ThePrimeagen
(04:27:54)
That started off as a French voice-
Lex Fridman
(04:27:55)
French voice?
ThePrimeagen
(04:27:56)
… and then it turns out ElevenLabs just cannot do a French lady. And when you do multilingual French lady, she starts talking. It’s like, “What? What is this?”
Lex Fridman
(04:28:08)
I tuned into one of your streams and there’s this lady in a sexualized way.
ThePrimeagen
(04:28:16)
It became too funny. And so we call her Not French Stormy Daniels.
Lex Fridman
(04:28:20)
Oh, nice.
ThePrimeagen
(04:28:21)
Yeah. But I want to go back to the AI and some of the aspects.
Lex Fridman
(04:28:24)
Sure.
ThePrimeagen
(04:28:25)
And so my big gripe with AI has nothing to do with its capabilities. It’s exactly capable, as it should be capable, because that’s what people programmed it as. The things that I really dislike is, A, there’s a whole group of people that are just like, “The end is nigh. AI is here, you just need to stop programming.” I cannot tell you, even you mentioned Peter levels earlier, he made some sort of tweet and one of the person’s responses was, “No one in 2025 or whatever should be acquiring hard skills. You should rely on everything for the AI effectively.” And it’s just like these are really damning pieces of advice for young people. Young people are being told that you should never become an expert in anything, you should always offload.

(04:29:04)
And the problem is that anyone worth any of their salt will tell you that AI, though can produce code, is going to get it wrong in a huge number of cases. And as the code becomes bigger or more complex or more input, it’s going to just start kind of sloshing back and forth between bugs. And so if you don’t have those hard skills and you’re not ultimately the driver at the end of the day, you’re going to really find some hard times, and your ability to progress will be directly bound to how good the LLMs are. So if you believe that the LLMs will be vastly superior to humans in the next year, maybe that’s a good bet. But if they aren’t, then your skill ceiling is bound to whatever they are.

(04:29:42)
And even beyond that, there’s just a level of information problem, which is like, “Do we even have enough compute power to be able to solve things at this real scale?” And even if we did, if everybody started using it right now, “Do we even have the compute power for everybody to use it right now?” There’s a lot of kind of bounding questions, there’s privacy concerns, and I just don’t want people to make the immediate, or what appears to be the obvious choice, where you don’t need hard skills, you don’t need these things, we just need to only think creatively. It’s like, no, I don’t think so. I think these hard skills are going to be around for quite some time even with a massive improvement in the AI, you’re going to really be needed to step in regularly for quite some time as far as I can tell.
Lex Fridman
(04:30:27)
But I also think even on top of that, just even acquiring the hard skills or whether that means, programming from scratch, for example, in the context of programming, that’s going to make you better at steering the AI, not just correcting the AI, but steering the AI. I think there is some kind of, if you know how a computer works, you can program Python better. It’s maybe counterintuitive, but if you know the low level abstractions, some intuition around that, you can steer the high level abstractions better.
ThePrimeagen
(04:31:01)
Yeah.
Lex Fridman
(04:31:01)
But that just seems to be the case. Unless of course AI becomes like truly super intelligent like many levels above, but it’s very unlikely in the short term. And in the long term it’s still good as it gets better and better and better to be able to ride the wave of the improvement.
ThePrimeagen
(04:31:19)
Yeah, I’m on that team very much so.
Lex Fridman
(04:31:21)
A lot of people have written to me, I think a lot of developers, programmers are really concerned about the future of their profession in the context of quickly improving AI systems. So do you think AI will eventually replace programmers?
ThePrimeagen
(04:31:37)
The hard part about that phrase is you use the term eventually, meaning do I think in five years, 10 years, a hundred years? What does that term actually mean? I think at some point if all things continue at the current rate of improvement, there does come a point where programming as a hard skill does become unnecessary. At some eventual point, way, way down the road, yes. I don’t know what that point looks like. I don’t know when it’s going to happen. I don’t even attempt to make predictions about that. But there are still some leaps and bounds we need to make.

(04:32:11)
I mean even just societally, there’s plenty of companies that don’t even allow you to use AI. I mean, there’s just practical problems that exist. So that’s a question I just try not to answer in the direct sense. There will come a day if humanity continues and all things continue in a good positive direction, where a lot of skills will go out the window due to immense computing systems. So, yeah, I’ll give you that one. But it’s just like if I don’t think it has anything in the near term, there’s been no computer improvement up to this date that did not result in more jobs.
Lex Fridman
(04:32:46)
Yeah, absolutely. I would say that I think it depends how you define programming also, because when the community moves from assembly to C, from C to, I don’t know, Python and JavaScript, that’s evolution. That’s really painful for a lot of people who are used to programming that lower-level language, so there’s going to be a continuous evolution. And maybe that means with AI, there’s going to be more and more evolution towards natural language as part of the tool chain like being able to learn how to write proper prompts. Yeah, because natural language is still a language. And in the long term, it’s possible that a large percentage of programming is natural language. There are probably still going to be some percentage that’s not, that’s going to be extremely structured language.
ThePrimeagen
(04:33:45)
Right now, I don’t think we are anywhere near natural language being possible because it’s ambiguous. And I think what we’ll end up seeing as people push really hard into this, you’re going to see some sort of pseudo-lang, which is going to be a language for AIs in which you prompt, which is going to be less ambiguous. People keep striving towards the less ambiguous state. And at that point you’re just programming yet another evolution into a higher order language. And perhaps that is a future in which people will have a more terse language. I’m just not sure how much more terse it can get.

(04:34:19)
I mean, all I see is that if you say natural language can be used in the pipeline, you’ve just made that many more people can become programmers, which means that much more software will eventually be created, which means there’s that much more software that will need to be maintained, and just becomes a real big snowballing effect.
Lex Fridman
(04:34:37)
But there’s just people who are programmers who are worried about their jobs. Not a complete replacement, but maybe a rapid evolution of what it means to be a programmer. Like you mentioned, if natural language becomes a way that you can communicate or you can program, that means the pool of people who can get programming jobs changes rapidly, so they’re really concerned.
ThePrimeagen
(04:35:01)
And to some extent, because no matter how much we want to say how good AI is, there comes a point where there exists a bug, there exists a large piece of software in which to describe the change requires just pages and pages of description to the point where it is significantly just faster or easier for someone to just whip something out. There’s definitely a balance there, it’s not like a perfect trade-off. And so I think people need to quit worrying and think about how they can integrate it and try, like prove it to themselves. Do they actually make themselves irrelevant?

(04:35:37)
And if you truly make yourself irrelevant, I would challenge you that you’re just doing something that was just slightly too complicated to automate. If you’re only writing just straight up crud apps from backend to front end and simple table displays, yeah, maybe we just couldn’t quite automate that away and now we just have something that can just do that a little bit better, so now that’s automated away, but that’s not really programming. That’s almost like building Legos at that point, where the designs are already set, you just simply have to move piece from bag into correct position.
Lex Fridman
(04:36:11)
Yeah. Is there something you recommend how a developer or programmer could avoid the situation where AI can automate them away?
ThePrimeagen
(04:36:23)
I think that the bigger the project you can manage, the bigger the thing you can build, the more understanding both down and up the stack you can go, the more valuable you become. Because if you understand how to build something in the front end, okay, well, now you kick off some LLM task of some sort, that’s going to go off and make a change to the front end. Okay, while it’s doing that, you can go and kick off something in the CLI tool, you can go and you can go kick off something somewhere else. And as these things come back with results, you can review the results, make sure it’s the way you want it, change it, commit it, go to the next. You only become more, as you said in the end, more productive if we reach this state where it’s truly able to do that.
Lex Fridman
(04:37:00)
I think there is like a skill to working together with AI, which is why I’m kind of excited to watch you keep trying to do it.
ThePrimeagen
(04:37:08)
Yeah.
Lex Fridman
(04:37:09)
It’s like we don’t know how it fits exactly, but it feels like AI should be a boost to productivity. And I definitely think it’s a boost to just the joy of programming. I think there’s a lot of people, yeah, it’s a job, but it’s also a source of meaning, a source of joy. Programming is fun, you’re creating something cool, and also potentially that a lot of people use.
ThePrimeagen
(04:37:34)
There’s this one thing that just really frustrates me, and this is kind of going into the Devin category, which is that I want an intern that cares.
Lex Fridman
(04:37:41)
Yeah.
ThePrimeagen
(04:37:42)
You don’t get that out of an LLM. It does not care, meaning that I don’t want it just to make a UI for me that displays these icons like I asked, I want it to care, I want it to think about it, I want it to present to me and me being like, “Oh yeah, yeah, that’s great.” And then me to make changes. And then later on it’s like, “Actually, I really rethought about this and actually it’d be way better if we change…” It doesn’t actually care about the craft. But when you work with an intern or you work with somebody else, they care. When they factor something, they actually go over and go, “Oh yeah, this is actually kind of bad. I’m going to come back to that.” They finish this, they go back over here and they make this even better. They actually care about the thing itself. It’s a completely different experience. I just want something that also cares that wants to make the thing better, not just simply accomplish the task.

(04:38:24)
And I know I’m asking way too much that’s not… Now we’re getting into Blade Runner’s level AI. I just want something that it just feels like I’m missing that, where it’s just like it will complete the task to whatever level it understood what I was prompting, but it doesn’t actually care about it.
Lex Fridman
(04:38:42)
I mean, there’s so many aspects to caring versus the trivial version of that is a kind of restlessness where you want to keep improving, and I think that is very much AI could do, where it constantly just ask itself, “Can I make this better?” And if it keeps doing that, it probably is going to take it to some ridiculous place, so actually it’s also knowing when to stop. I think developing something you can call taste, which is like trying, working extremely hard, constantly improving until it just feels right. This is it. And I think that is a thing that AI is not good at where it’s just like, “Yes, this is it.”
ThePrimeagen
(04:39:26)
I’ve had write iterated three times and three was the-
Lex Fridman
(04:39:27)
Yeah, that’s it. We’re now there. And I think ultimately that is what humans are amazing at, which is like knowing when something is right like, “This is it.” Especially as you understand, as you develop taste in a particular industry, in a particular context application, knowing like, “This is it. Yeah, the rounded corners on this button, that’s exactly that. That’s beautiful.” So it’s just a sense of beauty, a sense of function, and efficiency, and so on. Yeah, but humans could do almost like supervision of AI systems in that context.
ThePrimeagen
(04:40:08)
Yeah.
Lex Fridman
(04:40:09)
Yeah. You’ve ranted about Devin just full of rage.
ThePrimeagen
(04:40:15)
I mean, first off, the people that run Devin are extremely nice. I want that to be understood. I don’t have some sort of upsetness against them or anything like that. Second, Devin, it’s like the full package when it comes to programming. So it’s going to have, you’re going to give it a task and a repo, and it’s going to go through, it’s going to try to understand the repo and the task, make the change to the repo by exploring it, then actually make a commit to GitHub and explain what it did. So hopefully you have this whole offline thing, which is the other part of this AI part that I actually really like, where it’s just like, “Go fix this thing.” Then I can just go and unbroke and fix this one thing and come back and go, “Okay, good enough, merge, boom.” I want that kind of running, being able to complete things.

(04:41:00)
I think the ideal solution is that you can start giving it small bugs and it goes and fixes these bugs and you can just come back to these backlog tickets that no one ever does, and it actually starts going through these backlog tickets, and it’s actually a really amazing experience. So I love the idea. I think we can all agree that that sounds great, but every time I’ve done it and I’ve asked it for many and I try to keep narrowing down the problems, the more narrow the problem, the better it does. So if I’m like, “Just add one singular icon. And when it gets clicked, I want you to do this just console, click me. At least create me an SVG and place it so it’s nicely placed.” The more narrow the task, the more likely it’s to be successful.

(04:41:39)
There’s like a certain level of specifying where if you specify too much, it just like can’t do it. If you specify too little, it just does weird things. So it’s kind of like this very kind of fun, unique way you have to play the balance game. But so far, every time I do these things, I always end up going, “Gosh, I should just get better at Tailwind and write it myself,” because I always go back and I just rewrite it, and then it’s just like, “Dang it. What am I saving at the end?” I feel like I’m not saving anything yet. And it’s just like this, “I want it so bad.” I actually want AI to be great because then I can really go fast. I mean, I can go amazing fast, but then I always just go, “Gosh, I should have just learned Tailwind myself to like the nth degree and just go fast.”
Lex Fridman
(04:42:17)
Yeah. We should also mention that debugging, this might be intuitive or counterintuitive, AI is really bad at.
ThePrimeagen
(04:42:26)
Yeah.
Lex Fridman
(04:42:27)
Like that is one of the hardest… It actually makes you realize how special humans are and how difficult the task of debugging is. Obviously, for trivial debugging, maybe you can find bugs, but that is the real art of programming is finding bugs, logical bugs, extremely complicated rare bugs, edge cases. AI can assist, but man, the hard ones really require so much context, so much experience, so much intuition from, again, operating in a fog full of uncertainty. It’s hard.
ThePrimeagen
(04:43:06)
Yeah.
Lex Fridman
(04:43:07)
Of course AI could maybe create logs and do traces and do some kind of loading a huge amount of data that humans can’t, but ultimately that just means it could be a better assistant in debugging versus the actual lead debugger.
ThePrimeagen
(04:43:27)
Yeah. I mean, it’d be great if they could. I mean, the more it can do that, the better, because as far as I can tell, correct me where I’m wrong on this, current state debugging is really, it looks at the code, it looks at the bug problem, it just kind of tries to text-predict where it’s most likely accurate, and then just tries to fix that spot. And so it’s like it’s likely this spot, you said admin panel, it’s slightly off, this, this, this. It’s probably this location, which could actually be a really great way to do search, let me do semantic searching, point to me where this is, because maybe that is a really great way to navigate large code bases. It’s like smart intelligent search. As opposed to trying to make it do the thing, ask it to just help you do the thing in pinpointing problems. I’d love to see more of that, because that’s for me is like the exciting part.

(04:44:11)
And there’s this really great article by creator or maintainer of curl, it’s the I in the LLM stands for intelligence. And he writes curl and maintains curl. Curl has been inundated with security problems and all this, and it’s all from LLMs being like, “Oh, I found a security flaw. Here’s the security flaw,” details it out in the code. And he’s just like, “Okay, how did you reproduce that? Show me,” because if you look at the code right here, that’s actually an impossible situation you’re speaking of. And it’s just like going in these circles and security right now is being inundated, these bug bounty programs are being inundated by LLM-submitted responses because they can’t actually analyze the code beyond just like basic text prediction. “Oh, this is a stir copy. Stir copy is commonly referred, blah, blah, blah, blah. Boom, there you go. Here’s the bug.” And it’s just like, “No, that’s actually impossible because the if statement right beforehand leaves the function if the string is too long, so it’s like we don’t even run into this case. It’s impossible what you’re saying.” So debugging is very interesting.
Lex Fridman
(04:45:08)
Yeah. I mean, that for me would be the big, if it can solve that, not solve that, but improve that, that would be huge, whether it’s agents or just LLMs integrated into IDE.
ThePrimeagen
(04:45:19)
I think there’s this whole idea, I call it a denial of attention. I think there’s an entire attack vector that’s going to be happening. We’re using LLMs to generate fake bug reports, fake all these things to just actually effectively to demotivate and hurt open source maintainers. Polykill was the first bug that kind of had this experience, is this denial of attention where an active malicious maintainer just hounded the owner. And then a white knight came out and offered to buy some stuff from under them. And when they bought it, they actually replaced it with a malicious piece of code and then used it. So there’s this whole security world that’s developing around using these in a very aggressive format.
Lex Fridman
(04:46:04)
I mean, it’s a fascinating world we’re entering into, but I do agree with you that human developers will be a huge part of that world.
ThePrimeagen
(04:46:11)
Yeah.
Lex Fridman
(04:46:11)
That the job might evolve, but it’s going to be there. If I can, I didn’t really look at this page, I thought it would be cool to go over with you. This is, again, the-
ThePrimeagen
(04:46:21)
Stack overflow, my favorite
Lex Fridman
(04:46:23)
… Stack Overflow Developer Survey, talking about their sentiment and usage of AI systems. The general sentiment of, yes, 61% say yes, they use it and 25% say no, don’t plan to. So majority use it, majority have a favorable sentiment over it, favorable or very favorable or indifferent. That’s like looks like over 90%.
ThePrimeagen
(04:46:52)
That’s really surprising that that many people just have no plan in looking into AI. As much as I don’t like using it for coding, I hope one day I can use it more. And so it’s like, to me, I’m always looking for the next thing. I’m just surprised that people are, that, I guess obstinate for it. Obviously, the second one, the AI tool sentiment, it must be only the users who responded to the top two of that first one just given the amount of respondents.
Lex Fridman
(04:47:17)
I wonder if no and don’t plan to are people who have tried it and quickly built up the intuition like, “This really sucks.”
ThePrimeagen
(04:47:25)
Yeah.
Lex Fridman
(04:47:26)
So it could be like experienced programmers. They’re like, “No, this is not making me more productive.” 81% agree that increasing productivity is the biggest benefit that the developers identify for AI tools. Okay, so this is, what are the benefits? Increased productivity, speed up learning, greater efficiency, improve accuracy in coding, make workload more manageable, improve collaborate. Where’s the fun, increased fun? I would say that’s like number one for me.
ThePrimeagen
(04:47:55)
Maybe speed up learning is like a subcategory of fun. If you’re able to learn more and be able to become better. To me, that sounds good.
Lex Fridman
(04:48:05)
I don’t know. It’s different because productivity is part of fun too. There is just a lightness. I mean, maybe improved collaboration, all of these elements for sure.
ThePrimeagen
(04:48:16)
My time using Copilot, there was certainly a level of wonder that would happen for quite some time where it’s just like, it’s just amazing what it can do.
Lex Fridman
(04:48:23)
Yeah.
ThePrimeagen
(04:48:24)
I’m just super impressed by what it can do, even though I don’t use it. It’s amazing to me that we have something that can even get that close.
Lex Fridman
(04:48:31)
In terms of accuracy of AI tools, only 2.7% highly trust-
ThePrimeagen
(04:48:37)
I would say that you have to be very green to think that you should highly trust an AI output. You should be very skeptical.
Lex Fridman
(04:48:43)
Yeah, I don’t know where I stand. Probably somewhat distrust. Highly distrust seems aggressive.
ThePrimeagen
(04:48:47)
It does seem a little, true. You should always assume that there’s something wrong, and then from there you can go and challenge it.
Lex Fridman
(04:48:57)
And then estimation of whether AI can handle complex tasks, most people don’t think it can handle complex tasks. I mean, it seems like people have a good sense of what it’s able to handle and not.
ThePrimeagen
(04:49:07)
I would argue that people don’t have a good grasp of what complex is in programming.
Lex Fridman
(04:49:11)
Sure, yeah.
ThePrimeagen
(04:49:12)
If you say, “Write me quicksort,” some people will think quicksort’s super complex. But I would argue that that’s actually probably the simplest thing you could ask an AI to do. Things that are so well documented, it’s going to do a great job at that.
Lex Fridman
(04:49:26)
Yeah. Probably high-level design decisions, which people don’t even use AI for right now, I guess agents are supposed to be doing that kind of stuff. That’s probably the most difficult thing or the most impactful thing, or the most difficult thing is finding bugs.
ThePrimeagen
(04:49:42)
Yeah.
Lex Fridman
(04:49:43)
AI tools next year, writing code, and so on.
ThePrimeagen
(04:49:47)
Now, this one, the ethics part. I’m actually super curious your take on the ethics. Will we see Europe laying down some new regulations?
Lex Fridman
(04:49:56)
Oh, boy.
ThePrimeagen
(04:49:56)
What about artists? What about people that are really… Because the difference between coding and artists is very, very simple. If you gave me a sheet of paper, I could draw you a crab. You go, “That’s a crab.” But you can’t do that with coding. It’s like it’s right or it’s wrong. There’s not a variation of interpretation for what a crab is. It’s like, “No, you cannot make that statement.” It’s very bounded in what it can express. And I could see why artists, that’s a very frustrating point. And then who gets rewarded for all that?

(04:50:28)
And then there’s like the whole thing with coding and licenses. How much of it is GPL licenses, do you think, they’ve scraped and used as training data? GPL forces open source. What are you going to do with that one? That means your model might need to be open source. OpenAI may have to get forced open, all their previous stuff if there’s any hint of GPL.
Lex Fridman
(04:50:48)
Yeah, that’s a weird one. That’s a really weird one because most of these models I think are training on data they don’t technically have rights to be training on.
ThePrimeagen
(04:50:56)
Yeah, there’s a lot of questions.
Lex Fridman
(04:50:58)
There’s an unspoken, it’s a real Wild West.
ThePrimeagen
(04:51:01)
Because you could imagine that, I always use Europe because they tend to have like maybe the most consumer protection laws out there. You could imagine what happens if a law came down that said that if you used a model that produced GPL potential code, you have to open source? How many companies are going to be like, “Oh my gosh”? Like, “You have one year to get rid of all code that was generated that’s potentially GPL-sourced from a model.” You could imagine just the sheer panic that’s going to happen. It’d be a fire sale of code.

Advice for young programmers

Lex Fridman
(04:51:31)
So given all that, can you give advice to young programmers? This is another question from Reddit, the infinite wisdom of Reddit. “What should a person in their early 20s do to move forward in the tech industry?” And this is an interesting addition to the question, “And by doing it, will this be walking on someone else’s path?”
ThePrimeagen
(04:51:59)
I am going to try to answer that question, I guess the best I can, which I think that if you’re entering into the tech world, one of the hardest pieces of advice that I took a long time to learn was I became enamored and addicted. Obviously, we talked about that I program for way too many hours, forgetting to spend the time I needed with my wife, with my friends, all that stuff, like totally wrapping myself up into one activity. I think though it made me who I am, it was probably an unhealthy activity and probably not a wise activity. And so the best advice I can give is that you got to develop the love, the skill, the desire for it. Whether that’s just only using AI agents, programming yourself, using Zig or programming JavaScript, whatever that flavor is that’s going to get you coming back every single day, getting the reps in the gym, if you will for programming.

(04:52:53)
But also knowing how to value what is valuable and not getting lost in the sauce where you’re just so stuck on trying to make the next greatest startup that you sacrifice your health, you sacrifice your relationships. Or even worse, you sacrifice your own morals to take certain shortcuts that you probably shouldn’t be taking in life to be able to achieve these things, because I’m sure there’s hundreds of horror stories you could hear where people definitely shortcutted their morals for monetary success.
Lex Fridman
(04:53:22)
Yeah. I mean, the golden handcuffs comfort can destroy the soul in some sense. Yeah, so that’s really important to remember. There’s young people kind of thinking, “Do I even want to be a programmer now?” It seems like AI is getting better and better at programming. If they were trying to make that decision, would you still say, “Yeah, if this is something that fills you with joy”?
ThePrimeagen
(04:53:51)
I still want my kids to learn how to program if I can answer that, if that’s a good enough answer-
Lex Fridman
(04:53:51)
Yeah, that’s a really powerful answer.
ThePrimeagen
(04:53:57)
… in the sense that my kids are decades younger than a young person trying to learn how to program right now. And so I’m hoping that my kid can run and build whatever he wants in Roblox. I’m showing him ChatGPT and be like, “All right, let’s ask questions. How do we do this?” It’s still extremely confusing for him to do all these things. And so it’s like, “Let’s do this.” I want him to learn and be effective, and maybe one day he has to throw away all those skills in 20 years. But I bet you that whatever skills he threw away or whatever hard skills he had to throw away, an entirely new field that none of us have thought about, just like if you would have asked somebody in the ’70s about social networks, they’d be like, “What the heck are you even talking about?” Things will exist in the future that are going to be massively different, and crazy, and exciting.
Lex Fridman
(04:54:40)
Maybe in virtual reality.
ThePrimeagen
(04:54:42)
There you go.
Lex Fridman
(04:54:42)
Maybe all of us actually down the line will just be building video games.
ThePrimeagen
(04:54:46)
Just entertainment for all, the brave new world of our world?
Lex Fridman
(04:54:51)
Well, I think entertainment is a kind of trivialized version of what a video game could be. It’s like, what is the purpose of life anyway? I mean, it could be a deeply fulfilling video game. It doesn’t have to be just like dopamine rush. It could be educational, it could be scary, it could be challenging, forcing an evolution, the leap into adventure that it makes up a fulfilling life. That could be video games. Who knows? Especially in virtual reality. I tend to… That’s the other thing. I play a lot of video games. I think there’s a lot of room to make video games deeply fulfilling, like there’s a lot of space where that can go.
ThePrimeagen
(04:55:42)
I didn’t know you played a lot of video games, because when I asked you specifically, “Should I play World of Warcraft or do Advent of Code,” you’re like, “Advent of Code, Advent of Code.”
Lex Fridman
(04:55:50)
Oh, well, that might mean I’ve never played World of Warcraft because there’s certain games I avoid. Fortnite, by the way, I think was one of them because I was worried it’d become too addicted.
ThePrimeagen
(04:56:00)
Yeah, yeah.
Lex Fridman
(04:56:01)
So there’s certain games I just know I won’t get super addicted to. Like for example, I’m terrified of Civilization. I have never played a Civ’s game because I’m worried. I’m worried the dark path in my lead because there’s some games just really pull you in. I’m much better with, that’s why I play Skyrim. I can play these games or a Baldur’s Gate, and moderate how much I play. And they could be like a lifelong companion versus an addiction where it’s like sunrise and you’re like, “What’s happening with my life?” And I find myself naked behind a dumpster somewhere just wondering what happened. Yeah, so that’s how I choose my video games.
ThePrimeagen
(04:56:43)
You’re not the first person who has specifically called out Civilization.
Lex Fridman
(04:56:48)
Yeah.
ThePrimeagen
(04:56:48)
I’ve had more than one person, also very high up in the tech world, be like, “Civilization is my downfall. If I get near that game, I’m done.”
Lex Fridman
(04:56:56)
Yeah.
ThePrimeagen
(04:56:57)
I’ve never even played the game. Now it makes me be like, “Dude, I got to give this a try. That sounds crazy.”
Lex Fridman
(04:57:02)
Yeah. And the new one is actually supposed to be-
ThePrimeagen
(04:57:00)
Give this a try. That sounds crazy.
Lex Fridman
(04:57:02)
Yeah. And the new one is actually supposed to be really, really good. What were we talking about? Yes. For that same young developer, is there a trajectory through jobs that you could give advice on? So you started out with Schedulicity?
ThePrimeagen
(04:57:17)
Yeah, that was my first full-time…. When I had the government contracting one before, that wasn’t quite full-time. It was in C. It was a lot of fun. And then building my own startup for quite some time. So if you count either of those as full-time, then those would be the full-time. Schedulicity was the official on the docs.
Lex Fridman
(04:57:32)
Is there some value to jumping around, working in one company and another to try to figure out what brings you joy?
ThePrimeagen
(04:57:41)
I think there’s a lot to that because not every job you’re going to get is going to be great. Now, your first job you could get could make you think you hate programming. It happened, I did an internship at a place and I keep on surprising you with more kind of things I did in the past, did an internship at-
Lex Fridman
(04:57:59)
Fuck. You did so many things. It’s incredible.
ThePrimeagen
(04:58:02)
… a place called Total Information Management System. Remember when I talked about that hours ago, about healthcare and that and industrial shipping and all that? It was a C-sharp shop. It was so bad that after I did that, I went and changed my major to mechanical engineering first semester in college.
Lex Fridman
(04:58:02)
Oh, boy. Oh, boy.
ThePrimeagen
(04:58:17)
I thought I, “Okay. Actually I like computer science. I hate the programming.” So just because you’ve had a job doesn’t mean it’s going to be the one. And the thing is, here’s the best part though, if you get a job and you like it and you want to do it and it’s exciting, you don’t need to change. I think a lot of people are like, “Oh, I got to find the next thing. I’ve been here for two years.” There’s of this, you got to move around mindset. I don’t think you have to move around. I don’t think it hurts your career. Because if anything, you’ll gain more responsibility and you’ll be able to talk with way more authority. And the next time you interview, you’re going to be way more into like, “Oh yeah, I had to get these X people and these X people to be able to do all this stuff.”

(04:58:53)
And it’s like you can talk with much more authority if you stay at a place longer. And that’s nothing but benefits in my book. It’s only if you stay at a place because you’re afraid or you don’t want to… You already have something that works for you and you just never want to change and you’re just like, “I get to go in and just be completely mindless.” I think if you go mindless for a couple years, you’ll find yourself… That’s the only real danger. You just come out with nothing at all.
Lex Fridman
(04:59:15)
Especially when you’re younger, that’s the whole point. You’re like, “Take the risk. Take the leap out to the next thing, to the next thing.” And not for money, but for just personal joy, joy.
ThePrimeagen
(04:59:24)
And money could get at the end, that’s the best part. When you don’t strive for the money, sometimes the money just shows up anyways.
Lex Fridman
(04:59:29)
Yeah. And some of the, what makes life worth living is the people you work with, a good team. Some of it’s not to be generic, but culture matters. It’s whatever makes you happy. For example, I just had won’t call out places, but there’s certain companies where everybody is very nine to five. And even if the work is exciting, they don’t work hard enough I would say. I’m one of those people that likes to go all out, likes to be surrounded by people who are super passionate. Now to be fair, a lot of them don’t have families or don’t-
ThePrimeagen
(05:00:06)
Yeah.
Lex Fridman
(05:00:07)
It’s a fascinating choice. I really don’t want to talk down on any choice like work-life balance or not, I think both are beautiful paths. And if you really derive a lot of value from joy from your work, going all in, at least for some stretch of your life is a beautiful thing to do. Just all out, full-on passion, sacrifice a lot of social life, all that kind of stuff. I don’t know. That could also be beautiful.
ThePrimeagen
(05:00:39)
There could be something very, very exciting about that in some sense, especially if you’re building your own thing. I could imagine that would be very exciting. If I was Amazon, Jeff Bezos building Amazon, one could imagine that those early years were probably very rough and the amount of hours he probably put in we’re very, very rough. But I will say that there’s this unique aspect in our culture where we make this as an equal trade-off between family or work, like “Oh, you do or you don’t have to have kids.” And my only real notion with that one is that you will never know your capacity for love until you have kids. You just don’t know. And some people are like, “Oh yeah, but I love my dog.” It’s just like I loved my dogs too. And then I had kids and now my dogs are, “They’re all right. I like them.”
Lex Fridman
(05:01:26)
I get it.
ThePrimeagen
(05:01:26)
I could come home and I pet Indy and I’m like, “Oh, Indy.” And then I’m just like, “Okay, bye Indy.” I can’t even describe the difference between the two, it’s not even the same. And so that trade-off making is no one can tell you what it’s like because there’s a real reality that’s right now, and I’m sure, I’m 100% positive this is with my wife as well, where if right now we got news that said you have some medical procedure where if we do this, you will die, but your kid will live, there’s not a question in my soul that I wouldn’t do that. If I could look into the future and if I had to die right now knowing that my kids would’ve a better life, they would be happier, they’d be more fulfilled and all those things, I guarantee you either my wife or I would take that every single time.

(05:02:08)
It’s just like you’ll never be able to say that about most things. People will jokingly say that until it’s actually on the line. But it’s like with that, you just have this ferociousness. I can break out and sweat thinking about somebody fictionally pushing my kid to the ground, actually get real adrenal responses flowing through my body. So it’s just such a different world and it’s hard to explain. And you could never have convinced me when I was young that it’d be this big.
Lex Fridman
(05:02:32)
Yeah, yeah.
ThePrimeagen
(05:02:33)
Yeah. I thought I knew. I didn’t know.
Lex Fridman
(05:02:35)
But to add on top of that, some of the most successful people I know, some of the most productive people I know have kids. So I don’t know if it’s even a trade-off. That love you feel, it seems to be a catalyst to make sure you have less time, but you’re going to use that time better to be productive.
ThePrimeagen
(05:02:56)
I would argue that it definitely changed a lot of my life and how I approach problems and everything, in a very different way.

Reddit questions

Lex Fridman
(05:03:03)
Let me ask some random questions from Reddit. On a scale of one to 10, how much do you hate every product Microsoft has ever created and why is it a 10? I think we covered that.
ThePrimeagen
(05:03:16)
We haven’t technically covered it.
Lex Fridman
(05:03:18)
There you go. All right, go ahead. Go ahead.
ThePrimeagen
(05:03:20)
Okay.
Lex Fridman
(05:03:20)
Use your time.
ThePrimeagen
(05:03:22)
The only thing I’ll say is that I don’t like that Microsoft pretends to be the good guy when what they really wanted to get you addicted to their products, to get you to use their products as much as possible so they can extract as much money out of you.
Lex Fridman
(05:03:32)
Well, in this world, are there really good guys?
ThePrimeagen
(05:03:35)
That’s a great point. I would argue Neovim is a great guy. There’s no way they can make money. Justin Keyes is the benevolent dictator and he thinks deeply about the product and tries to make it the best as possible. Whereas something like Microsoft, they made VS Code as a loss leader. Copilot’s probably operating on a loss leader. These things are all getting you so tied into, GitHub, remote workspaces, CI, Copilot, you become this trapped in permanent person and if that price rises, the switching cost is so great at some point that you’ll never be able to switch. That’s my only fear is that Microsoft was once accused of EEE and it feels like they’re EEEing again.
Lex Fridman
(05:04:13)
Yeah, I’m nervous about criticizing a good thing because you could see an incentive to do that good thing, like Google creating all these services that don’t make money like Gmail for example, you can cynically say they’re only doing that to tie you into an ecosystem so they can basically keep you for life. But also it’s awesome that they created Gmail-
ThePrimeagen
(05:04:37)
Yeah.
Lex Fridman
(05:04:38)
… and they created an incredible product, so-
ThePrimeagen
(05:04:40)
I can side with you on that one. It is a good product. VS Code is a good product.
Lex Fridman
(05:04:44)
Yeah.
ThePrimeagen
(05:04:45)
Now, I think don’t put that on the… But it is fine. They did a great job.
Lex Fridman
(05:04:50)
Yeah. So there is going to be financial incentives behind some of these companies. And by the way, me defending, not defending, but saying positive things about Microsoft is just so I could talk shit to Prime. But that’s…
ThePrimeagen
(05:05:02)
I love that by the way.
Lex Fridman
(05:05:03)
Yeah, Linux is my first and last love, it definitely… The spirit of Linux and open source is a beautiful thing so I do think that when you have these large corporations, even when they try to do good, oftentimes the profit imperative just takes over and they can corrupt themselves and Microsoft has a long history of doing just that to themselves.
ThePrimeagen
(05:05:28)
Yeah.
Lex Fridman
(05:05:28)
That said, they’ve done, they have you could say for cynical reason because they want to seem like the good guy amongst developers, but they’ve done a lot to support open source. It’s just like, same with Meta, Meta has done insane amount-
ThePrimeagen
(05:05:43)
Yeah.
Lex Fridman
(05:05:43)
… to support open source. You can say, actually for that one, I don’t know if I can even make a financial or a cynical case for why Meta is open sourcing Llama and these-
ThePrimeagen
(05:05:55)
Yeah, that one’s confusing. It just seems great.
Lex Fridman
(05:05:56)
Maybe for hiring. But no, I think that’s legit, just an ethical, really powerful decision. And sometimes these companies, because they have a lot of cash, can make the right, do the right thing.
ThePrimeagen
(05:06:13)
Yeah. It’s a really positive way to look at it and I think that’s really nice.
Lex Fridman
(05:06:17)
Well, we should always be skeptical.
ThePrimeagen
(05:06:18)
Yeah, I mean because at the end of the day, companies, they’re not good, they’re not bad, they’re morally neutral. It’s the people that are running them, the decisions those people make that are really where the bad or the good comes from.
Lex Fridman
(05:06:28)
Another question, ask him if he knows how to milk a cow? I’ve already asked that. The answer is-
ThePrimeagen
(05:06:33)
No.
Lex Fridman
(05:06:34)
Oh, no, you don’t know.
ThePrimeagen
(05:06:35)
I’ve never milked a cow.
Lex Fridman
(05:06:36)
Never milked a cow.
ThePrimeagen
(05:06:37)
Almost been killed by a cow, but never milked a cow.
Lex Fridman
(05:06:40)
Did you ever ride a bull?
ThePrimeagen
(05:06:41)
No.
Lex Fridman
(05:06:42)
All right. Why male models?
ThePrimeagen
(05:06:46)
Okay, so I can explain that one. I will say something like, “I really dislike the color purple because the color purple makes me upset.” I don’t know, just something very benign. But then someone right afterwards will be like, “But why don’t you like the color purple?” And it’ll just be like… It’s just like Derek Zoolander. It’s just like I get done on a five-minute talk about it and then the next question’s like, “But seriously why though?” It was just like, “Why male models?”
Lex Fridman
(05:07:12)
Yeah. So that’s the Zoolander reference when there’s a long explanation why male models and he agrees and then forgets.
ThePrimeagen
(05:07:20)
Yep.
Lex Fridman
(05:07:24)
What is Ligma?
ThePrimeagen
(05:07:26)
I’ve died by Ligma quite a few times. So do you know the origin story of Ligma?
Lex Fridman
(05:07:30)
No.
ThePrimeagen
(05:07:31)
So Ninja, famous streamer, someone got him with Ligma and said like, “Oh,” something like, “Have you heard about Ligma?” And he was like, “No.” And he’s like, “Oh, Ligma balls.” And then after that Ninja got so hurt by getting had by that, he started banning anyone in chat who’s said the word Ligma or something like that. And so then if you don’t embrace the meme you get destroyed. So of course, gets destroyed and so then the whole goal is that, can people get me with Ligma? TJ did iladies. He’s like, “Oh, did you hear that E-girls got renamed to iladies?” And I just didn’t even see it coming. And I was just like, “What?” And he’s like, “iladies nuts on your face.” And then it’s just like, “Oh my gosh.” And then a pirate software has also got me like, “Oh, have you heard about Google SIMA,” which SEMA is a real product by Google>

(05:08:12)
And I’m like, “Oh yeah, I’ve heard about this. What is this again?” He’s like, “SIMA balls,” right? It’s just like, “Dang it,” how do I keep…? So I’ve just had it happen live on stream many, many times. I’ve died by Ligma the most.
Lex Fridman
(05:08:24)
Please ask him about the size of his dict.
ThePrimeagen
(05:08:28)
Okay, so that’s dict, that’s dictionary in Python.
Lex Fridman
(05:08:32)
Who doesn’t love dicts?
ThePrimeagen
(05:08:34)
Yeah, that’s a great question. Just a dicts party when you use Python.
Lex Fridman
(05:08:38)
I love dicts.
ThePrimeagen
(05:08:39)
That should be a T-shirt. That’s actually a hilarious T-shirt. So on Stack Overflow, you can ask any question you want, and I decided to craft a question one day on Stack Overflow that says how to measure your dict and bytes. And then I proceeded to really go to town and explain all the different things like, “Well, what about the cost of the strings and the references?” And when you really get both hands on your dict and really go after it, it’s very hard to, really threw in some innuendos. The Stack Overflow team deleted the question, and then someone hand wrote me an email explaining why they deleted the question and complimented me on how thoroughly and thoughtful the question was just to weave in innuendos and that the entire team was impressed, but it’s inappropriate and it had to be deleted and don’t do it again or we’re going to ban your account.

(05:09:31)
And so it was a very funny moment and so I was like, “Oh, that’s funny that happened.” That was about six years ago. Last year I was at a conference and there’s a guy wearing a Stack Overflow name tag and I was like, “Oh, you work at Stack Overflow?” He’s like, “Oh, yeah, I do.” I’m like, “Do I got a story for you.” And he goes-
Lex Fridman
(05:09:48)
Oh, no.
ThePrimeagen
(05:09:49)
… “No, wait a second. Are you the dict guy?”
Lex Fridman
(05:09:51)
Yeah.
ThePrimeagen
(05:09:51)
That was his only question was that. And I was just like, “Let’s go.” I didn’t even say anything about me and he already knew immediately I was the dict guy.
Lex Fridman
(05:10:01)
I should say in all seriousness, I think I’ve had a bunch of conversations in the Python world where I would have to mention the name of this data structure and it makes me uncomfortable every time.
ThePrimeagen
(05:10:10)
It’s a very unfortunate shortening of a word.
Lex Fridman
(05:10:13)
Dict. It’s just like when I go to the hardware store and ask for caulk and there’s always a nice old lady and I ask her where to find, and it’s very uncomfortable. I try to pronounce it as hard as I can.
ThePrimeagen
(05:10:27)
Really get that L in there, like caulk.

God

Lex Fridman
(05:10:30)
Caulk, just to be clear. And try to avoid eye contact the whole time. You said that God was a big part, was a big part of your life. Can you speak to that a little bit more? Who is God and what effect, what role did he play in your life?
ThePrimeagen
(05:10:47)
So I did talk about that one important evening where I, for whatever reason, gained my conscious that moment. So obviously for me that I grew up with a life where I would probably argue myself as a functional atheist. I went to church a handful of times. I can’t quite really remember actually going to church as a family in any sort of sense. So there wasn’t some super strong tie or anything like that to it. Pretty much anyone else growing up in America in the ’90s, you had some sort of impact or intersection with church at some point in your life, that was just a very normal thing I would probably say. And so when that happened, it was a fairly big surprise for me. I wasn’t necessarily going that direction or deciding to do any of those things.

(05:11:33)
And so for me, it’s obviously the turning point of my entire life. I cannot speak to who I would be now without that. I can just tell you that I wouldn’t have had the drive. I probably would not have completed college. I would’ve not have found my wife or had my kids. I wouldn’t know how to value people. I don’t think without that whole thing, my value for people would’ve been very, very small because I would’ve continued to just objectifying in the way I was. And then probably the biggest thing is there’s this one verse, I don’t even know where it’s at, it effectively says that we love because he first loved us. And so for me it’s like I don’t think I would’ve ever lived a life that was happy without this. And I just didn’t even know that that was an option for me.

(05:12:22)
And I never really, it was a very tough set of years for me and I was very, very sad and just always just constantly looking for something to fulfill me. And so it’s like I didn’t have any confidence, I didn’t have any joy. I felt very sad. And so that was this moment where for the first time ever didn’t, all of a sudden I just felt like I didn’t have to live up to a standard. The standards have already been paid for, everything’s already, that’s the free gift, that’s the exchange. And so it’s just like for the first time, I didn’t have to be the cool guy, I didn’t have to have all the right words, I didn’t have to feel, I didn’t have to go on the conquest, the sexual conquest to find validation. I didn’t have to do any of those things and it was exceptionally liberating.

(05:13:12)
And so who is God? That’s more of a catechism question perhaps. What is man, who is God? Those are much harder questions. I believe that anytime you try to get too deep into describing who God is, you typically fall into Christian heresy. But for you-
Lex Fridman
(05:13:29)
He gave you a chance to be happy.
ThePrimeagen
(05:13:31)
Yeah. He gave me a chance not just to be happy, but also made it so that the first time I can actually feel forgiven I guess in some sense, and able to forgive people that hurt me. For a long time I had this weight I’d carry around from the things I hated about high school and all that kind of stuff. And through that experience, I just wrote down every last person’s name and actually held them with me for quite some time and this was the list of people I forgave. And I read it a few times. I couldn’t let myself be angry or consumed by that kind of stuff because hate is so sticky, it sticks for a lifetime. And there really is only one cure for hate, which is forgiveness. I just don’t think you can get rid of it without that.

(05:14:17)
And so I just had choose to forgive these people and to move on, and it really kind of freed me. And I would never have thought forgiveness as a means for that change if I didn’t first experience it myself.
Lex Fridman
(05:14:30)
What’s the role of love in the human condition, to go to the philosophical, and what’s been the role of love in your life?
ThePrimeagen
(05:14:40)
It’s very obvious that every person wants or desires love. My wife has recently convinced me to watch Love is Blind with her one time. And you watch the show and if you’re not familiar with it, it feels like just a disaster of an experiment to just cause crazy filming. But anyways, the idea is that if you just don’t see somebody, you can fall in love with somebody and want to marry them after 10 days or some very small period of time. And what you really end up seeing is all these people who are just desperate for actually love. And there’s some part of it… I told my wife, “It’s like love gladiators.” We’re watching people battle it out for drama and really what they want is love. And it’s like they’re fighting to the death and love, if you will. And it’s this almost kind of sad aspect to watch.

(05:15:28)
And so I think that it’s hard to call, what is its role in the human experience, because I think it’s just something that we all naturally not just want, but need. And I don’t think that you can really progress, and when I say the word love, I would like to kind of narrow it down maybe a bit more. And I don’t mean Eros, the Greek word like sexy love, I think that paternal and friendship love are extremely important. And I think agape, God love is also very important. Agape love is the one that is superior to them all, but obviously different and also co-needed with the parental ones and all that. And so you kind of need this mixture of them all, and each one is different for each reason and where it’s applied.

(05:16:15)
And so I don’t think… I just don’t see a world in which is good of any kind without that as a very foundational piece. Because again, I didn’t come here trying to quote any sort of scripture, but it says that it’s not the nails that hung on there, it’s love. That’s the reason why these things happen. And so if forgiveness is the requirement to kind of pay off hate in some sense, then love has to be the motivation for forgiveness.
Lex Fridman
(05:16:48)
Yeah, that’s the tragic aspect of life. I think there’s a deep loneliness in all of us and a longing to be a part of this bigger thing. And that longing is a love and it has many names, but yeah. Yeah, the love aspect of it, it’s the beautiful aspect of life, the tragedies, the loneliness, and the unfortunate suffering that is a fundamental part of life and the beautiful aspect is the love.
ThePrimeagen
(05:17:22)
Yeah.
Lex Fridman
(05:17:24)
Which I think is a good time to mention more Reddit, the place for everlasting positivity and love. Somebody wrote, “Please thank him, you, for his everlasting positivity and give him a big hug for me.” So I won’t give you a big hug on camera because I’m afraid I’ll get a boner and that’ll be very unfortunate.
ThePrimeagen
(05:17:51)
Hey, let’s not bring dicts into this again, okay?
Lex Fridman
(05:17:53)
It’s my favorite data structure. Like I said, I love dicts, all kinds of dicts, ordered dicts-
ThePrimeagen
(05:18:00)
Unordered.
Lex Fridman
(05:18:01)
… unordered dicts. I don’t discriminate. Yeah, but just that to say big thank you from me. I listen to you a lot and I just really enjoy… I’ve been going through a lot of shit myself and just the positivity, even when you’re building the stupidest shit, it’s just the positivity radiates from you and you inspire me to be a good person. You inspire me to build stuff. So thank you. And I’m sure there’s many, many others who listen to you for the same reason. So thank you for your positivity. Thank you for being the light in many people’s lives, and thank you for talking to me, brother.
ThePrimeagen
(05:18:43)
Dang. That was very, very kind. I really do appreciate all those extremely nice words even from Reddit. That’s very surprising. But thank you. I mean, I know you know that there’s many people’s lives, and I’m sure you’ve received the letters that have been changed from actions and things you’ve said and things you’ve done. And so it’s one of the best parts about doing this side is that you get a chance to potentially improve somebody’s life. And you getting to interview a lot of people, there’s a lot of people that listened to Chris Latner and saw his excitement for Swift and probably went and learned Swift and then got really amazing jobs and it can be all origined back to back to you in that interview. And so those are amazing things. And so same goes back to you, you’ve done a lot of good stuff.
Lex Fridman
(05:19:34)
Right back at you brother. Thank you for talking today. Thanks for listening to this conversation with Michael Paulson, aka ThePrimeagen. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Paulo Coelho. “When we strive to become better than we are, everything around us becomes better too.” Thank you for listening and hope to see you next time.

Files for Narendra Modi Episode | Lex Fridman Podcast #456

Captions & Transcripts:

You can download captions & transcripts in English, Hindi & Russian here:
Dropbox Link

Videos:

If you need the video version with Original, Hindi, or Russian audio tracks for creating clips for YouTube or for any other video platform, the currently best-available method I know for sharing is hosting them on YouTube as unlisted videos. Hopefully these are useful for you:

The main video for watching and sharing (where all languages are available) is:
https://www.youtube.com/watch?v=ZPUtA3W-7_I

Transcript for Narendra Modi: Prime Minister of India – Power, Democracy, War & Peace | Lex Fridman Podcast #460

This is a transcript of Lex Fridman Podcast #460 with Narendra Modi.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

In this episode…

Narendra Modi
(00:00:00)
My strength lies not in my name, but in the backing of 1.4 billion Indians and thousands of years of timeless culture and heritage. So wherever I go, I carry with me the essence of thousands of years of Vedic tradition, the timeless teachings of Swami Vivekananda and the blessings, dreams, and aspirations of 1.4 billion Indians. When I shake hands with the world leader, it’s not Modi, but 1.4 billion Indians doing so. So this isn’t my strength at all. It is rather the strength of India. Whenever we speak of peace, the world listens to us, because India is the land of Gautama Buddha and Mahatma Gandhi, and Indians aren’t hardwired to espouse strife and conflict. We espouse harmony instead.

(00:01:06)
We seek neither to wage war against nature, nor to foster strife among nations. We stand for peace, and wherever we can act as peacemakers, we have gladly embraced that responsibility. My early life was spent in extreme poverty, but we never really felt the burden of poverty. You see, someone who is used to wearing fine shoes will feel their absence when they don’t have them, but for us, we had never worn shoes in our lives. So how would we even know that wearing shoes was a big deal? We weren’t in a position to compare. That’s just how we lived.

(00:01:56)
When I became Prime Minister, I especially invited Pakistan to my swearing-in ceremony so we could turn over a new leaf, yet every noble attempt at fostering peace was met with hostility and betrayal. We sincerely hope that wisdom prevails upon them and they choose the path of peace. I believe even the people of Pakistan long for peace. Look, regarding what you said about criticism and how I deal with it. If I had to summarize in one sentence, I welcome it. I have a strong belief that criticism is the soul of democracy. I want to tell all the young people the following. No matter how dark the night may seem, it is still just night, and morning is bound to come.

Introduction

Lex Fridman
(00:03:07)
The following is a conversation with Narendra Modi, the prime Minister of India. It was one of the most moving conversations and experiences of my life. Allow me here to say a few words about it. Please skip ahead straight to our conversation, if you like. Narendra Modi’s life story is incredible. He rose from poverty to lead a nation of 1.4 billion people, the biggest democracy in the world, where he won epic-scale elections for Prime Minister three times. As a leader, he fought for ideas that unite his nation of India, a nation that is composed of a large number of highly varied and disparate cultures and peoples, who have a long history marked by religious, social, and political frictions. He is known for taking decisive, at times controversial actions for which he is loved by hundreds of millions of people, and is also criticized by many. We discuss all of this at length in this conversation. On the world stage, he is respected as a peacemaker and friend by most major world leaders, even those whose nations are at war with each other, from the United States to China, to Ukraine and Russia, to Israel, Palestine and the Middle East, and everywhere else. Now, at this moment in history, it is clear, at least to me, that the flourishing of human civilization hangs in the balance, with several wars on the brink of escalation to regional and even global conflict, rising tensions between nuclear powers, technological developments from AI to nuclear fusion that aim to completely transform society and geopolitics as we know it, and of course, generally increasing political and cultural turmoil.

(00:05:05)
So now more than ever, we need great leaders, great peacemakers who build bridges, not destroy them, who may preserve the identity of their nations, but still celebrate the common humanity of all of us, all people on earth. For this and many other reasons, this conversation with Prime Minister Modi was one of the most remarkable I’ve ever had. You may hear such words and think that I’m just enamored by power or access. No, never was, never will be. I do not idolize anyone, especially those in power. I’m generally skeptical of power, money, and fame because of their natural corrupting influence on the mind, the heart, the soul of a person.

(00:05:57)
The whole point of all the conversations I’ve had in my life, on mic and off mic, is that I try to see and explore the full complexity of every human being, the good and the bad. I believe we’re all the same in a deep fundamental sense, all capable of good, all capable of evil, all carry stories of pain and stories of hope. Whether you’re a world leader or a truck driver, a coal miner, or a farmer in the American Midwest. And by the way, I will be talking to a lot of the latter kind of folk this year off mic, and maybe even on mic, as I travel the US and the world. My brief statements here about Narendra Modi are about both him as a leader, and especially him as a human being. In the extensive time I spent with him, I spoke with him off mic and on mic.

(00:06:56)
It was a deeply personal human interaction characterized by warmth, kindness, humor, inner and outer peace, and absolute focus on the conversation between us in the present moment as if nothing else existed. I have heard from many people that he treats everyone he meets in this empathic way, no matter where they come from or what their position is in this world. So for those and many other reasons, this really was an incredible experience. I will never forget. Oh, and by the way, we make captions and voiceover audio tracks available in English, Hindi and other languages. You can also listen to the original mixed language version where I speak English and Prime Minister Modi speaks Hindi. Separately, you can choose to turn on subtitles in your preferred language.

(00:07:52)
On YouTube, you can switch between language audio tracks by clicking the settings gear icon and clicking audio track, and then selecting the language you prefer. For fully English overdub, select English. For fully Hindi overdub select Hindi, and to listen to the original mixed language version where I speak English and Prime Minister Modi speaks Hindi, please select Hindi Latin audio track, so you can listen either to a version that is all one language or to the original mixed language version with subtitles in your preferred language. The default is English overdub. Our thanks to ElevenLabs and a great team of translators, we do our best to bring the Prime Minister’s voice to life with AI voice cloning in English.

(00:08:41)
I promise that we will continue to work very hard to break down the barriers that language creates, and try to make these conversations as accessible as possible to everyone in the world. Anyway, let me pause one more time to say a big thank you. What a wild ride this life has been. It’s an honor for me to be on it with all of you. I love you all. This is the Lex Fridman podcast, and now, dear friends, here’s the Prime Minister of India, Narendra Modi.

Fasting

Lex Fridman
(00:09:19)
So, I should also say I’m fasting right now. It’s been almost two days, 45 hours, so just water, no food in honor of this conversation, just to get in the right mindset, get into the spiritual level. I’ve read that you often fast for many days. Can you explain why you fast, and where does your mind go when you fast?
Narendra Modi
(00:09:42)
First of all, I’m truly pleasantly surprised and honored that you’re fasting, all the more because it feels like you’re fasting as a tribute of respect for me. So, I express my deepest gratitude to you for doing this. In India, our religious traditions are actually a way of life. Our Supreme Court once gave a brilliant interpretation of Hinduism. They have stated that Hinduism is not about rituals or methods of worship, but rather it’s a way of living, a philosophy that guides life itself. And in our scriptures, there is deep discussion on elevating the body, mind, intellect, soul, and humanity. They outline various paths, traditions, and systems to achieve this, and fasting is one of them, but fasting alone isn’t everything. In India, whether you see it culturally or philosophically, sometimes I see that fasting is a way to cultivate discipline.

(00:11:28)
If I put it in simple terms, or explain it to the viewers who are unfamiliar with India, it is a powerful tool to bring both the inner and outer self into balance. It shapes life in profound ways when you fast. You may have noticed, as you said, you’ve been fasting on water for two days. Every single one of your senses, especially smell, touch and taste, becomes highly sensitive. You may even notice the subtle aroma of water itself, something you probably never noticed before when drinking it. If someone walks past you carrying tea, you will catch its aroma just like you would with coffee. A small flower you’ve seen before, you will see it again today, but now you can perceive its details more vividly. Your senses become extra sharp, highly aware, and fully tuned in, and their capability to observe and to respond multiplies, and grows sharper. I have personally often experienced this. Another thing I’ve experienced is that fasting can greatly accelerate the thinking process, and can give a fresh perspective. You start thinking outside the box. I don’t know if everyone experiences this, but I certainly do. Most people assume that fasting simply means giving up food or not eating, but that’s just the physical aspect of fasting. If someone is forced to go without food due to hardship with an empty stomach, can we call that fasting? Fasting is actually a scientific process. Whenever I fast for an extended period, I prepare my body in advance. For five to seven days before the fast, I follow various Ayurvedic practices and yoga practices, along with other traditional cleansing methods to internally reset my system.

(00:14:07)
Before actually beginning the fast. I make sure to drink a lot of water, as much as possible. So, you could say that this detoxification process helps prepare my body in the best possible way. And once I begin fasting, for me, it’s an act of devotion. For me, fasting is a form of self-discipline. For me personally, even while I go about doing my daily activities during a fast, my mind remains deeply introspective and focused inward, and that experience is profoundly transformative for me. My practice of fasting didn’t come from reading books, listening to sermons, or following a tradition just because my family happened to have observed it. It came from my own personal experience.

(00:15:10)
During my school days, there was a movement inspired by Mahatma Gandhi’s vision, his vision of cow protection. The government had not enacted any laws at the time. At the time, people across the country observed a one-day fast by gathering in public places in silent protests. We were just kids, probably had just finished primary school. Something inside me said, “I should be part of this,” and that was the first time in my life I experienced fasting. At such a young age, I felt neither hunger nor any desire for food. Instead, I felt a new awareness, a surge of energy within me.

(00:16:04)
So, I became convinced that fasting is a science far beyond just skipping meals. It is something far greater than that. Then gradually, I refined my body and mind through various experiments. Over time, it became a long and disciplined journey for me, and one thing is certain. Fasting never slows me down. I work just as much as usual. Sometimes I even work more. And another fascinating thing I’ve noticed is that when I need to express my thoughts, I’m amazed at where they come from and how they flow. It’s truly an incredible experience.
Lex Fridman
(00:16:58)
So you still do meetings with world leaders, you still manage the affairs of India. You still carry out your role as a leader on the world stage, all fasted, and sometimes nine days?
Narendra Modi
(00:17:14)
Well, this practice has a long historical context. I hope it may be interesting for those listening. There is an ancient tradition in India called Chaturmas. During the monsoon season, we know that the digestion tends to slow down, and so in this season, many people in India follow the practice of eating only a single meal within 24 hours. For me, this starts around mid-June and goes on until after Diwali around November. For about four to four and a half months, I follow this tradition of eating only once in 24 hours. Then comes the Navratri Festival in India, which usually falls in September or October. During this time, the whole country celebrates Durga Puja, a festival of strength, devotion, and spiritual discipline. This lasts for nine days.

(00:18:17)
During this time, I completely abstain from food and only drink hot water. Although drinking hot water has always been a part of my daily routine, my past lifestyle was such that I naturally developed this habit over time. Then in March or April, another Navratri occurs called Chaitra Navratri. This year, it will likely begin around March 31st. During this nine-day fast, I eat only one specific fruit, once a day. So for those nine days, if let’s say I choose papaya, then for all nine days, I won’t touch anything else. Just papaya. That too, I eat only once a day. That’s how I follow my nine-day fasting routine. So there are numerous fasts I keep throughout the year, and this has become a deeply ingrained tradition in my life. Perhaps I can say that I’ve been following these practices for 50 to 55 years.
Lex Fridman
(00:19:25)
Has there been times when you met with a world leader and completely fasted, and maybe what do they think about that? What do they think about your ability to do that kind of thing? And you’re right, I should mention that from even my two days, my ability to be present, my ability to sense everything, sharply focus on this experience is elevated. But yes, is there stories with a world leader that maybe jumped to mind when you were fasted?
Narendra Modi
(00:19:55)
Well, most of the time I don’t even let people know about it. It’s my personal matter, so I never publicized it, but people gradually started finding out. It became known only after I became Chief Minister and Prime Minister. Otherwise, it was purely personal. But now that it’s out in the open, I don’t mind sharing. If someone asks, I tell them so it might be useful to them, because it’s not my personal property. It’s my experience, and if that can help someone, why not share it? After all, my life has always been devoted to the well-being of others.

(00:20:34)
For example, after I became Prime Minister, I had a bilateral meeting at the White House with President Obama, and he had also arranged a formal dinner. Then as discussions between the two governments progressed, someone said, “Please join us for dinner.” To which another replied, “But the Prime Minister doesn’t eat.” This left them a bit concerned. How do you host the leader of such a major nation at the White House without serving food? When we sat down, they brought me a glass of hot water. I turned to President Obama and jokingly said, “Look, my dinner has arrived,” as I placed the glass in front of me. Later when I visited again, he still remembered. He smiled and said, “Last time you were fasting, this time we’re having lunch. Since you’re not fasting, you’ll have to eat twice as much.”

Early life

Lex Fridman
(00:21:38)
Let’s go to the beginning. You rose from humble beginnings to lead the world’s largest democracy. So, I think there’s a lot of people for whom this is truly inspiring. Your family was a very modest means, and you grew up in a one-room house with a mud floor, your whole family living there. Tell me about your childhood. How did those humble beginnings shape your outlook on life?
Narendra Modi
(00:22:07)
My birthplace is in Gujarat, specifically in North Gujarat, in Mehsana district, in a small town called Vadnagar. Historically, this town holds great significance, and so Vadnagar is where I was born and completed my early education. Looking at the world as I understand it today, I can reflect on my childhood and the unique environment I grew up in. My village had certain fascinating aspects, some of which are quite rare, even globally. When I was in school, there was an elder in our village who would regularly tell students, “Listen, kids, wherever you go, if you find a carved stone, or you find a stone with inscriptions on it or anything with engravings, bring it and place it in this corner of the school.” Over time, my curiosity grew and I started to understand. I realized that my village had a rich and ancient history. Discussions at school often revealed more fascinating details about its past. Later I learned that China even made a film about it. I had read in a newspaper about a film that mentioned the Chinese philosopher Hiuen Tsang, who had spent a considerable amount of time in my village, having arrived there many centuries ago. Back then, it was a major center for Buddhist learning. That’s how I first learned about it. And perhaps around the 1400s, it was a prominent Buddhist educational hub.

(00:24:22)
There was a victory monument from the 12th century, a temple from the 17th century, and in the 16th century, two sisters, Tana and Riri, who were renowned musicians. As I uncovered these history traces, I grasped the depth of our heritage. So when I became chief minister, I initiated large-scale excavation projects. The findings from these very projects confirmed that thousands of Buddhist monks had at one time studied there. It was a place where Buddhist, Jain and Hindu traditions co-existed harmoniously, and for us, history wasn’t just confined to books. Every stone spoke. Every wall had a story to tell, and so when we began the large-scale excavation work, we uncovered findings that hold immense historical significance.

(00:25:40)
So far, they have discovered evidence dating back 2,800 years, proving that this city has remained unbroken and eternal for all those 2,800 years. They have discovered solid proof of how its development unfolded over these centuries. Now, an international-level museum has been established there, open to visitors, especially for archeology students. It has become a major area of study. So the place where I was born holds its own unique historical significance. I see it as my good fortune. Some things in life unfold beyond our understanding. Kashi became my realm of duty. Now, Kashi is also eternal. Kashi, also known as Benares, or Varanasi, is an eternal city that has remained vibrant and alive for centuries.

(00:26:45)
Perhaps it was some divine design that led a boy born in Vadnagar to eventually make Kashi his realm of duty, living in the embrace of Mother Ganga. When I think about my family, my father, my mother, my siblings, my uncles, aunts, grandparents, we all grew up together in a small house. The place we lived was likely even smaller than where we are sitting now. There was no window, just a small door. That’s where I was born. That’s where I grew up. Now, when people talk about poverty, it’s natural to discuss it in the context of public life, and by those standards, my early life was spent in extreme poverty, but we never really felt the burden of poverty.

(00:27:50)
You see, someone who is used to wearing fine shoes will feel their absence when they don’t have them. But for us, we had never worn shoes in our lives. So how would we even know that wearing shoes was a big deal? We weren’t in a position to compare. That’s just how we lived. Our mother worked incredibly hard. My father, too. He was extremely hardworking, and he was also extremely disciplined. Every morning around 4:00 or 4:30 AM he would leave the house, walk long distances, visit several temples, and then reach his shop.

(00:28:39)
He wore traditional leather shoes, handmade in the village. The shoes were very tough and sturdy, making a distinct tock, tock, tock sound when he walked. People in the village used to say that they could tell the time just by hearing his footsteps. ” Oh, yes,” they would say, “Mr. Damodar is on his way.” Such was his discipline. He worked tirelessly, late into the night. Our mother too, ensured that we never felt the struggles of our circumstances, but despite everything, these challenging circumstances of living in scarcity never left a mark on our minds. I remember in school, the idea of wearing shoes never even crossed my mind.

(00:29:40)
One day, while I was on my way to school, I ran into my uncle on the way he saw me and was surprised, “Hey, you go to school like this, without shoes?” So at that time, he bought me a pair of canvas shoes and made me wear them. Back then, they must have cost around 10 or 12 rupees. But here’s the thing. They were white canvas shoes, and they would quickly get stained. So what did I do? In the evening, after school was over, I would stay back for a while. I would go from classroom to classroom, collecting leftover pieces of chalk that the teachers had discarded. I would take the pieces of chalk home, soak them in water, mix them into a paste and polish my canvas shoes with it, making them bright white again.

(00:30:49)
For me, those shoes were a treasured possession, a symbol of great wealth, and I don’t exactly know why, but from childhood, our mother was extremely particular about cleanliness. Perhaps that’s where we inherited that habit, too. Not sure how I picked up the habit of dressing neatly, but it’s been there since childhood. Whatever I wore, I made sure it looked proper. Back then, as you can imagine, we didn’t have any arrangements for ironing clothes. So instead, I would heat up water in a copper pot, hold it with tongs, and press my clothes myself. Then I’d head off to school. That’s how I lived, and I found joy in it. We never thought about being poor, or judged about how others lived or what their struggles were.

(00:31:45)
We lived carefree, enjoying whatever little we had and kept working hard. Never once did we complain about these things. And all these aspects of my life, whether you call it fortune or misfortune, unfolded in such a way in politics that they started coming to light. Because when I was taking my oath as Chief Minister, TV reporters went to my village, questioned my childhood friends, went to capture videos of my home. That’s when people started asking, “Who is this and what background he is coming from?” Before that, very few knew much about my life. That’s just how my journey has been. My mother possessed an innate spirit of caring for others’ well-being. It was woven into the very fabric of her being. She possessed knowledge of traditional remedies and healing practices, and would treat children with these home remedies. Every morning before sunrise, around five o’clock, she would start treating them, so all the children and their parents would gather at our home, little children crying, and we had to wake up early because of it.
Narendra Modi
(00:33:02)
And we had to wake up early because of it. Meanwhile, my mother would continue treating them with care. This spirit of service, in a way, was nurtured through these experiences. A sense of empathy for society, the desire to do good for others, these values were instilled in me from my family. I believe that my life has been shaped by my mother, my father, my teachers, and the environment I grew up in.

Advice to Young People

Lex Fridman
(00:33:33)
There’s a lot of young people listening to this that are truly inspired by your story. From those humble beginnings to the leader of the biggest democracy in the world, what can you tell to those young folks who are struggling, who are lost in the world, who are trying to find their way? What advice could you give them?
Narendra Modi
(00:34:00)
I want to tell all the young people the following. No matter how dark the night may seem, it is still just night, and morning is bound to come. That’s why we need patience and self-confidence. Yes, the challenges are real, but I am not defined by my circumstances. I am here for a purpose, sent by a higher power, and I am not alone. The one who sent me is always with me. This unwavering faith should always remain within us. Difficulties are a test of endurance. They are not meant to defeat me. Hardships exist to make me stronger, helping me grow and improve, not to leave me feeling hopeless or discouraged. Personally, I see every crisis, every challenge as an opportunity.

(00:35:12)
So, to all young people, I say, have patience. There are no shortcuts in life. At our railway stations, there hangs a sign for those who habitually cross the tracks instead of using the bridge, it reads, “Shortcut will cut you short.” I would tell young people the same, shortcut will cut you short. There are no shortcuts in life. Patience and perseverance are essential. Whatever responsibility we are given, we must pour our heart into it. We should live it with passion. Enjoy the journey and find fulfillment in it. I truly believe that if this mindset is cultivated, it transforms life. Similarly, abundance alone is not enough. There is no guarantee of success. Even a wealthy person who indulges in comfort and idleness will eventually wither away.

(00:36:09)
Instead, he must decide, “Yes, I may have resources around me, but I must use my abilities to grow them further. I must contribute more to society with my own strength. Even if I am in a good position, there is still so much more to do. Even if I am not in a good position, there is still so much work to do.” That’s what I believe. I have also noticed that some people tend to think, “I’ve learned enough. That’s it.” But one should never let the student within them die. Learning should never stop. I believe that as long as I am alive, I must have a purpose. Perhaps I exist to keep learning, to keep growing. Now, my mother tongue is Gujarati, and we were not very familiar with the Hindi language, nor did we know how to speak it eloquently or communicate effectively. But as a child, I used to sit at my father’s tea shop, and at that young age, I got the chance to meet so many people. And every time, I learned something from them, I observed their ways of speaking, their expressions. These things taught me a lot, even though I wasn’t in a position to apply it then, I thought, “If I ever get the chance, why not? Why shouldn’t I present myself well?” So, I believe the desire to learn should always remain alive. And another thing I’ve observed is that most people dream about achieving something or becoming someone. They set big targets and when they fall short, they feel disappointed.

(00:38:19)
That’s why whenever I get a chance to talk to my friends, I tell them, instead of dreaming about getting and becoming, dream of doing something. If you focus on doing something, and let’s say your goal is to reach 10, but you make it to eight, you won’t feel discouraged. You’ll still work toward 10 with determination. But if your dream is only to become something and it doesn’t happen, even your achievements may feel like a burden. That’s why we must adjust our mindset in life. Instead of thinking about what I got or didn’t get, the mindset should be what can I give? Because true contentment doesn’t bloom on its own. It grows from the depth of what you give.

Journey in the Himalayas

Lex Fridman
(00:39:16)
And I should say that this young kid, one of the things I’ve dreamed of doing is to do this very thing, to talk to you today. So, this is very surreal. At 17, another fascinating part of your life. You left home and spent two years roaming in the Himalayas, searching for purpose, for deeper truth, for God. So, not much is known about this period of your life. You lived a nomadic, minimalist existence, very much like a yogi, often sleeping without a roof over your head. What are some memorable spiritual moments, rituals, experiences from that time?
Narendra Modi
(00:39:59)
It seems like you’ve put in a lot of effort. Look, I don’t usually talk much about this, but I can share a few external aspects of it. I grew up in a very small town. Our life was all about being part of a community. We lived among people, surrounded by them. That was just how life was. There was a library in the village, and I used to go there often to read books. Whenever I read something from the books, I often found myself feeling inspired, thinking, “Why shouldn’t I shape my own life like that?” That desire was always there. When I used to read about Swami Vivekananda or read about Chhatrapati Shivaji Maharaj, I would often wonder, “How did they live? How did they build such remarkable lives?” And for that, I constantly experimented on myself. Most of my experiments were physical in nature, testing my body’s limits.

(00:41:20)
For example, where I lived, winters weren’t too harsh, but December nights could get quite cold. But still, at night, the cold would bite. It was natural. So, sometimes I would decide to sleep outside in the open with nothing to cover myself, just to see how my body endures the cold. So, from a very young age, I would often experiment with my body, and this became a regular thing for me. For me, going to the library, reading extensively, visiting the pond, washing the family’s clothes, and swimming became part of my routine. Swimming was my main physical activity. All of these things were deeply connected to my life. Later as I read Vivekananda, I became even more drawn to his teachings. One time I read about Swami Vivekananda, his mother was ill. So, he went to Sri Ramakrishna Paramahamsa for guidance. He would argue with him, debate with him. In his early days, he would often argue with him, questioning everything intellectually.

(00:43:10)
He said, “My mother is sick. If I were earning, I could take better care of her.” Sri Ramakrishna said, “Don’t bother me with all this. Go to Goddess Kali. She is there. Ask her for what you need.” And so, Swami Vivekananda went and sat before Goddess Kali’s idol for hours, and he immersed himself into deep meditation. After a few hours when he returned, Ramakrishna asked him, “So, did you request the goddess?” Swami Vivekananda replied, “No, I didn’t.” Ramakrishna said, “Go again tomorrow. She will fulfill your request. Ask her.” He went the next day and then again the day after. But each time, he found himself unable to ask for anything. His mother was unwell and he needed help, but when he sat before Goddess Kali, he was completely absorbed in her presence, and yet he could not bring himself to ask for anything.

(00:44:29)
Each time he returned empty-handed. He told Sri Ramakrishna, “I came back empty-handed. I didn’t ask for anything.” To stand before the divine goddess and not be able to ask for anything, that moment, that experience lit a flame inside him. There was a spark in his life. And from that came the spirit of giving. I believe that perhaps that small incident in Vivekananda’s life left an impression on me too. The thought of, “What can I give to the world?” Maybe true contentment comes from giving. If my heart is only filled with the hunger to receive, that hunger will never end. And within that realization came the idea of Shiva and living being as one. If you wish to serve Shiva, serve all living beings. Recognize the unity between the divine and the living. True non-duality is experience through this realization. I would often lose myself in such thoughts. My mind naturally drifting in that direction.

(00:45:58)
I remember an incident in the neighborhood where we lived, just outside there was a Lord Shiva temple. One day a saint came to stay there. So, that saint used to engage in meditation and spiritual practices. I started feeling drawn to him thinking perhaps he possessed some spiritual energy. I had only read about Swami Vivekananda, never seen such figures in real life. During Navratri, he was fasting and he had placed sorghum grains on his hand, a common tradition in our culture. In a way, sprouting seeds on your palms and sleeping like that for nine or 10 days, it was a kind of spiritual vow, and this saint was observing it. During those same days, my maternal uncle’s family was preparing for my aunt’s wedding. Everyone from my home was going to my uncle’s house for the wedding.

(00:47:07)
Now, for any child, visiting an uncle’s house is always exciting. But I told my family, “I’m not going. I’ll stay here and I will take care of Swamiji. Since he has these grains on his hand, he can’t eat or drink, so I will take care of him.” So, as a child, I chose not to attend the wedding. I stayed back serving Swamiji instead. Somehow my mind was naturally drawn in that direction. At times, whenever soldiers from my village came home during their holidays, they would walk around in their uniforms with such pride. I would run behind them all day thinking, “Look at them. They are serving the nation.” So, there was always a strong feeling inside me to do something meaningful. I didn’t fully understand what it would be, and I didn’t have a roadmap. There was a hunger within me, a deep longing to understand life, to explore its meaning. So, I just set out and began the journey.

(00:48:11)
During my time in the mission, I came across remarkable saints. They showered me with love and blessings. Among them, I formed a special bond with Swami Atmasthanandji. He lived for nearly 100 years, a life full of wisdom and service. In his final years, I deeply wished for him to stay with me at the Prime Minister’s residence, but his responsibilities were vast and he couldn’t come. However, back when I was chief minister, he used to visit and I was fortunate to receive his blessings and guidance. He once looked at me and said, “Why have you come here? You have a greater purpose to fulfill. Is your priority your own well-being? Or is it the welfare of society? Whatever Swami Vivekananda said was for the betterment of society. He said you are meant to serve others.” So, I remember feeling a bit disheartened at that moment. I had come seeking guidance, but all I got were words.

(00:49:20)
So, I continued on my journey, wandering from place to place. I spent time in the Himalayas embracing the solitude of the mountains. I met many remarkable individuals along the way. Some were great ascetics, people who had renounced everything, but still my mind remained restless. Perhaps it was my age of curiosity, of wanting to learn, to understand. It was a new experience, a world shaped by the mountains, by ice, by the towering snow-covered peaks. But all of this played a huge role in shaping me. It strengthened me from within and enabled me to discover my inner power. Practicing meditation, waking up in the sacred pre-dawn hours, bathing in the cold, serving people with devotion and naturally tending to elderly saints became a seamless part of who I was. Once, a natural calamity struck the region, and I immediately devoted myself to helping the villagers. So, these were the saints and spiritual masters with whom I stayed from time to time. I never remained in one place for long, I kept moving, constantly wandering. That was the kind of life I lived.

Becoming a monk

Lex Fridman
(00:50:45)
And for people who don’t know, that moment in the Ramakrishna Mission Ashram with the monk, Swami Atmasthananda, as you mentioned, he helped steer you towards a life of service. So, there’s another possible life that could have been where you take Sannyasa, you give away everything, and you’re a monk. So, we could have had a monk, Narendra Modi, and Prime Minister Narendra Modi, and he helped you take the decision to live a life of service at every scale
Narendra Modi
(00:51:21)
From the outside, people may call me a leader. Some call me the prime minister. Others call me the chief minister, and that’s how they see me from their perspective. But deep within, there is only an unwavering spiritual commitment. The Modi, who lovingly helped his mother care for children during their treatments, tending to them with patience and compassion. The Modi who wandered through the Himalayas. And the Modi who now works from this seat of responsibility. They are all tied together by the same inner consistency. Every action is dedicated to serving others. People may see a stark difference between a saint and a leader, but to me, there is no real difference. Yes, the attire changes, the way of life changes. The words spoken throughout the day shift and the nature of work evolves. But the core of my being remains unchanged, carrying out every responsibility with the same sense of calm, focus, and dedication.

RSS and Hindu nationalism

Lex Fridman
(00:52:33)
Another part of your life of who you are is you’ve spoken your whole life about putting your nation of India above all else. When you were eight, you joined the RSS, which espouses the idea of Hindu nationalism. Can you tell me about RSS and what impact they had on who you are and the development of your political ideas?
Narendra Modi
(00:52:58)
Ever since childhood, I always had the habit of staying engaged in something or the other. I remember there was a man named Makoshi. I don’t quite recall his full name. I think he was part of the service group, Makoshi Soni or something like that. He used to carry a small drum-like instrument called the Tambourine with him, and he used to sing patriotic songs in his deep, powerful voice. Whenever he came to our village, he would hold programs in different places. I would run after him like a crazy fan just to listen to his songs. I would spend entire nights listening to their patriotic songs. I enjoyed it. I don’t even know why, but I just did. In our village, there was a branch of the Rashtriya Swayamsevak Sangh, where we played sports and sang patriotic songs. Something about those songs touched me deeply. They stirred something inside me, and that’s how I eventually became part of the RSS. One of the core values that were instilled in us at RSS was, whatever you do, do it with a purpose. Even while studying, study with the goal of learning enough to contribute to the nation. Even when you exercise, do it with the purpose of strengthening your body to serve the nation. This is what we were taught. And today, RSS is a massive organization. It is now nearing its 100th anniversary. Such a massive volunteer organization likely doesn’t exist anywhere else in the world. Millions of people are connected to it, but understanding RSS is not that simple. One must make an effort to truly grasp the nature of its work. More than anything, the RSS provides you with a clear direction toward what can truly be called a purpose in life. Secondly, the nation is everything, and serving the people is akin to serving God.

(00:55:11)
This is what has been said since the Vedic Era. What our sages have said, what Vivekananda said and what the RSS echoes. A volunteer is told that the inspiration he gains from RSS is not just about attending the one-hour session or wearing the uniform. What matters is what you do for society. And today, inspired by that spirit, many initiatives are thriving. Like some volunteers established an organization called Seva Bharati. This organization serves the slums and settlements where the poorest people live, which they call service communities. To my knowledge, they run approximately 125,000 service projects without any government assistance, solely through community support. They spend time there, teach the children, care for their health, instill good values and work towards improving cleanliness in these communities. Running 125,000 social service projects is no small feat.

(00:56:20)
Similarly, some volunteers nurtured by RSS are dedicated to serving tribal communities through Vanvasi Kalyan Ashram. They live among the tribal people, working for their welfare. They have established over 70,000 one-teacher schools in remote tribal regions. There are also some people in America who show their support for this cause and contribute donations of about $10 or $15. And they say, “Skip a Coca-Cola this month. Don’t drink Coca-Cola and donate that money to a one-teacher school instead.” Now, imagine 70,000 one-teacher schools dedicated to educating tribal children. Some volunteers have founded Vidya Bharati to revolutionize education. Today they run nearly 25,000 schools educating around 3 million students, and I believe that millions of students have benefited from this initiative, receiving quality education at an incredibly low cost.

(00:57:28)
Alongside education, values are prioritized, and students remain grounded, learning skills so they don’t become a burden on society. That is in every aspect of life, whether it’s women, youth, or even laborers, the RSS has played a role. In terms of membership size, if I may say so, we have the Indian Labor Union. It has around 50,000 unions with millions of members across the country. Perhaps in terms of scale, there is no bigger labor union in the world. But what’s interesting is the approach they take. Historically, leftist ideologies have fueled labor movements worldwide. And what has been their slogan? “Workers of the world unite.” The message was clear. Unite first and then we’ll deal with everything else. But what do the labor unions run by RSS-trained volunteers believe in? They say, “Workers unite the world.” Others say, “Workers of the world unite.” And we say, “Workers unite the world.” It may seem like just a small shift in words, but it represents a huge ideological transformation.

(00:58:58)
The volunteers who come from the RSS follow their own interests, nature and inclination, and in doing so, they strengthen and promote these kinds of activities. When you observe these initiatives, you’ll see how over the past 100 years, the RSS has dedicated itself with the discipline and devotion of a seeker staying away from the glare of mainstream attention. I feel blessed to gain life’s values from such a sacred organization. Through the RSS, I found a life of purpose. Then I was fortunate to spend some time among the saints, which gave me a strong spiritual foundation. I found discipline and a life of purpose. And through the guidance of saints, I gained spiritual grounding. Swami Atmasthananda and others like him have held my hand throughout my journey, constantly guiding me at every step. The teachings of Ramakrishna Mission, Swami Vivekananda, and the service-driven philosophy of the RSS have played a crucial role in shaping me.

Explaining India

Lex Fridman
(01:00:17)
But they’ve also helped push the idea of India. What is the idea that unifies India? What is India as a nation? What is the foundational idea that unites all of these disparate worlds and communities and cultures? What would it be?
Narendra Modi
(01:00:40)
Look, India is a cultural identity. It is a civilization that dates back thousands of years. Consider the vastness of India, over 100 languages, thousands of dialects. India is so diverse that we have a saying that every 20 miles, the language changes, customs change, cuisine changes, even clothing styles shift from region to region. From the south to the north, you will see immense diversity across the country. But if you dig a little deeper, you will find a common thread. For example, the stories of Lord Ram can be heard everywhere in India. His name echoes in every corner of the country. But if you look closely from Tamil Nadu to Jammu and Kashmir, you will always find people whose names include Ram in some form. In Gujarat, he might be called Rambhai. In Tamil Nadu, Ramachandra, and in Maharashtra, Rambhau.

(01:02:07)
This unique cultural bond is what unites India as one civilization. Take something as simple as bathing in water. We have a ritual where all the rivers of India are remembered. They chant, ” I am bathing with the waters of all these rivers: Ganga, Yamuna, Godavari, Saraswati, Narmada, Sindhu, Kaveri.” It’s a sentiment that unites a nation, and we have a long tradition of making such resolutions at the beginning of important events and rituals. And the resolution itself can present a historical record. And in so doing serve as a way of collecting and preserving historical data. It has been an incredibly unique system, guided meticulously by our scriptures. When someone makes a resolution, performs a Puja, or even during weddings, we start by invoking the entire universe, beginning with Jambudweep, Bharatkhand, Aryavrat, and gradually narrow it down to the village, then mentioning the specific family, and finally, we invoke the family deity. This practice is still alive, happening daily in every corner of India, but sadly, the Western and global models began viewing nations only as administrative systems. India, however, has had a variety of administrative systems throughout history, many systems were fragmented, scattered and varied across regions. Kings and rulers were numerous, but India’s unity lay in cultural bonds. Pilgrimage traditions played a key role in preserving this unity. Shankaracharya established the four pilgrimage sites. Even today, millions of people travel from one place to another for pilgrimage. In Kashi, you’ll find people who bring water from Rameshwaram to Kashi and take water from Kashi to Rameshwaram. Even if you look at our Hindu calendar, you’ll find so many things across the country that you can’t even imagine.

Mahatma Gandhi

Lex Fridman
(01:04:27)
If we look at the historical foundation of modern India, along with yourself, Mahatma Gandhi is one of the most important humans to have ever lived, but certainly one of the most important humans to the history of India. What do you admire about Mahatma Gandhi?
Narendra Modi
(01:04:55)
As I had mentioned before, I was born in Gujarat, and Gujarati is my mother tongue. Mahatma Gandhi was also born in Gujarat. His native language was Gujarati too. He pursued a career as an attorney and lived overseas for several years. He had plenty of great opportunities, but the deep sense of duty within him, along with the values instilled by his family, led him to give up all comforts and devote his life to serving the people of India. He joined the struggle for India’s independence, and to this day, he continues to deeply influence the life of every Indian in some way. Mahatma Gandhi tried to live by his principles and practiced what he preached. For example, he strongly advocated for cleanliness and practiced it-
Narendra Modi
(01:06:03)
He strongly advocated for cleanliness and practiced it himself, and he made it a point to discuss cleanliness wherever he went. Another key factor to consider is India’s fight for independence. India was ruled by the Mughals, the British, and several other foreign powers. Despite being bound by the shackles of colonial rule for centuries, the flame of independence burned brightly in every corner and nook of India, never fading, always fueling the desire for freedom. Millions of people sacrificed their lives so the light of freedom could shine on India. They laid down their lives for freedom, sacrificing their youth behind prison walls. Mahatma Gandhi also fought for India’s independence, but in his own way, indeed the other freedom fighters were brave warriors and devoted sons of Mother India. They came, they fought and their martyrdom immortalized them, and they did indeed have a lasting impact.

(01:07:15)
But it was Mahatma Gandhi who awakened the nation, leading a mass movement fueled by truth and he wove even a sweeper into the very fabric of the freedom struggle. He told teachers their work was part of the freedom struggle. He told the people spinning thread and weaving clothes, they were freedom fighters. He told those tending to lepers that their service was a step toward India’s freedom. He viewed every task as a vital thread in the fabric of India’s independence movement, and this transformed India’s common man into a soldier in the quest for freedom. Gandhi forged a mass movement so immense that the British could never fully grasp it. The British never imagined that a pinch of salt from the Dandi March could spark a massive revolution, and he made it happen. And his life, presence, style, mannerisms all left a profound impact, and I have seen many of his stories evolve into timeless legends.

(01:08:24)
I recall an incident from a Roundtable Conference. Yeah, I believe he was attending a Roundtable Conference. He was supposed to meet King George at the Buckingham Palace draped in his breechcloth. Mahatma Gandhi made his way to the palace. Many people were amazed that he had showed up in that attire to meet the King. Gandhi remarked he didn’t need to wear a lot of clothes. He said, “Your king is wearing enough clothes for the both of us.” This was the whimsical charm of his nature. Mahatma Gandhi possessed many remarkable qualities. His call for unity and recognition of the people’s strength still resonates with me. In everything I do, I strive to include the common man and to ensure the participation of as many as possible. I don’t believe in leaving everything to the government. I am a firm believer in the power of social change.
Lex Fridman
(01:09:46)
So he was probably one of the greatest leaders of the 20th century. You are one of the greatest leaders of the 21st century. Those two centuries are very different, and you have been masterful in the game, in the art of geopolitics. So let me ask you, you have found a balance. So when negotiating on the world stage with super-powerful nations, is it better to be loved or feared? It seems like you are a masterclass of being loved by everybody, but everybody knows and feels the strength, so finding that balance. Can you speak to that balance?
Narendra Modi
(01:10:41)
First and foremost, I don’t think this is a fair comparison. Mahatma Gandhi wasn’t just a 20th-century leader. His relevance transcends centuries. Mahatma Gandhi’s legacy will last for centuries to come, and he remains relevant to date. As far as I am concerned, I have a responsibility to fulfill, yet the weight of that responsibility pales in comparison to my country. I am nowhere near as great as my country, and my strength lies not in my name, but in the backing of 1.4 billion Indians and thousands of years of timeless culture and heritage. So wherever I go, I carry with me the essence of thousands of years of Vedic tradition, the timeless teachings of Swami Vivekananda, and the blessings, dreams and aspirations of 1.4 billion Indians. When I shake hands with a world leader, it’s not Modi, but 1.4 billion Indians doing so. So this isn’t my strength at all. It is rather the strength of India.

(01:12:10)
You see, I recall something that happened all the way back in 2013. It was when my party declared I would be their prime ministerial candidate. My critics often tried to corner me on one point. It became a topic of widespread discussion, Modi is nothing more than a state leader. What does he know of foreign policy? Does he even understand global geopolitics? This was on everyone’s lips, and I was asked this question in every interview. I gave a very well-thought-out answer at the time. I said, “I won’t lay out my entire foreign policy in an interview, nor is it needed.” That said, India will neither allow itself to be looked down upon, nor will it ever look up to anyone. India will now see eye-to-eye with her counterparts. This was my belief in 2013, and it still lies at the heart of my foreign policy. For me, the country always comes first. However, to belittle someone or speaking ill of others is neither part of my cultural values nor my traditions. Moreover, our culture upholds and advocates for the welfare of mankind.

(01:13:45)
India has always championed the ideas of global peace and brotherhood. For centuries, we have envisioned the world as one big family. Our noble ancestors envisaged the welfare of the whole world and universe, and that’s why you must have noticed, the nature of our conversations as well as the ideas that I have presented on the global stage, which are rooted in respect and positivity. For example, I spoke about the environment in one of my speeches. I proposed the concept of one sun, one world, one grid. During the COVID pandemic, I delivered a speech at the G20 summit. I put forth the vision of one health, where humans and nature would live in harmony, and I have always worked towards this.

(01:14:47)
We hosted the G20 Summit with the motto, One Earth, One Family, One Future. We have inherited this timeless wisdom and it’s our duty to share it with the world. To give you an example, I have advocated for embracing renewable energy. We founded the International Solar Alliance with the motto, One Sun, One World, One Grid. Even when it comes to global healthcare, I had proposed One Earth, One Health. This initiative extends not only to humans, but also to all flora and fauna. I have always aimed to initiate efforts that foster global well-being, and the global community needs to join hands to accomplish that.

(01:15:40)
We must also understand that the world has become one small village today. No country can thrive in isolation. Today we all depend upon one another. No one can make it far by themselves. That is why you must learn to synchronize with everyone and everyone else must learn to synchronize with you. That’s the only way to propel this initiative forward. Organizations like the United Nations came into being after the First World War, but they failed to evolve with the times, and this inability to adapt has sparked a global debate on their relevance.

Path to peace in Ukraine

Lex Fridman
(01:16:23)
You have spoken about, you have the experience, you have the skill, you have the geopolitical leverage to be the biggest peacemaker in the world today, on the world stage, and there’s several wars going on. Can you maybe explain how you approach the process of making peace, helping make peace between two warring nations, for example, Russia and Ukraine?
Narendra Modi
(01:16:51)
Well, I represent the country that is the land of Lord Buddha. I represent the country that is the land of Mahatma Gandhi. These are the great souls whose teachings, words, actions, and behavior are entirely dedicated to peace. And that is why culturally and historically our background is so strong that whenever we speak of peace, the world listens to us. Because India is the land of Gautam Buddha and Mahatma Gandhi and Indians aren’t hardwired to espouse strife and conflict. We espouse harmony instead. We seek neither to wage war against nature, nor to foster strife among nations. We stand for peace and wherever we can act as peacemakers, we have gladly embraced that responsibility.

(01:18:03)
Returning to your example, I have a close relationship with Russia and Ukraine alike. I can sit with President Putin and say that this is not the time for war, and I can also tell President Zelensky in a friendly way that brother, regardless of how many people stand with you in the world, there will never be a resolution on the battlefield. The resolution will only come when both Ukraine and Russia come to the negotiating table. Ukraine may hold countless discussions with their allies, but it will bear no fruit. Discussions must include both parties instead. Initially, it was challenging to find peace, but now the current situation presents an opportunity for meaningful and productive talks between Ukraine and Russia. There has been a lot of suffering. Even the global south has suffered. The world has been grappling with a food, fuel and fertilizer crisis. So the global community should unite in the pursuit of peace. As for me, I have always maintained that I stand with peace. I am not neutral. I have a stance and that is peace, and peace is what I strive for.

India and Pakistan

Lex Fridman
(01:19:37)
Another difficult historic relationship and conflict is between India and Pakistan, it’s one of the most tense conflicts in the world, two nuclear powers with strong ideological differences. You are a great peacemaker. Looking out into the future as a visionary, what do you see as the path for friendship, for peace, for good relations between India and Pakistan?
Narendra Modi
(01:20:09)
I would like to delve into periods of our history the world may be unfamiliar with. Before 1947, during the struggle for independence, everyone was fighting side-by-side, shoulder-to-shoulder, and the nation was eagerly waiting to celebrate the freedom, the joy of independence. Now, we could have a lengthy discussion on what led to the events that unfolded, but the fact remains that the policymakers of the time agreed to India’s partition and they agreed to the Muslim side’s demand of carving out a separate nation. With hearts weighed down by grief and silent tears, Indians embraced this painful reality. However, what unfolded was an immediate heartbreaking saga of bloodshed. Trains filled with bloodied, wounded people and corpses started arriving from Pakistan. It was a harrowing sight. After getting their own way, we expected them to live and let live and yet, they chose not to foster a harmonious coexistence. Time and again, they decided to be at odds with India. They have waged a proxy war against us.

(01:21:57)
Don’t mistake this for ideology. What kind of ideology thrives on bloodshed and the export of terror, and we are not the sole victims of this menace. Wherever terror strikes in the world, the trail somehow leads to Pakistan. Let’s take the September 11th attacks, for example. The main mastermind behind it, Osama bin Laden, where did he eventually emerge from? He had taken refuge in Pakistan. The world has recognized that in a way terrorism and the terrorist mindset are deeply rooted in Pakistan. Today, it stands as an epicenter of turmoil, not just for India but for the world. And we have repeatedly asked them what good can come from this path? We have urged them to abandon the path of state-sponsored terrorism for good, “What do you hope to gain by surrendering your nation to lawless forces?” I even personally traveled to Lahore in the pursuit of peace.

(01:23:11)
When I became Prime Minister, I specially invited Pakistan to my swearing-in ceremony so we could turn over a new leaf. Yet, every noble attempt at fostering peace was met with hostility and betrayal. We sincerely hope that wisdom prevails upon them and they choose the path of peace. I believe even the people of Pakistan long for peace because even they must be weary of living in strife and unrest, they must have grown weary of relentless terror where even innocent children are killed and countless lives are destroyed.
Lex Fridman
(01:23:54)
Is there some memorable stories from your past attempts to try to improve relations with Pakistan that could guide the path forward into the future?
Narendra Modi
(01:24:09)
Like I mentioned, my first attempt at improving bilateral relations was when I invited my Pakistani counterpart to my swearing-in. It was a gesture of goodwill. It was a diplomatic gesture unlike any in decades. The very people who once questioned my approach to foreign policy were taken aback when they learned I had invited all SAARC heads of state and our then president, Mr. Pranab Mukherjee beautifully captured that historic gesture in his memoir. This was a testament to how clear and confident India’s foreign policy had become. This sent a clear message to the world about India’s commitment to peace and harmony, but we didn’t get the desired outcome.

Cricket and Football

Lex Fridman
(01:25:16)
Maybe to ask a little bit of a lighter question, who has the better cricket team, India or Pakistan? The two teams have an epic rivalry on the pitch and more seriously, given the geopolitical tensions that you spoke to, what role do sports and cricket and football play in fostering better relations?
Narendra Modi
(01:25:45)
I think sports have the power to energize the entire world. The spirit of sports brings people together across different nations. That’s why I would never want to see sports being discredited. I truly believe that sports play a major role in human evolution. They’re not just games, they connect people on a deeper level. Now, coming to the question of who’s better and who’s not, when it comes to techniques in sports, I’m not an expert. Only those who specialize in the technical aspects can judge which techniques are superior and who the best players really are. But sometimes the results speak for themselves. Just a few days ago, India and Pakistan played a match. The result reveals, which is the better team. That’s how we know.
Lex Fridman
(01:26:51)
Yeah. I’ve watched this series called The Greatest Rivalry, India versus Pakistan, that describes so many incredible players, so many incredible games. It’s always beautiful to see a great rivalry. You’ve also spoken about football. Football is very popular in India. So another tough question, who is the greatest football player of all time? We’ve got Messi, Pele, Maradona, Cristiano Ronaldo, Zidane, who do you think is the greatest football player to have ever played?
Narendra Modi
(01:27:22)
It’s absolutely true that many regions in India have a strong football culture. Our women’s football team is performing really well and the men’s team is also making great progress. But if we talk about the past, back in the 1980s, one name that always stood out was Maradona. For that generation, he was seen as a true hero, and if you ask today’s generation, they’ll immediately mention Messi.

(01:27:55)
Now that you’ve asked, another interesting memory just came to mind. There’s a state in India called Madhya Pradesh, right in the center. There’s a district called Shahdol, a completely tribal region where a large tribal community resides. I really enjoy interacting with people from such communities, especially the self-help groups run by tribal women. So I decided to visit them and have a conversation. But when I got there, I noticed something fascinating. Around 80 to 100 young boys, kids and even some older youth all dressed in sports uniforms standing together. Naturally, I walked over to them. So I asked them, “Where are you all from?” And they replied, “We’re from mini Brazil.” I was surprised and said, “What do you mean by mini Brazil?” They said, “That’s what people call our village.” Curious, I asked, “Why do they call it mini Brazil?” They explained, “In our village, football has been played for four generations. Nearly 80 national level players have come from here. Our entire village is dedicated to football.” They also told me when we host our annual football match, nearly 20,000 to 25,000 spectators come from nearby villages to watch.

(01:29:20)
I see the growing craze for football in India these days as a positive sign because it not only fuels passion, but also builds true team spirit.

Donald Trump

Lex Fridman
(01:29:32)
Yeah, football is one of the great sports that unites not just India, the whole world, and that just shows the power of what sport can do. You recently visited the United States and reinvigorated your friendship with Donald Trump. What do you like about Donald Trump as a friend, as a leader?
Narendra Modi
(01:29:55)
I’d like to share with you an event that stands out in my memory. Perhaps from that, you’ll get a better understanding of the point I’m trying to convey. For example, we had an event in Houston, Howdy Modi. Both President Trump and I were there and the entire stadium was completely packed. A massive crowd at an event in the US is a huge moment. While packed stadiums are common in sports, this was extraordinary for a political rally. The Indian diaspora had gathered in large numbers. Both of us delivered speeches and he sat down below listening to me speak. Now, that’s his humility. The President of the United States sitting in the audience while I spoke from the stage, that was a remarkable gesture on his part. After finishing my speech, I stepped down and as we all know, security in the US is extremely strict and thorough. The level of scrutiny there is on a completely different level. I went over to thank him and casually said, “If you don’t mind, why don’t we take a lap around the stadium? There are so many people here. Let’s walk, wave and greet them.”

(01:31:16)
In American life, it’s almost impossible for the President to walk into a crowd of thousands, but without even a moment’s hesitation, he agreed and started walking with me. His entire security detail was thrown off guard, but for me that moment was truly touching. It showed me that this man had courage. He makes his own decisions, but also he trusted me and my lead in that moment enough to have walked with me into the crowd. It was that sense of mutual trust, a strong bond between us that I truly witnessed on that day and the way I saw President Trump that day walking into a crowd of thousands without even asking security, it was truly amazing. And if you watch the video now, you’ll be amazed.

(01:32:15)
When he was shot during the recent campaign, I saw the same resilient and determined President Trump, the one who walked hand-in-hand with me in that stadium. Even after being shot, he remained unwaveringly dedicated to America. His life was for his nation. His reflection showed his America First spirit, just as I believe in nation first. I stand for India first and that’s why we connect so well. These are the things that truly resonate. And I believe that across the world politicians are covered so much by the media that people mostly perceive each other through its lens. People rarely get the chance to truly meet or personally know one another and perhaps third-party intervention is the real cause of tensions.

(01:33:20)
When I visited him in the White House for the first time, there was already a lot written about President Trump in the media. At that time, he was still new to office and the world had a rather different perception of him. Even I had been briefed in many different ways before meeting him, but to my surprise, the very moment I stepped into the White House, he broke all formal protocols right away. And then, he personally took me on a tour of the White House. As he showed me around, I noticed something striking, he wasn’t holding any notes or cue cards, nor was anyone accompanying him to assist. He pointed things out himself. “This is where Abraham Lincoln lived,” he said. He even explained why the courtroom was designed so long. He would point at the table and tell me which President signed here and on what date. I found that incredibly impressive. It showed how much he honored the Presidency and how respectful and deeply connected he was to America’s history. I could feel that. And he spoke to me freely, discussing many things openly. That was my experience from our first meeting.

(01:34:39)
Later, when his first term ended, and President Biden won, four years passed, but during that time whenever someone we both knew met him, and this must have happened dozens of times, he would say, “Modi is my friend, convey my regards.” That kind of gesture is rare. Even though we didn’t meet physically for years, our direct and indirect communication, our closeness and the trust between us remained unshaken.
Lex Fridman
(01:35:22)
He said that you’re a much tougher, much better negotiator than he is. He said this recently when you visited. What do you think of him as a negotiator and what do you think he meant about you being a great negotiator?
Narendra Modi
(01:35:39)
Now, that’s not something I can comment on. Since it’s his graciousness and humility, it is very kind of him that he openly appreciates me on various occasions and in different contexts. But about negotiation, I always put my country’s interests first. That’s why in every forum, I speak up for India’s interest, not to harm anyone but in a positive manner and because of that, no one takes offense. People know that if Modi is present, he will strongly advocate for these things. After all, the people of India have given me this responsibility. For me, my nation is my high command and I will always honor their will.
Lex Fridman
(01:36:30)
You’ve also had a bunch of productive meetings with several other folks on your visit to the United States, Elon Musk, JD Vance, Tulsi Gabbard, Vivek Ramaswamy. What are some things that stood out from those meetings? Maybe key takeaways, key memories.
Narendra Modi
(01:36:47)
Look, I can say this, I have observed President Trump both during his first term and now in his second run. This time, he seems far more prepared than before. He has a clear roadmap in his mind with well-defined steps, each one designed to lead him toward his goals. I also had the chance to meet members of his team, and I truly believe he has put together a strong, capable group and with such a strong team, I feel they are fully capable of implementing President Trump’s vision based on my interactions with them. I met several people, Tulsi Gabbard, Vivek Ramaswamy, Elon Musk, and there was a family-like atmosphere, everyone had come with their families. As for Elon Musk, I have known him since my time as chief minister. He was there with his family and children, so naturally the atmosphere felt warm and friendly. Of course, we had discussions and we talked about many different topics. Now, with his DOGE mission, he is incredibly excited about how it’s progressing and honestly, it makes me happy too because when I took office in 2014, I wanted to free my country from the deep-rooted issues and harmful practices that have crept in, and I’ll continue striving to eliminate as many of them as I possibly can. For example, after I took office in 2014, I observed that back then we weren’t part of many global discussions, not like how President Trump and DOGE are being talked about today. But let me give you an example so you can see the kind of work that was done. I noticed that the benefits of certain government schemes, especially welfare programs, were being exploited by so many people who never even existed in real life…
Narendra Modi
(01:39:01)
… who never even existed in real life. There were ghost names, pensions being issued to fake people. Widow pensions were being granted even before marriages took place, and disability pensions were given without any real disabilities. Then I launched a scrutiny process, and you’ll be shocked to know what we found. A hundred million people, a hundred million people, that’s a hundred million fake or duplicate names that I removed from the system. And because of that, we saved massive amount of money.

(01:39:40)
Then I introduced direct benefit transfer, ensuring that every rupee sent from Delhi reached the rightful person without leakage. As a result, my country saved nearly 3 trillion rupees that would have otherwise ended up in the wrong hands. Just because of direct benefit transfer through technology, we eliminated middlemen, ensuring transparency in the system.

(01:40:03)
I also introduced the GeM portal for government purchases, which has helped save both time and money. It has increased competition and improved quality. In India, we had an overwhelming burden of compliances. I eliminated 40,000 unnecessary compliances, and removed nearly 1500 outdated laws that served no purpose.

(01:40:29)
So in a way, my efforts have been about freeing governance from unnecessary dominance and inefficiency. And naturally, when bold changes happen, just like DOGE’s mission, they become a topic of discussion worldwide.

China and Xi Jinping

Lex Fridman
(01:40:52)
You and Xi Jinping have considered each other friends. How can that friendship be reinvigorated to help de-escalate some of the recent tensions, and resume dialogue and cooperation with China?
Narendra Modi
(01:41:06)
Look, the relationship between India and China isn’t something new. Both nations have ancient cultures and civilizations. Even in the modern world, they play a significant role. If you look at historical records, for centuries, India and China have learned from each other. Together, they have always contributed to the global good in some way. Old records suggest that at one point India and China alone accounted for more than 50% of the world’s GDP. That’s how massive India’s contribution was. And I believe our ties have been extremely strong, with deep cultural connections. If we look back centuries, there’s no real history of conflict between us. It has always been about learning from each other and understanding one another. At one time, Buddhism had a profound influence in China, and that philosophy originally came from here.

(01:42:40)
Our relationship should remain just as strong in the future. It should continue to grow. Of course, differences are natural. When two neighboring countries exist, occasional disagreements are bound to happen. Even within a family, not everything is always perfect. But our focus is to ensure that these differences don’t turn into disputes. That’s what we actively work toward. Instead of discord, we emphasize dialogue, because only through dialogue can we build a stable cooperative relationship that serves the best interests of both nations.

(01:43:35)
It is true that there have been ongoing border disputes between us. And in 2020, the incidents along the border created significant tensions between our countries. However, after my recent meeting with President Xi, we have seen a return to normalcy at the border. We are now working to restore conditions to how they were before 2020. Slowly but surely, trust, enthusiasm, and energy will return. But of course, it will take some time, since there’s been a five-year gap. Our cooperation isn’t just beneficial, it’s also essential for global stability and prosperity. And since the 21st century is Asia’s century, we want India and China to compete in a healthy and natural way. Competition is not a bad thing, but it should never turn into conflict.
Lex Fridman
(01:44:35)
The world is worried about a brewing global war. The tensions between China and the United States, in Ukraine, Russia and Europe, in Israel, the Middle East. What can you say about how we in the 21st century can avoid a global war, avoid an escalation towards more conflict, more war?
Narendra Modi
(01:45:07)
Look, COVID exposed the limitations of every nation. No matter how much we consider ourselves as a great nation, no matter how progressive we think we are, or how scientifically advanced we believe we’ve become, everyone has their own way of looking at things. In the end, we all found ourselves on the same ground. Every country in the world faced this reality. At that time, it felt like the world would learn from it, that we would move toward a more unified world. Just as a geopolitical order emerged after World War II, many thought something similar would happen post-COVID. But unfortunately, instead of moving towards peace, the world became even more fragmented, ushering in a period of uncertainty, and the wars have only made it worse.

(01:46:16)
I believe that modern wars are no longer just about resources or interests. Today I see so many kinds of conflicts happening. Physical battles often get discussed. Struggles are happening in every domain. International organizations that were once powerful have become almost irrelevant. No real reforms are happening. Institutions, like the UN, are failing to fulfill their roles. People, who disregard international laws and rules, continue to act freely, and no one can stop them. In such situations, the prudent choice for everyone is to let go of conflict and move toward cooperation. And a development-driven approach is the way forward. Expansionism will not work. As I’ve said before, the world is interdependent and interconnected. Every nation needs one another, no one can stand alone. And from all the different forums I attend, one thing is clear: Everyone is deeply worried about these conflicts. We can only hope that peace is restored very soon.

Gujarat riots in 2002

Lex Fridman
(01:47:59)
I’m not very good at this.
Narendra Modi
(01:48:00)
You keep looking at your watch.
Lex Fridman
(01:48:02)
No, no, no. I barely know what I’m doing, Prime Minister. I’m not very good at this. Okay. You’ve been… through your career and through your life, you have seen a lot of difficult situations in the history of India. One of them, the 2002 Gujarat riots, they’re one of the most challenging periods of modern Indian history, when there was violence between Hindu and Muslim citizens of the Gujarat that led to over 1000 deaths. It revealed the intensity of religious tensions in the region. You were, as you mentioned, chief minister of Gujarat at the time. Looking back, what lessons do you draw from that time? And we should also say that India’s independent Supreme Court upheld twice, in ’12 and ’22, that you had no involvement in the violence of the 2002 Gujarat riots. But I was wondering if you could speak to the broad lessons you draw from that time?
Narendra Modi
(01:49:07)
Look, regarding your first point, when you humbly said that you don’t know what you’re doing, that you’re not good at this, I disagree and personally feel you’ve put in tremendous care. You’ve done extensive research and have dived deeply into every small detail. So I think you’ve done very well, and all the efforts you’ve put in during our conversation and in all your conversations are appreciated. And rather than simply interviewing me, I feel you’re trying to deeply understand India. That’s why I strongly feel there’s genuine honesty in your sincere effort to uncover the truth. And for that sincere approach, I genuinely congratulate you.
Lex Fridman
(01:50:06)
Thank you.
Narendra Modi
(01:50:06)
Regarding the earlier events that you mentioned, like the 2002 riots in Gujarat, I’d like to paint you a clearer picture of the 12 to 15 months leading up to that, so you can fully understand the atmosphere of that time. For instance, take December 24, 1999, roughly three years earlier, an Indian flight from Kathmandu to Delhi was hijacked, redirected to Afghanistan and landed in Kandahar. Hundreds of Indian passengers were held hostage. It caused massive turmoil across India as people faced life and death uncertainty.

(01:51:02)
Then, in the year 2000, the Red Fort in Delhi was attacked by terrorists. Yet another crisis struck the nation, intensifying fear and turmoil. On September 11th, 2001, the Twin Towers in America faced a devastating terror attack, once again shocking the entire world. Because ultimately, the people behind these attacks are driven by a similar mindset. Then in October 2001, terrorists attacked the Jammu and Kashmir Assembly. Soon after, on December 13th, 2001, India’s parliament was targeted.

(01:51:41)
Within just eight to 10 months, these major global terrorist attacks took place, violent incidents that led to bloodshed and the loss of innocent lives. In such a tense environment, even the smallest spark can ignite unrest. The situation had already become extremely volatile. In such times, suddenly, on October 7th, 2001, I was given the responsibility of becoming chief minister of Gujarat. This was an enormous challenge.

(01:52:21)
At that time, Gujarat was recovering from a devastating earthquake, the largest of the previous century, which had left thousands dead. My first major task as chief minister was overseeing the rehabilitation of survivors. This was a crucial task, and from day one after my oath, I immersed myself in it. I was a person who had absolutely no prior experience with government. I had never been part of any administration, never even served in government before. I had never contested an election, never even been a state representative. For the first time in my life, I had to face elections.

(01:53:04)
On February 24, 2002, I became a state representative, an elected representative for the first time. And it was only around February 24th, 25th, or 26th that I stepped into the Gujarat Assembly for the very first time. On February 27th, 2002, we were seated in the Assembly for the budget session. And that same day, it had been just three days since I’d become a state representative, when suddenly the horrific Godhra incident occurred. It was a tragedy of unimaginable magnitude, people were burned alive. You can imagine, against the backdrop of incidents like the Kandahar hijacking, the attack on parliament, or even 9/11, and then to have so many people killed and burned alive, you can imagine how tense and volatile the situation was. Of course, this was tragic for everyone. Everyone prefers peace.

(01:54:27)
The perception that these were the biggest riots ever is actually misinformation. If you review the data from before 2002, you will see that Gujarat faced frequent riots. Curfews were constantly being imposed somewhere. Communal violence could erupt over trivial issues, like kite flying contests or even minor bicycle collisions. Before 2002, Gujarat witnessed over 250 significant riots. The riots in 1969 lasted nearly six months. So there was a long history, long before I was in the picture.

(01:55:15)
But that one tragic incident in 2002 became a sparking point, leading some people towards violence. Yet, the judiciary thoroughly investigated the matter. At that time, our political opponents were in power, and naturally they wanted all allegations against us to stick. Despite their relentless efforts, the judiciary analyzed the situation meticulously twice and ultimately found us completely innocent. Those who were truly responsible have faced justice from the courts.

(01:55:55)
But the most important thing is, in Gujarat, where riots used to happen some way or the other every year, but after 2002, in 22 years, there hasn’t been a single major riot in Gujarat. Gujarat remains completely peaceful. Our approach has always been to avoid World Bank politics. Instead, our mantra has been, together with everyone, development for all, trust from all, and efforts by everyone. We’ve shifted away from the politics of appeasement to the politics of aspiration. Because of this, anyone who wishes to contribute joins us willingly. We’ve continuously strived to turn Gujarat into a well-developed state. And today, Gujarat is actively contributing toward building a developed India as well.
Lex Fridman
(01:56:55)
A lot of people love you. I’ve got to hear from a lot of them. But there is also people who criticize you, including from the media. And folks in the media have criticized you over this 2002 Gujarat riots. What’s your relationship like with criticism? How do you deal with critics? How do you deal with criticism coming from the media or in your own inner circle or just in your own life?
Narendra Modi
(01:57:24)
Look, regarding what you said about criticism and how I deal with it, if I had to summarize in one sentence, I welcome it. I have a strong belief that criticism is the soul of democracy. If democracy truly runs in your veins, you must embrace it. In our scriptures it’s said, always keep your critics close. Critics should be your nearest companions, because through genuine criticism, you can improve quickly, and work democratically with better insights. In fact, I believe we should have more criticism and it should be sharp and well-informed.

(01:58:23)
But my real complaint is that nowadays what we see isn’t real criticism. Genuine criticism requires thorough study, in-depth research and careful analysis. It demands finding the truth from falsehoods. Today, people look for shortcuts, avoid proper research and skip deep analysis. Instead of identifying genuine weaknesses, they jump straight to accusations. There’s a big difference between allegations and criticism. The references you’re giving, they are allegations, not criticism. For a strong democracy, genuine criticism is necessary. Allegations benefit no one, they just cause unnecessary conflicts. That’s why I always welcome criticism openly. And whenever false accusations arise, I calmly continue serving my country with complete dedication.
Lex Fridman
(01:59:28)
Yeah, the thing you speak of is very important to me because I admire great journalism. And unfortunately, in modern day, a lot of journalists seek clickbait headlines, make accusations, because they operate under incentive, because they want the headline, the cheap shot. I think there is room and desire and hunger for great journalists, and that requires deep understanding. And it saddens me how often… I don’t think I’m very good at this, but one of the reasons I really wanted to talk to you is because I don’t see enough high effort, deep dive research. I don’t know how many books I’ve read. I’ve read a lot in preparing, just to experience, just to try to understand. It requires a lot of preparation, a lot of work, and I would love to see great journalists do that more. And from that place, you can criticize. From that place, you can really investigate the complexity of a situation, of people in power; their strengths, their flaws, the mistakes they have made. But that requires great, great, great preparation. So I wish there was more of that, of great journalism.
Narendra Modi
(02:00:46)
Yes. Clear, well-directed and specific criticism genuinely helps in the process of effective policymaking. It leads to a clear-cut policy vision. I specifically pay close attention to such constructive criticism. In fact, I actively welcome it.

(02:01:06)
Regarding your point about journalistic headlines, if someone’s attracted by catchy headlines or plays with words, I honestly don’t mind that much. But when there’s a deliberate agenda behind the actions and the truth is deliberately ignored, that can cause damage that lasts for decades. If someone focuses on pleasing their readers or viewers with attractive headlines, maybe we can compromise a little. But if there’s a hidden motive, or if things are intentionally twisted to serve an agenda, that’s a serious issue worth worrying about.
Lex Fridman
(02:01:50)
And in that, the truth suffers, I think.
Narendra Modi
(02:01:54)
I remember once in London, I was invited to give a speech at an event organized by a Gujarati newspaper there. So during my speech, I casually said, since it was an event attended by journalists, “What kind of journalism should we have? Should it be like a fly or a bee?” I explained, “A fly sits on dirt and spreads the dirt around, but a bee lands on flowers, collects nectar, and then shares that sweetness everywhere. Yet if something wrong occurs, the bee can sting so powerfully that you’d have to hide your face for three days straight.”

(02:02:56)
However, some people selectively picked up just half of my analogy and created a huge controversy out of it. Honestly, was I being negative about anyone? Not at all. I was simply highlighting the incredible strength of a bee, that even its small sting can leave such an impact that can make someone hide their face for days. You can’t show your face. That’s the power journalism should have. But unfortunately, some people prefer the fly approach instead.

Biggest democracy in the world

Lex Fridman
(02:03:34)
I now have a new life goal of becoming the bee. You mentioned democracy, and not knowing much about government until 2002, but from 2002 to today, you won eight elections that I could count. So many of the elections, over 800 million people vote in India. What does it take to win an election like that, and to win an election of 1.4 billion people where you get to represent those people, the biggest democracy in the world?
Narendra Modi
(02:04:17)
Well, I’ve been actively involved in politics for years. Before stepping into active politics, my focus was primarily on organizational work. This also included managing elections and strategizing campaigns, so that was where I dedicated my time. For 24 years, the people of Gujarat and India have placed their trust in me to lead with unwavering dedication and a deep sense of duty. I try to fulfill the sacred duty entrusted to me by the people I revere as divine. I remain committed to honoring their trust, ensuring it never falters. And they see me for what I truly am.

(02:05:18)
My government is committed to ensuring welfare schemes reach every citizen. Every scheme must reach its intended beneficiaries. Every beneficiary must be treated equally. No one should face discrimination on caste, creed, faith, wealth, or ideology. We must strive to ensure the well-being and prosperity of everyone. This way, even those not directly benefiting, never feel left out or treated unfairly. They find comfort in knowing they too will benefit in due course. This fosters a deep sense of trust, and trust is the cornerstone of my governance model.

(02:06:09)
Secondly, my governance flows beyond the ebb and flow of elections. My governance is rooted in the people, not the polls. It is committed to the well-being of my citizens and the greater good of the nation. As you may know, I had once set out on a quest for spiritual awakening, so now I revere my nation as the Divine itself, and I now revere the people as a manifestation of the Divine. Like a devoted priest, my heart is set on serving the people. I don’t distance myself from the people. I live among them as one of them. And I tell everyone I work with, if you work hard, I will work harder. People see this and it builds trust. Besides, I have no conflicts of interest. I have no friends or relatives who stand to gain from the position I hold. The common man appreciates this lack of vested interests, and that’s probably just one reason why.

(02:07:21)
Moreover, I come from a party with millions of dedicated volunteers. Volunteers who are completely devoted to the welfare of India and its citizens. They have had no stake whatsoever in politics. They’ve held no title, nor strayed where influence holds sway. My party is blessed with millions of volunteers who work tirelessly. I am proud to belong to the world’s largest political party, and it’s not like my party has been around forever. It reflects the hard work of millions of volunteers. Their selfless service is widely recognized and valued by the community. It fortifies people’s trust in the BJP, echoed in the election results. I never tallied our election triumphs, but we have had people’s blessings.
Lex Fridman
(02:08:18)
I was wondering if you could speak to the incredible logistics, that blew my mind, of running the elections in India. So there’s a lot of interesting anecdotes that arise, for example, that no voter should be more than two kilometers away from a polling station. The result of that is you have these stories, of voting machines having to be carried to remote regions of India, is really incredible. Just every single voter counts, and the machinery of having 600 plus million people vote. Is there some anecdote you could speak to that is particularly impressive to you? Or maybe you could speak generally to the logistics of what it makes to run an election that big, a democracy that big.
Narendra Modi
(02:09:10)
First and foremost, I am truly grateful for your insightful question. Anyone who believes in democracy should listen to what I’m about to share with you. We often discuss election results, but all the behind the scenes work is overlooked. Let’s take the recently concluded 2024 general elections as an example. There were 980 million registered voters. Each of those voters had a registered ID, and all the necessary details in a vast database. And this number is twice the entire population of North America. It even surpasses the total population of the entire European Union. Out of the 980 million registered voters, 646 million people ventured out, and braved the intense heat of May, determined to cast their votes. Temperatures soared to 40 degrees in some areas, yet they chose to go, and the size of this voter base is double the population of the United States.

(02:10:36)
We had more than a million polling booths. Can you fathom all the manpower involved in this? My country has more than 2,500 registered political parties. This staggering figure, with such a vast array of political parties, has the power to leave the entire world in awe. My country has more than 900 round-the-clock news channels. There are over 5,000 newspapers published daily. They each play a role in upholding democracy in their own way. Even the humblest villagers here embrace technology with remarkable swiftness. While some other countries take months to declare election results, we announced the results within a day, despite the mind-boggling number of voters.

(02:11:35)
And you were spot on in saying that even the remotest villages have polling stations. We even use helicopters to transport polling stations. I believe Arunachal Pradesh has the highest altitude polling booth on record. In Gujarat, a polling booth was set up in the Gir Forest for just one voter, who was in the middle of nowhere, but we ensured a polling booth was set up for them. All I mean to say…
Narendra Modi
(02:12:03)
… ensure the polling booth was set up for them. All I mean to say is that we truly leave no stone unturned in our unwavering commitment to strengthening democracy, ensuring that we are always fully prepared and election ready no matter the circumstances. I firmly believe that India’s Election Commission sets a global standard as a benchmark for conducting free and fair elections. It is the Apex decision-maker. This in itself is such a remarkable story that the world’s top universities should consider it a case study. They should analyze the management behind it as a valuable learning example. Given the sheer number of voters, can you truly grasp the immense depth of political awareness? All of this will make for an excellent case study for the younger generation worldwide.
Lex Fridman
(02:13:07)
To me, I love democracy. This is one of the main reasons I love the United States, but there’s just nothing quite as beautiful as democracy when it functions in India. Like you said, 900 million people registered to vote. It really is a case study. It’s beautiful to see that many people come together willingly, passionately casting a vote for some person to represent them like they’re putting their heart in that. It’s really important for a person to feel like their voice is going to be heard. It’s beautiful. Speaking of which, you are loved by a lot of people. You are one of the most powerful humans in the world. Do you sometimes think about whether this much power has a corrupting effect on your mind, especially across the many years that you’ve been in power?

Power

Narendra Modi
(02:14:03)
Well, I don’t think the word powerful quite reflects the journey of my life. I can never claim to be powerful. For all I am is a humble servant, I even identify myself as not the Prime Minister, but the prime servant, and service is the guiding principle of my work ethic.

(02:14:28)
As far as power is concerned, it is something I have never bothered about. I entered politics not to play power games, but to serve. Rather than seeking power, I stay committed to doing and getting work done. I am more focused on productivity than power. I have always dedicated myself to serving the people. I have always devoted myself to bringing about a positive change in their lives.
Lex Fridman
(02:15:04)
Like you mentioned, you work a lot. You give your whole soul to your work. Do you ever get lonely?
Narendra Modi
(02:15:20)
Look, I never experience loneliness because I am an ardent believer in the philosophy of one plus one, and this philosophy of one plus one is aligned with my moral compass, and whenever I am asked to elaborate on this outlook, I say, the first one represents Modi and the other one represents the Almighty. I am never alone for he is always there to keep me company. This is just how I function. Having wholeheartedly embraced the ideals of Swami Vivekananda, I firmly believe that service to mankind is service to God. For me, the nation itself is divine and mankind is a reflection of the divine. I walk this path with the conviction that serving the people is serving the divine. That’s why the very notion of struggling with loneliness has never even remotely crossed my mind.

(02:16:41)
Like at the time of pandemic, with the lockdown and traveling restrictions in place, I came up with a way to make the most of my time. I designed a governance model that operated seamlessly through videoconferencing. I kept myself occupied with remote work and virtual meetings.

(02:17:16)
Another thing I did was to connect with the people I had worked with throughout my life. Among my party’s volunteers nationwide, I made a list of those who were 70 and older. Some of the volunteers had very humble beginnings and came from very modest backgrounds. I personally called every volunteer aged 70 and above. I made it a point to inquire about their health and their family’s well-being. I inquired about how their area was coping. I asked these questions to make sure they were doing well. This allowed me to build rapport with them, and we would go down the memory lane. They were touched that the Prime Minister would check on them in the pandemic.

(02:18:08)
I made around 40 calls every day without fail. I carried on with this throughout the pandemic. It gave me the chance to reconnect and relive old memories with familiar faces. Loneliness is never a concern for me, as I always find ways to stay engaged, and I have long been at peace with myself. The time I spent in the Himalayas have helped me develop this.

Hard work

Lex Fridman
(02:18:35)
I’ve heard from many people that you are the hardest worker they know. What’s your philosophy behind that? Maybe you put in crazy hours every single day. Do you ever get tired? What’s your source of strength and perseverance through all of that?
Narendra Modi
(02:18:56)
Look, first of all, I don’t believe I’m the only one working. I look at the people around me and always think these people work harder than I do. When I think about farmers, I realize how hard they work. They toil and sweat under the open sky day after day. When I look at our country’s soldiers, I think about how many hours someone spends working tirelessly in snow, deserts, or even underwater day and night. When I see a laborer, I think about how hard they’re working. I always think about how hard our mothers and sisters work in every family for the happiness of the family. They’re the first to wake up and the last to go to bed, taking care of everyone in the family, while also managing social responsibilities. Thinking of all this, I’m in awe of how hard people work. So I think, “How can I sleep? How can I relax?” So naturally the motivation is right in front of my eyes. Those very things around me keep me motivated.

(02:20:04)
Secondly, my responsibilities keep pushing me forward. The responsibilities entrusted to me by my fellow citizens always remind me that I’m not here to enjoy privileges. I will always give it my absolute best. Perhaps there are a couple of things I may not accomplish, but there will never be a lack of effort or hard work from me. When I was campaigning in 2014, I had made a promise first in Gujarat and later across India. I promised my fellow citizens that I will never fall behind in hard work for my country. Secondly, I promised I would never act with bad intentions, and thirdly, I vowed I’d never do anything for personal gain.

(02:21:05)
Today it’s been 24 years. For such a long period, the people have entrusted me as head of government. I’ve continuously held myself to these three standards, and I still live by them today. My inspiration comes from serving 1.4 billion people from understanding and fulfilling their aspirations and addressing their needs. I’m always determined to do as much as I can, work as hard as possible. Even today, my energy remains just as strong.

Srinivasa Ramanujan

Lex Fridman
(02:21:42)
Me as an engineer, as a person who loves mathematics, I have to ask Srinivasa Ramanujan is an Indian mathematician from a century ago. He’s widely considered to be one of the greatest mathematicians of all time. Self-taught, grew up in poverty. You have often spoken about him. What do you find inspiring about him?
Narendra Modi
(02:22:07)
Look, I deeply respect him, and everyone in my country respects him as well, because I strongly believe there’s a deep connection between science and spirituality. If you closely observe many scientifically advanced minds, you’ll find they’re often spiritually advanced too. They’re not disconnected from spirituality. Srinivasa Ramanujan once said that his mathematical ideas came from the goddess he worshiped, meaning ideas emerge from spiritual discipline, and discipline is more than just hard work. It means fully devoting yourself to a task and completely immersing yourself into it so much that you become one with your work.

(02:23:10)
You see, the more open we are to new and different sources of knowledge, the more new ideas we’ll have. I think it’s important for us to clearly understand the difference between information and knowledge. Some people mistakenly confuse information with knowledge, carrying around large amounts of information, but I don’t believe information alone equals knowledge. Knowledge is something deeper. It gradually evolves through processing, reflection and understanding. Recognizing this difference is important in how we handle both.

Decision-making process

Lex Fridman
(02:23:49)
You have a reputation for being a decisive leader. So can you walk me through on this topic of ideas, how you make decisions? What’s your process? So for instance, when facing a high stakes choice with no clear precedence, a lot of uncertainty, having to balance input, how do you make decisions?
Narendra Modi
(02:24:13)
There are many factors to my decision making. First, I’m perhaps the only politician in India who has stayed overnight in around 85 to 90% of the districts across the country. This was before my current role. I used to travel extensively. I learned a lot from those experiences. They gave me firsthand knowledge of the ground realities and grassroots level issues, not something asked or heard or learned merely from books. Secondly, from a governance perspective, I carry no baggage of any kind. I don’t carry any baggage that weighs me down or forces me to act a certain way. Thirdly, I have a simple yardstick for decisions. My country first. I always question if what I’m doing harms my nation in any way. Additionally, Mahatma Gandhi once said that, if you’re ever unsure when making a decision, think of the poorest person’s face. Remember them and ask yourself, “Will this help them?” Then your decision will be right. That wisdom always guides me, remembering ordinary citizens and considering how my actions affect them.

(02:25:54)
Another factor in my approach is that I’m very well-connected in my administration. My officials know this well and probably feel overwhelmed by it, by the fact that my information channels are numerous and are very active. Because of that, I receive lots of insights from various sources. So when someone comes to brief me, that’s not my only source of information. I always have additional perspectives available to me.

(02:26:37)
Another thing, I maintain a learner’s mindset. Suppose I’m not familiar with something and an official explains it to me. I approach them like a student and ask, “Can you clarify this? How does it work? Then what happens next and how?” Whenever I have different information, I deliberately play devil’s advocate and ask challenging questions. I thoroughly analyze the issue from multiple angles, hoping that careful evaluation will yield something valuable. Then once I converge toward a decision or action that is worth taking, I share the idea informally with like-minded people just to gauge their reactions and see how they respond, gathering insights and feedback before moving forward, until finally, I have a strong conviction that my decision is right.

(02:27:46)
This entire decision-making process doesn’t actually take much time. My speed is very fast. Let me share an example. How did I make decisions during COVID? I had Nobel Prize winners advising me, giving countless economic examples from around the world. They’d say, “This country is doing this, that country did that. You should do it too.” Renowned economists constantly bombarded me with suggestions. Political parties pressured me relentlessly, urging me to spend huge amounts of money, but I didn’t act immediately. I paused and reflected. What exactly should I do? Then considering the unique conditions of my own country, I made a clear decision. I wouldn’t let the poor sleep hungry. I wouldn’t allow social tensions to arise over basic daily needs.

(02:28:43)
These core principles guided my approach. The entire world was in lockdown. Global economies were collapsing. Everyone pressured me to empty the treasury, print more currency and flood money everywhere, but I decided that this was not the right economic route to follow. And so instead, the path I chose, after carefully listening to experts, understanding their opinions without opposing them and combining their advice with my own country’s situation and my personal experiences, created a system that worked effectively. As a result, when the whole world suffered from severe inflation immediately after COVID, India did not.

(02:29:38)
Today, my country is steadily advancing at a rapid pace, emerging as one of the fastest growing major economies in the world. The main reason is that during that crisis, with patience and discipline, I resisted the temptation to apply every global theory blindly. We didn’t worry about what newspapers would say, whether they’d praise or criticize. Ignoring all that, I stayed focused on basic fundamentals, and by doing so, we succeeded and kept moving ahead. So ultimately, my economy benefited as well. My approach has always been to stay focused on these fundamentals.

(02:30:22)
Another strength is my risk-taking capacity. I don’t worry about potential losses for myself. If something is right for my country, for the people, I’m always prepared to take the risk. Secondly, I take ownership of my decisions. If something goes wrong, I don’t ship blame to others. I stand up, take responsibility, and own the outcome. When you take ownership, your team also becomes deeply committed. They know this person won’t let us down, won’t abandon us. He’ll always stand with us because they see I’m making honest decisions, not for myself, but for the nation. I’ve openly told the country from the start, I’m human. I can make mistakes, but I won’t act with bad intentions. People remember those words clearly. Even if something doesn’t go as planned, they trust that Modi’s intentions were right. They think he probably meant to do something good even if it didn’t work out. So society sees and accepts me just as I am.

AI

Lex Fridman
(02:31:35)
You gave a powerful speech on AI a few weeks ago at the AI Summit in France. In it you spoke about the talent pool for AI engineers in India. I think it’s probably one of the biggest pools of brilliant engineers in the world. So how can India become the leader in the space of AI? Currently lags behind the United States. What does it take for India to start winning and leading the world in AI?
Narendra Modi
(02:32:09)
One thing I’m about to say might sound strong, and it may even upset some people, but since you’ve asked, I’ll speak openly from my heart. No matter what the world does with AI, it will remain incomplete without India. I’m making this statement very responsibly. Tell me, you’ve heard my speech at the AI Summit in Paris on global cooperation. What do you think? Can anyone develop AI entirely on their own? What is your perspective on this?
Lex Fridman
(02:32:56)
You gave, actually, in your speech a brilliant example of the positive impact of AI and the limitations of AI. I think the example you gave is when you ask it to generate an image of a person writing with their left hand-
Narendra Modi
(02:33:17)
Left hand.
Lex Fridman
(02:33:17)
… it’s always going to generate a person writing with their right hand. So in that way, the West creating an AI system where India is not part of that process is always going to generate the person with the right hand is an essential part of what the world is historically, but especially in the 21st century.
Narendra Modi
(02:33:43)
I agree. I believe AI development is fundamentally a collaboration. Everyone involved supports one another through shared experiences and learning. India isn’t just developing theoretical AI models. It is actively working on and bringing to life AI-driven applications for very specific use cases to ensure that GPU access is available to every section of society. We have already created a unique marketplace-based model to ensure its broad accessibility. A significant mindset shift is taking place in India, though historical influences, traditional government procedures or the lack of strong support infrastructure made us appear as lagging behind to others.

(02:34:49)
Take 5G for example. The world initially believed we were far behind, but once we started, we became the fastest nation globally to roll out comprehensive 5G networks. Recently, an American company executive visited me and shared his experiences about this very fact. He told me that if I were to advertise in the US for engineers, I would only receive enough applicants to fill a single room at best. But if I do the same in India, even a football field wouldn’t be enough to hold them. This indicates that India has access to an extraordinarily vast pool of talent, and that’s our greatest strength. After all, artificial intelligence is fundamentally powered, shaped, and guided by human intelligence. Without genuine human intelligence, AI can’t thrive or progress sustainably, and that real intelligence exists abundantly in India’s youth and talent pool, and I believe that’s our greatest asset.
Lex Fridman
(02:36:10)
But also if you look, many of the top tech leaders, first of all, tech talent, but tech leaders in the US are of Indian origin, Sundar Pichai, Satya Nadella, Aravind Srinivas. You’ve met with some of them. What spirit of their Indian origins do you think they carry in them that enables them to be so successful?
Narendra Modi
(02:36:35)
Look, Indian culture emphasizes that there should be equal respect for the place where you’re born and the place where you work. There should be no difference. As much as there is dedication to the land of birth, there should be the same sense of dedication to the land of work. You should always give your best wherever you are. Because of these rich cultural values, every Indian strives to give their best effort, regardless of their role or position. They don’t wait until they’re in senior roles, even in smaller roles.

(02:37:10)
Secondly, they never get involved in anything questionable or unethical. They tend to remain dedicated to what’s right and ethical. Their nature is collaborative. They easily get along with others, eventually for success. Just having knowledge isn’t enough. The ability to work effectively as part of a team matters significantly more. Understanding people and harnessing their abilities is an incredibly valuable skill.

(02:37:38)
Generally, people raised in India, especially those coming from joint families and brought up in an open society, find it easier to lead complex tasks and large teams effectively. That’s why today, in major corporations across the globe, you’ll find Indians holding key leadership positions. The problem-solving abilities, along with the analytical thinking of Indian professionals are truly exceptional. I believe this capability is so strong, it makes Indians globally competitive and extremely valuable on the international stage. This is the reason why in fields like innovation, entrepreneurship, startups and boardrooms, you’ll find Indians achieving extraordinary results everywhere.

(02:38:37)
Take our space sector, for example. Previously, it was entirely government-controlled, but just a couple of years ago, I opened it up to the private sector, and now we already have 200 startups working in space technology. Moreover, our missions like Chandrayaan are extremely cost-effective. India’s Chandrayaan mission costs less than what Hollywood spends making a single blockbuster film. So when the world sees how cost-effective our work is, they naturally think why not partner with India? This automatically generates respect for Indian talent globally. I believe this is a hallmark of our civilizational ethos.
Lex Fridman
(02:39:28)
You spoke about this human intelligence. Do you worry that AI, artificial intelligence, will replace us humans?
Narendra Modi
(02:39:38)
It’s true that in every era, a competitive atmosphere was created between technology and humanity. At times, it was even portrayed as conflict. It was often portrayed as if technology would challenge human existence itself. But every time, as technology advanced, humans adapted and stayed a step ahead. It has always been the case. After all, it is humans who find the best ways to use technology to their advantage. I believe that with AI, humans are now being forced to reflect on what it truly means to be human. This is the real power of AI. Because of the way AI functions, it has challenged how we perceive work itself. But human imagination is the fuel. AI can create many things based on that, and in the future it may achieve even more. Still, I firmly believe that no technology can ever replace the boundless creativity and imagination of the human mind.
Lex Fridman
(02:41:05)
I agree with you. It does make me and a lot of people wonder what makes humans special because it seems that there’s a lot that makes humans special. The imagination, the creativity, the consciousness, the ability to be afraid, to love, to dream, to think outside of the box, outside of the box of the box of the box, take risks, all of those things.
Narendra Modi
(02:41:32)
Now, look, humans have an innate ability to care for each other, the natural tendency to be concerned about one another. Now, can someone tell me, is AI capable of this?

Education

Lex Fridman
(02:41:46)
This is one of the big open questions of the 21st century. Every year you host the Pariksha Pe Charcha where you interact directly with young students and give them advice on how to prepare for exams. I watched a bunch of them. So you give advice on how to succeed in exams, how to manage stress, all those kinds of things. Can you explain at a high level the different exams that students in India need to take in their education journey and why it’s so stressful?
Narendra Modi
(02:42:19)
By and large, a strange mindset has developed in society today. Even schools measure their success by students’ rankings. Families too feel pride when their child achieves a high rank because they believe it improves their educational and social status. This mentality has resulted in increased pressure on children. Kids also began feeling that their entire lives depend on 10th and 12th grade exams.

(02:42:53)
We’ve introduced significant changes in our new education policy to address this issue. But until those changes take effect on the ground, I feel another responsibility. If our children face challenges, it’s my duty to listen to them, understand them, and ease their burden. In a way, when I conduct Pariksha Pe Charcha, I get insights directly from the students, understand their parents’ mindset, as well as the perspectives of people in the educational field. So these discussions don’t just benefit the students. They benefit me too.

(02:43:37)
Exams are valuable for assessing knowledge in a specific domain, but they can’t become the sole measure of someone’s overall potential. Many people may not score high academically, yet can hit a century in cricket because that’s where their true strength lies. When the focus shifts to actual learning, scores tend to naturally improve.

(02:44:08)
I remember when I was a student, I had a teacher whose learning techniques still greatly appeal to me today. He would give us children specific instructions. To one child he would say, “Tomorrow, bring exactly 10 chickpeas from home.” To another, he might request, ” Bring 15 grains of rice. No more. No less.” A third child could be told, “Bring 21 mung beans. Precisely that number.” Different students got different quantities and varieties. So each child would think, “I need to get exactly 10.” Counting them at home helped them memorize numbers naturally. Then they’d learn what chickpeas were, and after returning to school, they’d pool it all. The teacher would then ask, ” Take out 10 chickpeas, three chickpeas, two mung beans-“
Narendra Modi
(02:45:03)
… take out 10 chickpeas, three chickpeas, two mung beans. This way children learned math and could identify chickpeas and mung beans effortlessly. I’m talking about early childhood education here. Such learning techniques educate children without burdening them, and we’ve incorporated similar methods into our new education policy.

(02:45:22)
When I was in school, I observed one of my teachers using an innovative idea. On his very first day, he placed a diary on the table and said, “Whoever arrives earliest each morning will write one sentence in this diary along with their name.” The next student would then need to write a related sentence.

(02:45:52)
At first, I’d rush to school very early every day. Why? So that I could write the first sentence. I once wrote something like, “Today’s sunrise was magnificent. It filled me with energy.” I’d write my name and whoever arrived after me had to write something connected to the sunrise as well.

(02:46:20)
After a few days, I realized my creativity wasn’t improving much from this. Why? Because I’d arrived with the fixed thought already in mind and simply write it down. So I decided I’d start going last instead. What happened then was that I’d read what others had written first and then try to give my very best.

(02:46:48)
As a result, my creativity began to improve even more. Sometimes teachers do these small simple activities that greatly impact your life. These experiences combined with my own background in organizational work, made human resource development a key area of focus for me.

(02:47:14)
That’s why I engage with children through events once or twice each year and over time, these efforts have resulted in a book that’s benefiting thousands of children serving as a valuable reference for them.
Lex Fridman
(02:47:27)
Can you speak a little bit more by way of advice to students of how to be successful on their path in their career, how to find the career and how to find success in India and just to all the people across the world who find inspiration in your words?
Narendra Modi
(02:47:50)
I believe that whatever task you get, if you perform that task with complete dedication and sincerity, they inevitably become an expert sooner or later, and their enhanced capabilities open doors to success.

(02:48:23)
While working, one must continually strive to improve their skills and should never underestimate their ability to learn. When someone constantly neglects their learning ability, they limit their growth, but those who look beyond their tasks and observe what others around them are doing, their capacity can double or even triple.

(02:48:54)
To young people I’d say this clearly, there’s no need to feel discouraged. There’s certainly some task out there destined just for you. Don’t worry. Focus on enhancing your skills and opportunities will come.

(02:49:15)
You may think, “I wanted to be a doctor, but became a teacher. My life is wasted.” Thinking like that won’t help you at all. All right, you didn’t become a doctor, but as a teacher, you can shape 100 doctors.

(02:49:28)
If you had become a doctor, you’d serve only your patients. But now, as a teacher, you can inspire students to fulfill their dreams of becoming doctors so that both you and your students can together serve millions of patients. Then he gains a new perspective on life.

(02:49:44)
“I couldn’t become a doctor so I was miserable and I was unhappy being a teacher, but now I realize as a teacher, I can create doctors.” Connecting your life to a greater purpose brings a sense of inspiration and meaning. I have always believed that God has given everyone unique capabilities. Never lose faith in your own abilities. You should always maintain trust in your own abilities. Keep believing in yourself and trust that when the opportunity comes, you’ll perform and you’ll succeed. That confidence makes a person deliver results.
Lex Fridman
(02:50:26)
How do those students deal with stress, with struggle, with difficulties along that path?
Narendra Modi
(02:50:33)
Parents must first understand that life is not just about taking exams. Families should understand that their children aren’t trophies meant to be displayed or models to show off in society. It’s not about saying, “Look, my kid scores so high.” Parents really need to stop using their kids just as status symbols.

(02:50:55)
Secondly, students should always keep themselves well-prepared beforehand. Only then can they appear for exams feeling stress-free and confident. They should have complete trust in themselves and their abilities.

(02:51:12)
Sometimes I see students panic over the smallest issues during exams. They take papers or other things, and when their pen suddenly stops working causing anxiety, sometimes they feel uneasy thinking, “Oh no, I don’t like sitting next to this person.” If the bench wobbles, their whole attention goes there, indicating self-doubt.

(02:51:33)
Those lacking confidence constantly keep looking for distractions. But if you are confident and have genuinely worked hard, just take a minute, take some deep breaths, relax your mind and refocus your attention calmly.

(02:51:48)
Slowly read through the questions and allocate your time systematically. “I have this amount of time I’ll dedicate these minutes per question.” In my experience, students who regularly practice writing test papers can easily overcome such situations without any trouble at all.

Learning and focus

Lex Fridman
(02:52:05)
And you said, “Always focus on learning.” What’s your approach to learning? What advice can you give on how to learn best, not just when you’re young, throughout your life?
Narendra Modi
(02:52:19)
Let me share a personal example with you. I used to learn a lot from reading, but these days, more and more, I learn by being fully present. Whenever I meet someone, I am fully present in the moment. I give them my full attention. This complete focus allows me to grasp new concepts quickly.

(02:52:48)
When I’m with you, I’m fully present, grounded in the moment. No calls or messages can pull me away from this moment with you. I am fully present focused on the here and now. That’s why I always believe this is a habit everyone should embrace.

(02:53:06)
It will sharpen your mind and improve your learning ability. Besides, knowledge alone cannot light the way. You must immerse yourself in the flow of practice. You cannot master driving merely by reading the life stories of great drivers. You must get behind the wheel and take the road yourself.

(02:53:26)
You must dare to take risks. You can never master the road if fear of accident or death holds you back. I truly believe that those who live in the present are the ones who live their life to the fullest. That’s because they know that every moment lived has already slipped into the past.

(02:53:53)
So you must embrace the moment before it fades into the past. Otherwise, chasing the future only turns the present into the past. It’s not a trade worth making. Most people stress so much about the future that their present quietly slips away. Before they know it the moment has already faded into the past.
Lex Fridman
(02:54:19)
Yeah, I’ve heard a lot of stories of you having meetings with people and it’s usually all the distractions, there’s no distractions. It’s just two human beings just like this and just focused on the moment and the interaction. That’s a really beautiful thing. And today really is a gift that you would give that focus to me, so thank you.

(02:54:42)
Let me ask maybe a difficult, maybe a human question. Do you contemplate your mortality? Are you afraid of death?
Narendra Modi
(02:54:54)
Can I ask you a question instead?
Lex Fridman
(02:54:57)
Sure.
Narendra Modi
(02:54:58)
I have a very interesting question. Life and death are two sides of a coin, but which of the two is more certain?
Lex Fridman
(02:55:09)
Death.
Narendra Modi
(02:55:10)
Exactly. Now, with that out of the way, we know for a fact that life itself is a whispered promise of death, and yet life is also destined to flourish. So again, in the dance of life and death, only death is certain, so why fear what is certain? That’s why you must embrace life instead of fretting over death.

(02:55:43)
That’s how life will evolve and flourish, for it is uncertain. That’s why you must commit to enriching, refining and elevating your life so you can live fully and with a purpose before death comes knocking.

(02:56:04)
That’s why you must let go of the fear of death. After all, death is inevitable and there’s no use worrying about when it will arrive. It will arrive when it’s meant to.
Lex Fridman
(02:56:17)
What gives you hope about the future? Not just of India, but all of human civilization, all of us humans here on earth?
Narendra Modi
(02:56:27)
Well, I am an optimist at heart. Pessimism and negativity are simply not ingrained in my mindset. They don’t align with the way I think. That’s why I always gravitate toward the bright side instead.

(02:56:48)
If we take a moment to reflect on the history of mankind, we see the incredible crisis humans have overcome with resilience and strength. We also see the major changes humanity has embraced to evolve with the times, and this continuous transformation has carried on for millennia.

(02:57:10)
In every era, it is in human nature to adapt to the ever-flowing current of change. And while our progress has gone through cycles of ups and downs, it is those people who can break free from the constraints of these historical cycles and outdated thinking patterns.

(02:57:35)
It is they who can help humanity achieve extraordinary positive breakthroughs with an unshackled speed and grace transcending the limitations of the old way of doing things by embracing change.

Mantra

Lex Fridman
(02:57:56)
In this moment, I was wondering if you could guide me perhaps through a Hindu prayer or meditation for a few moments. I learned, I’m trying to learn the Gayatri Mantra. In my fast I was trying to do the chants. Perhaps I could try chanting. You could tell me about the importance of this mantra and maybe others in your life, in your spirituality? Should I try?
Narendra Modi
(02:58:31)
Yes, please.
Lex Fridman
(02:58:34)
[foreign language 02:58:34]. How did I do? It’s okay?
Narendra Modi
(02:58:51)
You did great. [foreign language 02:58:58]. This mantra is dedicated to the radiant power of the sun and is considered a powerful tool for spiritual enlightenment.

(02:59:13)
Many mantras in Hindu philosophy are deeply intertwined in some intricate and interesting ways with science and nature, each woven into different facets of life. Chanting mantras on a daily basis brings profound and lasting benefits.

Meditation

Lex Fridman
(02:59:41)
In your own spirituality and your quiet moments, when you are with God, where does your mind go? What role do mantras play when you’re fasting, when you’re just alone with yourself?
Narendra Modi
(02:59:57)
The word meditation has been overused to the point that it feels like a cliché. In Indian languages, we usually refer to it as dhyan. If I associate dhyan to meditation, it might seem burdensome to some. One might think, “This is too difficult. I’m not an enlightened being.” But it’s not rocket science. It just means freeing yourself from distraction.

(03:00:32)
For example, even when you’re in class, your mind wanders to recess. All you think about is lunch, not the lesson. Meditation is simply being present in the moment. I recall an incident from my time living in the Himalayas. There I encountered this wise sage. He taught me a simple practical technique. It was nothing spiritual.

(03:01:06)
There are several little streams in the Himalayas. He positioned a large leaf to catch water from one of those streams and placed an upside down bowl below so water would drip rhythmically from the leaf onto the bowl.

(03:01:33)
He asked me to focus only on the dripping water, ignoring all other sounds. “Ignore the chirping birds and the soft rustle of the breeze.” He would place the leaf and I would meditate there for hours. I felt my mind slowly tuning into the rhythmic sounds of the water droplets falling onto the bowl like a melody guiding me into deep focus.

(03:02:09)
And it’s not like I was chanting mantras or reciting God’s name. I like to call it the divine resonance. It was by tuning into that divine resonance that I learned the art of concentration. This practice slowly evolved into meditation. Sometimes you happen to stay at a fancy hotel, you get a lavish, luxurious room. The decor is impeccable and you have been wholeheartedly fasting, but there’s a dripping faucet in the bathroom. That faint sound is enough to make a luxurious room feel worthless.

(03:02:57)
At times, we realize the value of concentration in life’s inner journey. We come to appreciate the difference a little bit of concentration can make. One very interesting concept comes to mind from our scriptures.

(03:03:21)
Since we spoke about life and death, I would like to quote a mantra, [foreign language 03:03:28]. In other words, all life is part of a complete circle, and this mantra emphasizes the path to achieve that completeness.

(03:03:44)
Similarly, Hindus never focus solely on individual well-being. [foreign language 03:03:52]. In other words, we wish for the well-being and prosperity of all. [foreign language 03:04:00]. This mantra encompasses the idea of universal well-being and prosperity.

(03:04:11)
And you know how this mantra ends? [foreign language 03:04:13]. Every Hindu mantra ends on the same note. Peace, peace, peace. These ancient and powerful rituals born in India have emerged from thousands of years of the spiritual practice of sages. They connect us to the essence of life.
Lex Fridman
(03:04:40)
[foreign language 03:04:40]. Thank you for this honor. Thank you for this incredible conversation. Thank you for welcoming me to India, and I can’t wait to break the fast with some Indian food tomorrow. Thank you so much, Prime Minister. This was an honor.
Narendra Modi
(03:04:56)
I am thankful for the opportunity to have this conversation with you. After fasting for two days I recommend you ease into eating slowly and I hope you reap great benefits from this fasting experience.

(03:05:11)
I’ve explored several new realms of thought for the first time ever with you today. I had long kept those thoughts tucked away within myself, but today you brought those thoughts to light. I hope [inaudible 03:05:30].
Lex Fridman
(03:05:29)
Thank you.
Narendra Modi
(03:05:30)
I hope your viewers may enjoy this. It was a great pleasure speaking with you. Thank you.

Lex visiting India

Lex Fridman
(03:05:38)
Thank you. Thank you for listening to this conversation with Prime Minister Narendra Modi. And now, let me ask you some questions and try reflect on and articulate some things I’ve been thinking about. If you would like to submit questions or get in touch with me for whatever reason, go to lexfridman.com/contact.

(03:05:58)
First, let me give a shout out to the amazing team around the Prime Minister. Everyone was super kind, excellent at what they do, efficient, great communication, and just great people all around.

(03:06:13)
And since I spoke English and Prime Minister Modi spoke Hindi, I have to comment on the interpreter who was doing simultaneous interpreting for both of us. She was absolutely amazing. I can’t sing her enough praises. From the equipment used to the quality of the translation, to just the human touch of it all.

(03:06:36)
And in general, my travels around Delhi and India revealed to me some early glimpses of what felt like another world, almost like another planet, different culturally from anything I’ve experienced before. A chaos of human interactions, out there, big dynamic personalities and characters.

(03:06:58)
Obviously, India’s composed of many distinct subcultures and Delhi represents just one slice. Much like neither New York or Texas or Iowa alone represent America. They’re all different flavors of America.

(03:07:13)
On my visit, I walked around and rode rickshaws everywhere, just aimlessly wandering the streets, looking to talk to people about life. Of course, like many places on earth, there are always some people, especially those that have something to sell, who will at first see me as a tourist, a foreign traveler, one with some money to spend.

(03:07:36)
Like always, I avoided such shallow interactions and went straight past the small talk to the meaningful conversations. Shooting the shit about what they love, what they fear, what kind of hardship and triumph they’ve experienced in their lives.

(03:07:52)
I think the cool thing about people anywhere on earth is they quickly do see the real you past the facades that strangers put up for each other, if you’re vulnerable and honest enough to let them. And I tried to do just that and I should say that for the most part, everyone was super kind in the genuine human way.

(03:08:17)
Even when they didn’t speak English, it was always easy to understand. Probably more than any other peoples I’ve interacted with in India, people’s eyes, faces, body language, all communicate a lot of information, a lot of emotion, not reserved at all.

(03:08:33)
When I traveled through Eastern Europe, for example, in contrast, reading a person is much tougher. The meme does have some truth to it. There’s often a protective layer between the heart of the person and the outside world. In India, it’s all there on full display. So I had a lot of epic conversations and interactions as I walked around Delhi for a couple of weeks.

(03:08:58)
In general, on the topic of reading people, I do believe the eyes can often say more than words can. We humans are a fascinating bunch. There really is a deep turbulent ocean behind the surface waves we show the world. In some sense, what I try to do in conversations, on and off the mic, is to get to that depth.

(03:09:20)
Anyway, the few weeks I spent in India were a magical experience. Traffic alone was a wild time, like the world’s most difficult test for self-driving cars.

(03:09:32)
It reminded me of watching nature documentary videos of swarms of fish when it’s thousands of them swimming around at insane speeds, seemingly in complete chaos. And yet when looking at the big picture of it it all works like a perfectly tuned orchestra.

(03:09:52)
I will, most certainly, travel around India with my friend Paul Rosalie in the near future, maybe with some other friends all around from the north of India to the south.

Siddhartha


(03:10:04)
Now, allow me to also comment about one of the books that first drew me toward India and to its deep history of philosophical and spiritual traditions. The book is Siddhartha by Hermann Hesse.

(03:10:19)
I first read most of Hesse’s major work as a teenager, but then reread them again through the years. It first found me, Siddhartha, when I was immersed in a very different kind of literature of Dostoevsky, Camus, Kafka, Orwell, Hemingway, Kerouac, Steinbeck and so on. Many of these explore the same human condition that puzzled me when I was a young man and still puzzles me today, even more so.

(03:10:49)
But Siddhartha was my introduction to the Eastern way of looking at these puzzles. It was written by Hermann Hesse. And by the way, please allow me this pronunciation of his last name. I’ve heard some people say Hesse, but my whole life I’ve always said Hesse.

(03:11:04)
So, anyway, it was written by Hermann Hesse, a German-Swiss Nobel Prize winning writer during one of the darkest periods of his own life. His marriage was failing, World War I has shattered his pacifist ideals and he suffered from debilitating headaches, insomnia, and depression.

(03:11:22)
During this period, he began psychoanalysis with Carl Jung, which, in part, led him to explore Eastern philosophies as a way to heal his fractured psyche. Hesse immersed himself in translations of ancient Hindu and Buddhist texts studying the Upanishads and the Bhagavad Gita.

(03:11:42)
And so, the writing of Siddhartha was in itself, for him, a journey that paralleled that of the main character in the book. Hesse started writing the book in 1919 and finished three years later experiencing an extensive psychological crisis in the middle. The book follows Siddhartha, a young man in ancient India, as he leaves behind wealth and comfort to search for meaning. You can feel his personal struggle in every page. Siddhartha’s restlessness, his dissatisfaction with conventional wisdom, his need to find truth, the direct experience.

(03:12:20)
Again, the book wasn’t simply a philosophical exploration for Hesse, it was psychological survival. He was writing his way out of suffering and towards his own enlightenment. I won’t go into a deep analysis of the book here, but I will mention two key lessons I took away and carry with me to this day.

(03:12:41)
First lesson comes from the scene in the book that to me is one of the great scenes in all of literature. Siddhartha is sitting by a river just listening, and in that river he hears all of life. All sounds, all voices, all of time, past, present, future, flowing together as one.

(03:13:03)
That scene gave me the experience and the notion that while in some grounded human sense, the linear arrow of time does exist. In another sense, time is a kind of illusion that, in fact, everything exists simultaneously. That our lives are both momentary and eternal. It is hard to describe these ideas with words. I think they must be experienced as personal revelations.

(03:13:32)
I’m reminded of the fish story that David Foster Wallace, another one of my favorite writers, described in a commencement speech 20 years ago. The story goes, “Two young fish are swimming along when they encounter an older fish swimming the opposite way. The older fish nods and says, ‘Morning boys, how’s the water?’ The young fish swim on, and eventually one turns to the other and asks, ‘What the hell is water?'”

(03:14:04)
The illusion of the forward progress of time is water in this metaphor. As humans, we’re fully immersed in it, but enlightenment, in part, involves being able to step back and get a glimpse at another deeper perspective on reality where all things are inextricably interconnected across both time and space.

(03:14:28)
Another key lesson from the novel that was especially formative to me as a young man was that one should not blindly follow others or learn about the world exclusively through books. But rather forge your own path and thrust yourself into the world where the lessons of life can only be learned by experiencing them directly.

(03:14:49)
And every experience, both positive and negative, mistakes, suffering and even seemingly wasted time is all an essential part of growth. To this point, Hesse draws a distinction between knowledge and wisdom. Knowledge can be taught by others. Wisdom can only be gathered through experiencing the full mess of life yourself.

(03:15:13)
In other words, the path to understanding isn’t through rejection of the world, but through complete immersion in it. Those are my early steps in seeing the world through the lens of Eastern philosophy. But many of Hesse’s books had an impact on me. I would recommend to read Demian, when you are younger, Steppenwolf, when you are older, Siddhartha throughout your life, especially in moments of crisis, and The Glass Bead Game if you want to take on Hesse’s magnum opus, that rigorously explores the ways the human mind and human civilization can engage in the pursuit of knowledge, wisdom, and meaning.

(03:15:55)
But Siddhartha is the only one I’ve returned to more than twice. In my own life, when faced with a difficult situation, I often return to the moment in the book when Siddhartha is asked what skills he possesses, and his answer is simply, “I can think, I can wait, I can fast.”

(03:16:15)
Let me elaborate. Indeed, for the first part, “I can think.” As Marcus Aurelius said, “The quality of your life is determined by the quality of your thoughts.” For the second part, “I can wait.” Patience and waiting often is indeed the optimal decision when facing a problem. Time does bring clarity and depth of understanding.

(03:16:39)
For the third part, “I can fast.” When needed, being able to live and flourish with less is a prerequisite of being free when the mind, the body, and society all are trying to put you in cages.

(03:16:56)
All right, friends, now sadly, our time together in this episode has come to a close. As always, thank you for being here and thank you for your support through the years.

(03:17:08)
Let me leave you with a few words from the Bhagavad Gita. “He who experiences the unity of life, sees his own self in all beings and all beings in his own self and looks unto everything with an impartial eye.” Thank you for listening. I hope to see you next time.

Transcript for DeepSeek, China, OpenAI, NVIDIA, xAI, TSMC, Stargate, and AI Megaclusters | Lex Fridman Podcast #459

This is a transcript of Lex Fridman Podcast #459 with Dylan Patel and Nathan Lambert.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Lex Fridman
(00:00:00)
The following is a conversation with Dylan Patel and Nathan Lambert. Dylan runs SemiAnalysis, a well-respected research and analysis company that specializes in semiconductors, GPUs, CPUs, and AI hardware in general. Nathan is a research scientist at the Allen Institute for AI and is the author of the amazing blog on AI called Interconnects. They are both highly respected, read and listened to by the experts, researchers and engineers in the field of AI. And personally, I’m just a fan of the two of them, so I used the DeepSeek moment that shook the AI world a bit as an opportunity to sit down with them and lay it all out from DeepSeek, OpenAI, Google XAI, Meta, Anthropic to NVIDIA and DSMC, and to US-China-Taiwan relations and everything else that is happening at the cutting edge of AI. This conversation is a deep dive into many critical aspects of the AI industry.

(00:01:08)
While it does get super technical, we try to make sure that it’s still accessible to folks outside of the AI field by defining terms, stating important concepts explicitly, spelling out acronyms, and in general, always moving across the several layers of abstraction and levels of detail. There is a lot of hype in the media about what AI is and isn’t. The purpose of this podcast in part is to cut through the hype, through the bullshit and the low resolution analysis and to discuss in detail how stuff works and what the implications are. Let me also, if I may comment on the new OpenAI o3-mini reasoning model, the release of which we were anticipating during the conversation and it did indeed come out right after. Its capabilities and costs are on par with our expectations as we stated. OpenAI o3-mini is indeed a great model, but it should be stated that DeepSeek-R1 has similar performance on benchmarks, is still cheaper and it reveals its chain of thought reasoning, which o3-mini does not. It only shows a summary of the reasoning, plus R1 is open weight and o3-mini is not.

(00:02:29)
By the way, I got a chance to play with o3-mini and anecdotal vibe check wise, I felt that o3-mini, specifically o3-mini high is better than R1. Still for me personally, I find that Claude Sonnet 3.5 is the best model for programming except for tricky cases where I will use o1 Pro to brainstorm. Either way, many more better AI models will come including reasoning models both from American and Chinese companies. They’ll continue to shift the cost curve, but the quote “DeepSeek moment” is indeed real. I think it will still be remembered five years from now as a pivotal event in tech history due in part to the geopolitical implications, but for other reasons to, as we discuss in detail from many perspectives in this conversation. This is the Lex Fridman podcast, to support it please check out our sponsors in the description. And now, dear friends, here’s Dylan Patel and Nathan Lambert.

DeepSeek-R1 and DeepSeek-V3

Lex Fridman
(00:03:33)
A lot of people are curious to understand China’s DeepSeek AI models, so let’s lay it out. Nathan, can you describe what DeepSeek-V3 and DeepSeek-R1 are, how they work, how they’re trained? Let’s look at the big picture and then we’ll zoom in on the details.
Nathan Lambert
(00:03:50)
DeepSeek-V3 is a new mixture of experts, transformer language model from DeepSeek who is based in China. They have some new specifics in the model that we’ll get into. Largely this is a open weight model and it’s a instruction model like what you would use in ChatGPT. They also released what is called the base model, which is before these techniques of post-training. Most people use instruction models today, and those are what’s served in all sorts of applications. This was released on, I believe, December 26th or that week. And then weeks later on January 20th, DeepSeek released DeepSeek-R1, which is a reasoning model, which really accelerated a lot of this discussion.

(00:04:38)
This reasoning model has a lot of overlapping training steps to DeepSeek-V3, and it’s confusing that you have a base model called V3 that you do something to to get a chat model and then you do some different things to get a reasoning model. I think a lot of the AI industry is going through this challenge of communications right now where OpenAI makes fun of their own naming schemes. They have GPT-4o, they have OpenIA o1, and there’s a lot of types of models, so we’re going to break down what each of them are. There’s a lot of technical specifics on training and go through them high level to specific and go through each of them.
Lex Fridman
(00:05:14)
There’s so many places we can go here, but maybe let’s go to open weights first. What does it mean for a model to be open weights and what are the different flavors of open source in general?
Nathan Lambert
(00:05:24)
This discussion has been going on for a long time in AI. It became more important since ChatGPT or more focal since ChatGPT at the end of 2022. Open weights is the accepted term for when model weights of a language model are available on the internet for people to download. Those weights can have different licenses, which is effectively the terms by which you can use the model. There are licenses that come from history and open source software. There are licenses that are designed by companies specifically all of Llama, DeepSeek, Qwen, Mistral, these popular names in open weight models have some of their own licenses. It’s complicated because not all the same models have the same terms. The big debate is on what makes a model open weight. It’s like, why are we saying this term? It’s a mouthful. It sounds close to open source, but it’s not the same.

(00:06:17)
There’s still a lot of debate on the definition and soul of open source AI. Open source software has a rich history on freedom to modify, freedom to take on your own, freedom for many restrictions on how you would use the software and what that means for AI is still being defined. For what I do, I work at the Allen Institute for AI, we’re a nonprofit, we want to make AI open for everybody and we try to lead on what we think is truly open source. There’s not full agreement in the community, but for us that means releasing the training data, releasing the training code, and then also having open weights like this. And we’ll get into the details of the models and again and again as we try to get deeper into how the models were trained, we will say things like the data processing, data filtering data quality is the number one determinant of the model quality.

(00:07:09)
And then a lot of the training code is the determinant on how long it takes to train and how fast your experimentation is. Without fully open source models where you have access to this data, it is hard to know… Or it’s harder to replicate. We’ll get into cost numbers for DeepSeek-V3 on mostly GPU hours and how much you could pay to rent those yourselves. But without the data, the replication cost is going to be far, far higher. And same goes for the code.
Lex Fridman
(00:07:37)
We should also say that this is probably one of the more open models out of the frontier models.
Nathan Lambert
(00:07:43)
Yes.
Lex Fridman
(00:07:45)
In this full spectrum where probably the fullest open source, like you said, open code, open data, open weights, this is not open code, this is probably not open data and this is open weights and the licensing is MIT license or it’s… There’s some nuance in the different models, but it’s towards the free… In terms of the open source movement, these are the good guys.
Nathan Lambert
(00:08:13)
Yeah. DeepSeek is doing fantastic work for disseminating understanding of AI. Their papers are extremely detailed in what they do and for other teams around the world, they’re very actionable in terms of improving your own training techniques. And we’ll talk about licenses more, the DeepSeek-R1 model has a very permissive license. It’s called the MIT license. That effectively means there’s no downstream restrictions on commercial use, there’s no use case restrictions. You can use the outputs from the models to create synthetic data.

(00:08:47)
And this is all fantastic. I think the closest peer is something like Llama where you have the weights and you have a technical report. And the technical report is very good for Llama. One of the most read PDFs of the year last year is the Llama 3 paper, but in some ways it’s slightly less actionable. It has less details on the training specifics. I think less plots and so on. And the Llama 3 license is more restrictive than MIT. And then between the DeepSeek custom license and the Llama license, we could get into this whole rabbit hole, I think. We’ll make sure we want to go down the license rabbit hole before we do specifics.
Lex Fridman
(00:09:22)
It should be stated that one of the implications that DeepSeek, it puts pressure on Llama and everybody else on OpenAI to push towards open source. And that’s the other side of open source is that you mentioned is how much is published in detail about it, so how open are you with the insights behind the code? How good is the technical reports? Are there hand wavy or is there actual details in there? And that’s one of the things that DeepSeek did well is they published a lot of the details.
Nathan Lambert
(00:09:52)
Especially in the DeepSeek-V3, which is their pre-training paper. They were very clear that they are doing interventions on the technical stack that go at many different levels. For example, on their to get highly efficient training, they’re making modifications at or below the CUDA layer for NVIDIA chips. I have never worked there myself and there are a few people in the world that do that very well, and some of them are at DeepSeek. These types of people are at DeepSeek and leading American frontier labs, but there are not many places.
Lex Fridman
(00:10:25)
To help people understand the other implication of open weights, just there’s a topic we’ll return to often here. There’s a fear that China, the nation might have interest in stealing American data, violating privacy of American citizens. What can we say about open weights to help us understand what the weights are able to do in terms of stealing people’s data?
Nathan Lambert
(00:10:55)
These weights that you can download from Hugging Face or other platforms are very big matrices of numbers. You can download them to a computer in your own house that has no internet and you can run this model and you’re totally in control of your data. That is something that is different than how a lot of language model usage is actually done today, which is mostly through APIs where you send your prompt to GPUs run by certain companies. And these companies will have different distributions and policies on how your data is stored, if it is used to train future models, where it is stored, if it is encrypted, and so on. The open weights are you have your fate of data in your own hands, and that is something that is deeply connected to the soul of open source.
Lex Fridman
(00:11:37)
It’s not the model that steals your data, it’s whoever is hosting the model, which could be China if you’re using the DeepSeek app or it could be Perplexity. You’re trusting them with your data or OpenAI, you’re trusting them with your data. And some of these are American companies, some these are Chinese companies, but the model itself is not doing the stealing, it’s the host. All right, so back to the basics. What’s the difference between DeepSeek-V3 and DeepSeek-R1? Can we try to lay out the confusion potential?
Nathan Lambert
(00:12:11)
Yes. For one, I have very understanding of many people being confused by these two model names, so I would say the best way to think about this is that when training a language model, you have what is called pre-training, which is when you’re predicting the large amounts of mostly internet text you’re trying to predict the next token. And what to know about these new DeepSeek models is that they do this internet large scale pre-training once to get what is called DeepSeek-V3 base. This is a base model, it’s just going to finish your sentences for you. It’s going to be harder to work with than ChatGPT. And then what DeepSeek did is they’ve done two different post-training regimes to make the models have specific desirable behaviors. What is the more normal model in terms of the last few years of AI, an instruct model, a chat model, a quote unquote “aligned model”, a helpful model. There are many ways to describe this is more standard post-training. This is things like instruction tuning, reinforcement learning from human feedback.

(00:13:12)
We’ll get into some of these words and this is what they did to create the DeepSeek-V3 model. This was the first model to be released and it is very high performant, it’s competitive with GPT-4, Llama 405B and so on. And then when this release was happening, we don’t know their exact timeline or soon after they were finishing the training of a different training process from the same next token prediction based model that I talked about, which is when this new reasoning training that people have heard about comes in in order to create the model that is called DeepSeek-R1. The R through this conversation is good for grounding for reasoning. And the name is also similar to OpenAI’s o1, which is the other reasoning model that people have heard about. And we’ll have to break down the training for R1 in more detail because for one we have a paper detailing it, but also it is a far newer set of techniques for the AI community, so it is a much more rapidly evolving area of research.
Lex Fridman
(00:14:11)
Maybe we should also say the big two categories of training of pre-training and post-training. These are umbrella terms that people use, so what is pre-training and what is post-training and what are the different flavors of things underneath the post-training umbrella?
Nathan Lambert
(00:14:28)
Pre-training, I’m using some of the same words to really get the message across is you’re doing what is called autoregressive prediction to predict the next token in a series of documents. This is done over standard practice is trillions of tokens, so this is a ton of data that is mostly scraped from the web. And some of DeepSeek’s earlier papers, they talk about their training data being distilled for math. I shouldn’t use this word yet, but taken from Common Crawl and that’s a public access that anyone listening to this could go download data from the Common Crawl website. This is a crawler that is maintained publicly. Yes, other tech companies eventually shift to their own crawler and DeepSeek likely has done this as well as most frontier labs do. But this sort of data is something that people can get started with and you’re just predicting text in a series of documents.

(00:15:18)
This can be scaled to be very efficient and there’s a lot of numbers that are thrown around in AI training like how many floating-point operations or flops are used. And then you can also look at how many hours of these GPUs that are used. And it’s largely one loss function taken to a very large amount of compute usage. You set up really efficient systems and then at the end of that you have the base model and pre-training is where there is a lot more of complexity in terms of how the process is emerging or evolving and the different types of training losses that you’ll use. I think this is a lot of techniques grounded in the natural language processing literature. The oldest technique which is still used today is something called instruction tuning or also known as supervised fine-tuning. These acronyms will be IFT or SFT.

(00:16:16)
People really go back and forth throughout them, and I’ll probably do the same, which is where you add this formatting to the model where it knows to take a question that is, explain the history of the Roman Empire to me or a sort of question you’ll see on Reddit or Stack Overflow. And then the model will respond in a information-dense but presentable manner. The core of that formatting is in this instruction tuning phase. And then there’s two other categories of loss functions that are being used today. One I’ll classify as preference fine-tuning. Preference fine-tuning is a generalized term for what came out of reinforcement learning from human feedback, which is RLHF. This reinforcement learning from human feedback is credited as the technique that helped ChatGPT break through. It is a technique to make the responses that are nicely formatted like these Reddit answers more in tune with what a human would like to read.

(00:17:14)
This is done by collecting pairwise preferences from actual humans out in the world to start and now AIs are also labeling this data and we’ll get into those trade-offs. And you have this contrastive loss function between a good answer and a bad answer. And the model learns to pick up these trends. There’s different implementation ways. You have things called reward models. You could have direct alignment algorithms. There’s a lot of really specific things you can do, but all of this is about fine-tuning to human preferences. And the final stage is much newer and will link to what is done in R1 and these reasoning models is I think OpenAI’s name for this, they had this new API in the fall, which they called the reinforcement fine-tuning API. This is the idea that you use the techniques of reinforcement learning, which is a whole framework of AI.

(00:18:02)
There’s a deep literature here to summarize, it’s often known as trial and error learning or the subfield of AI where you’re trying to make sequential decisions in a certain potentially noisy environment. There’s a lot of ways we could go down that, but fine-tuning language models where they can generate an answer and then you check to see if the answer matches the true solution. For math or code you have an exactly correct answer for math, you can have unit tests for code. And what we’re doing is we are checking the language model’s work and we’re giving it multiple opportunities on the same questions to see if it is right. And if you keep doing this, the models can learn to improve in verifiable domains to a great extent. It works really well. It’s a newer technique in the academic literature. It’s been used at frontier labs in the US that don’t share every detail for multiple years. This is the idea of using reinforcement learning with language models and it has been taking off especially in this DeepSeek moment.
Lex Fridman
(00:19:00)
And we should say that there’s a lot of exciting stuff going on again across the stack, but the post-training probably this year, there’s going to be a lot of interesting developments in the post-training. We’ll talk about it. I almost forgot to talk about the difference between DeepSeek-V3 and R1 on the user experience side. Forget the technical stuff, forget all of that, just people that don’t know anything about AI, they show up. What’s the actual experience, what’s the use case for each one when they actually type and talk to it? What is each good at and that kind of thing?
Nathan Lambert
(00:19:32)
Let’s start with DeepSeek-V3, again it’s more people would tried something like it. You ask it a question, it’ll start generating tokens very fast and those tokens will look like a very human legible answer. It’ll be some sort of markdown list. It might have formatting to help you draw to the core details in the answer and it’ll generate tens to hundreds of tokens. A token is normally a word for common words or a sub word part in a longer word, and it’ll look like a very high quality Reddit or Stack Overflow answer. These models are really getting good at doing these across a wide variety of domains, I think. Even things that if you’re an expert, things that are close to the fringe of knowledge, they will still be fairly good at, I think.

(00:20:19)
Cutting edge AI topics that I do research on, these models are capable for study aid and they’re regularly updated. Where this changes is with the DeepSeek- R1, what is called these reasoning models is when you see tokens coming from these models to start, it will be a large chain of thought process. We’ll get back to chain of thought in a second, which looks like a lot of tokens where the model is explaining the problem. The model will often break down the problem and be like, okay, they asked me for this. Let’s break down the problem. I’m going to need to do this. And you’ll see all of this generating from the model. It’ll come very fast in most user experiences. These APIs are very fast, so you’ll see a lot of tokens, a lot of words show up really fast, it’ll keep flowing on the screen and this is all the reasoning process.

(00:21:06)
And then eventually the model will change its tone in R1 and it’ll write the answer where it summarizes its reasoning process and writes a similar answer to the first types of model. But in DeepSeek’s case, which is part of why this was so popular even outside the AI community, is that you can see how the language model is breaking down problems. And then you get this answer, on a technical side they train the model to do this specifically where they have a section which is reasoning, and then it generates a special token, which is probably hidden from the user most of the time, which says, okay, I’m starting the answer. The model is trained to do this two stage process on its own. If you use a similar model in say, OpenAI, OpenAI’s user interface is trying to summarize this process for you nicely by showing the sections that the model is doing and it’ll click through, it’ll say breaking down the problem, making X calculation, cleaning the result, and then the answer will come for something like OpenAI.
Lex Fridman
(00:22:03)
Maybe it’s useful here to go through an example of a DeepSeek-R1 reasoning.
Nathan Lambert
(00:22:10)
If you’re looking at the screen here, what you’ll see is a screenshot of the DeepSeek chat app, and at the top is thought for 157 seconds with the drop-down arrow. Underneath that if we were in an app that we were running, the drop-down arrow would have the reasoning.
Lex Fridman
(00:22:25)
In this case, the specific question which I’m philosophically/ pothead inclined, so this is asking DeepSeek-R1 for one truly novel insight about humans. And it reveals the reasoning and basically the truly novel aspect is what’s pushing the reasoning to constantly… The model asking itself, is this truly novel? It’s actually challenging itself to be more novel, more counterintuitive, less cringe, I suppose. Some of the reasoning says… This is just snapshots. Alternatively, humans have a unique meta-emotion where they feel emotions about their own emotions e.g. feeling guilty about being angry. This recursive emotional layering creates complex motivational drives that don’t exist in other animals. The insight is that human emotions are nested. It’s reasoning through how humans feel emotions, it’s reasoning about meta emotions.
Nathan Lambert
(00:23:21)
It’s going to have pages and pages of this. It’s almost too much to actually read, but it’s nice to skim as it’s coming.
Lex Fridman
(00:23:27)
It’s a James Joyce like stream of consciousness. And then it goes, wait, the user wants something that’s not seen anywhere else. Let me dig deeper. And consider the human ability to hold contradictory beliefs simultaneously. Cognitive dissonance is known, but perhaps the function is to allow flexible adaptation, so on and so forth. That really captures the public imagination that holy shit, this isn’t intelligence/almost like an inkling of sentience because you’re thinking through, you’re self-reflecting, you’re deliberating. And the final result of that after 157 seconds is, humans instinctively convert selfish desires into cooperative systems by collectively pretending abstract rules, money, laws, rights are real. These shared hallucinations act as quote “games” where competition is secretly redirected to benefit the group, turning conflict into society’s fuel. Pretty profound.
Nathan Lambert
(00:24:31)
This is a potential digression, but a lot of people have found that these reasoning models can sometimes produce much more eloquent text. That is a at least interesting example I think depending on how open-minded you are, you find language models interesting or not, and there’s a spectrum there.
Lex Fridman
(00:24:49)
We’ll talk about different benchmarks and so on but some has just a vibe. That in itself is a, let’s say quote “fire” tweet. If I’m trying to produce something where people are like, “Oh, shit.” Okay, so that’s a chance probably return to it more. How were they able to achieve such low cost on the training and the inference? Maybe you could talk to the training first.

Low cost of training

Dylan Patel
(00:25:16)
There’s two main techniques that they implemented that are probably the majority of their efficiency, and then there’s a lot of implementation details that maybe we’ll gloss over or get into later that contribute to it. But those two main things are, one is they went to a mixture of experts model, which we’ll define in a second. And then the other thing is that they invented this new technique called MLA, latent attention. Both of these are big deals. Mixture of experts is something that’s been in the literature for a handful of years. And OpenAI with GPT-4 was the first one to productize a mixture of experts model. And what this means is when you look at the common models around that most people have been able to interact with that are open, think Llama. Llama is a dense model i.e. every single parameter or neuron is activated as you’re going through the model for every single token you generate.

(00:26:10)
Now, with a mixture of experts model, you don’t do that. How does the human actually work? Is like, oh, well my visual cortex is active when I’m thinking about vision tasks and other things. My amygdala is when I’m scared. These different aspects of your brain are focused on different things. A mixture of experts, models attempts to approximate this to some extent. It’s nowhere close to what a brain architecture is, but different portions of the model activate. You’ll have a set number of experts in the model and a set number that are activated each time. And this dramatically reduces both your training and inference costs because now if you think about the parameter count as the total embedding space for all of this knowledge that you’re compressing down during training, one, you’re embedding this data in instead of having to activate every single parameter, every single time you’re training or running inference, now you can just activate on a subset and the model will learn which expert to route to for different tasks.

(00:27:07)
And so this is a humongous innovation in terms of, hey, I can continue to grow the total embedding space of parameters. And so DeepSeek’s model is 600 something billion parameters, relative to Llama 405B, it’s 405 billion parameters, relative to Llama 70B, it’s 70 billion parameters. This model technically has more embedding space for information to compress all of the world’s knowledge that’s on the internet down. But at the same time, it is only activating around 37 billion of the parameters, so only 37 billion of these parameters actually need to be computed every single time you’re training data or inferencing data out of it. Versus again, the Llama model, 70 billion parameters must be activated or 405 billion parameters must be activated, so you’ve dramatically reduced your compute cost when you’re doing training and inference with this mixture of experts architecture.
Nathan Lambert
(00:27:57)
Should we break down where it actually applies and go into the transformer? Is that useful?
Lex Fridman
(00:28:02)
Let’s go. Let’s go into the transformer.
Nathan Lambert
(00:28:03)
The transformer is a thing that is talked about a lot, and we will not cover every detail. Essentially the transformer is built on repeated blocks of this attention mechanism and then a traditional dense fully connected multilayer perception, whatever word you want to use for your normal neural network. And you alternate these blocks. There’s other details and where mixture of experts is applied is at this dense model. The dense model holds most of the weights if you count them in a transformer model, so you can get really big gains from those mixture of experts on parameter efficiency at training and inference because you get this efficiency by not activating all of these parameters.
Lex Fridman
(00:28:44)
We should also say that a transformer is a giant neural network.
Nathan Lambert
(00:28:48)
Yeah.
Lex Fridman
(00:28:49)
And then there’s, for 15 years now, there’s what’s called the deep learning revolution. Network’s gotten larger and larger. At a certain point, the scaling laws appeared where people realized-
Dylan Patel
(00:29:00)
This is a scaling law shirt by the way.
Lex Fridman
(00:29:02)
Representing scaling laws. Where it became more and more formalized that bigger is better across multiple dimensions of what bigger means. But these are all neural networks we’re talking about, and we’re talking about different architectures of how to construct these neural networks such that the training and the inference on them is super efficient.
Nathan Lambert
(00:29:24)
Yeah. Every different type of model has a different scaling law for it, which is effectively for how much compute you put in the architecture will get to different levels of performance at test tasks. And mixture of experts is one of the ones at training time even if you don’t consider the inference benefits, which are also big. At training time, your efficiency with your GPUs is dramatically improved by using this architecture if it is well implemented. You can get effectively the same performance model and evaluation scores with numbers like 30% less compute, I think. There’s going to be a wide variation depending on your implementation details and stuff. But it is just important to realize that this type of technical innovation is something that gives huge gains. And I expect most companies that are serving their models to move to this mixture of experts implementation. Historically, the reason why not everyone might do it is because it’s an implementation complexity, especially when doing these big models.

(00:30:21)
This is one of the things that DeepSeek gets credit for is they do this extremely well. They do a mixture of experts extremely well. This architecture for what is called DeepSeek MoE, MoE is the shortened version of mixture of experts, is multiple papers old. This part of their training infrastructure is not new to these models alone. And same goes for what Dylan mentioned with multi-head latent attention. This is all about reducing memory usage during inference and same things during training by using some fancy low rank approximation math. If you get into the details with this latent attention, it’s one of those things I look at and it’s like, okay, they’re doing really complex implementations because there’s other parts of language models such as embeddings that are used to extend the context length, the common one that DeepSeek used is rotary positional embeddings, which is called RoPE.

(00:31:12)
And if you want to use RoPE with a normal MoE, it’s a sequential thing, you take two of the attention matrices and you rotate them by a complex value rotation, which is a matrix multiplication. With DeepSeek’s MLA, with this new attention architecture, they need to do some clever things because they’re not set up the same and it just makes the implementation complexity much higher. They’re managing all of these things, and these are probably the sort of things that OpenAI these closed labs are doing. We don’t know if they’re doing the exact same techniques, but they actually shared them with the world, which is really nice to be like, this is the cutting edge of efficient language model training.
Lex Fridman
(00:31:49)
And some of this requires low level engineering, just it is a giant mess in trickery. As I understand they went below CUDA, so they go super low programming of GPUs.
Dylan Patel
(00:32:01)
Effectively, Nvidia builds this library called NCCL, in which when you’re training a model, you have all these communications between every single layer of the model, and you may have over a hundred layers.
Nathan Lambert
(00:32:12)
What does NCCL stand for? It’s NCCL.
Dylan Patel
(00:32:14)
Nvidia Communications Collectives Library.
Lex Fridman
(00:32:16)
Nice. Damn.
Dylan Patel
(00:32:18)
And so when you’re training a model, you’re going to have all these allreducers and allgathers, between each layer, between the multilayer perceptron or feed-forward network and the attention mechanism, you’ll have basically the model synchronized. Or you’ll have allreduce and allgather. And this is a communication between all the GPUs in the network, whether it’s in training or inference, so Nvidia has a standard library. This is one of the reasons why it’s really difficult to use anyone else’s hardware for training is because no one’s really built a standard communications library. And Nvidia has done this at a sort of a higher level. DeepSeek because they have certain limitations around the GPUs that they have access to, the interconnects are limited to some extent by the restrictions of the GPUs that were shipped into China legally, not the ones that are smuggled but legally shipped in that they used to train this model, they had to figure out how to get efficiencies. And one of those things is that instead of just calling the NVIDIA library NCCL, they scheduled their own communications, which some of the labs do.

(00:33:27)
Meta talked about in Llama 3, how they made their own custom version of NCCL. They didn’t talk about the implementation details. This is some of what they did, probably not as well as… Maybe not as well as DeepSeek because DeepSeek, necessity is the mother of innovation and they had to do this. OpenAI has people that do this sort of stuff, Anthropic, et cetera. But DeepSeek certainly did it publicly and they may have done it even better because they were gimped on a certain aspect of the chips that they have access to. And so they scheduled communications by scheduling specific SMs. SMs you could think of as the core on a GPU. There’s hundreds of cores or there’s a bit over a hundred cores SMs on a GPU. And they were specifically scheduling, hey, which ones are running the model? Which ones are doing allreduce? Which one are doing allgather? And they would flip back and forth between them. And this requires extremely low level programming.
Nathan Lambert
(00:34:22)
This is what NCCL does automatically or other Nvidia libraries handle this automatically usually.
Dylan Patel
(00:34:26)
Yeah, exactly. And so technically they’re using PTX which is, you could think of it as an assembly type language. It’s not exactly that or instruction set, like coding directly to assembly or instruction set. It’s not exactly that, but that’s still part of technically CUDA. But it’s like, do I want to write in Python, PyTorch equivalent and call Nvidia libraries? Do I want to go down to the C level and code even lower level, or do I want to go all the way down to the assembly or ISO level? And there are cases where you go all the way down there at the very big labs, but most companies just do not do that because it’s a waste of time and the efficiency gains you get are not worth it. But-
Dylan Patel
(00:35:00)
It’s a waste of time and the efficiency gains you get are not worth it. But DeepSeek’s implementation is so complex, especially with their mixture of experts. People have done mixture of experts, but they’re generally eight, 16 experts and they activate two. So, one of the words that we like to use is sparsity factor or usage.

(00:35:19)
So, you might have 1/4th of your model activate, and that’s what Mistral’s Mixtral model, right? They’re a model that really catapulted them to like, “Oh, my God. They’re really, really good.” OpenAI has also had models that are MoE and so have all the other labs that are major closed. But what DeepSeek did that maybe only the leading labs have only just started recently doing is have such a high sparsity factor, right? It’s not 1/4th of the model, right? Two out of eight experts activating every time you go through the model, it’s eight out of 256.
Nathan Lambert
(00:35:51)
And there’s different implementations for mixture of experts where you can have some of these experts that are always activated, which this just looks like a small neural network, and then all the tokens go through that and then they also go through some that are selected by this routing mechanism.

(00:36:08)
And one of the innovations in DeepSeek’s architecture is that they change the routing mechanism and mixture of expert models. There’s something called an auxiliary loss, which effectively means during training, you want to make sure that all of these experts are used across the tasks that the model sees.

(00:36:26)
Why there can be failures in mixture of experts is that when you’re doing this training, one objective is token prediction accuracy. And if you just let turning go with a mixture of expert model on your own, it can be that the model learns to only use a subset of the experts. And in the MoE literature, there’s something called the auxiliary loss which helps balance them.

(00:36:50)
But if you think about the loss functions of deep learning, this even connects to The Bitter Lesson, is that you want to have the minimum inductive bias in your model to let the model learn maximally. And this auxiliary loss, this balancing across experts could be seen as intention with the prediction accuracy of the tokens.

(00:37:09)
So we don’t know the exact extent that the DeepSeek MoE change, which is instead of doing an auxiliary loss, they have an extra parameter in their routing, which after the batches, they update this parameter to make sure that the next batches all have a similar use of experts. And this type of change can be big, it can be small, but they add up over time. And this is the sort of thing that just points to them innovating.

(00:37:31)
And I’m sure all the labs that are training big MoEs are looking at this sort of things, which is getting away from the auxiliary loss. Some of them might already use it, but you keep accumulating gains. And we’ll talk about the philosophy of training and how you organize these organizations. And a lot of it is just compounding small improvements over time in your data, in your architecture, in your post-training and how they integrate with each other.

(00:37:54)
DeepSeek does the same thing and some of them are shared, or a lot. We have to take them on face value that they share their most important details. I mean, the architecture and the weights are out there, so we’re seeing what they’re doing and it adds up.
Dylan Patel
(00:38:05)
Going back to the efficiency and complexity point, right? It’s 32 versus a four, right, for Mixtral and other MoE models that have been publicly released? So this ratio is extremely high. And what Nathan was getting at there was when you have such a different level of sparsity, you can’t just have every GPU have the entire model, right? The model’s too big, there’s too much complexity there. So you have to split up the model with different types of parallelism, right?

(00:38:31)
And so you might have different experts on different GPU nodes, but now what happens when this set of data that you get, “Hey, all of it looks like this one way and all of it should route to one part of my model.” So when all of it routes to one part of the model, then you can have this overloading of a certain set of the GPU resources or a certain set of the GPUs and then the rest of the training network sits idle because all of the tokens are just routing to that.

(00:39:00)
So this is the biggest complexity, one of the big complexities with running a very sparse mixture of experts model i.e., this 32 ratio versus this four ratio, is that you end up with so many of the experts just sitting there idle. So how do I load balance between them? How do I schedule the communications between them? This is a lot of the extremely low-level, detailed work that they figured out in the public first, and potentially second or third in the world and maybe even first in some cases.
Lex Fridman
(00:39:29)
What lesson do you, in the direction of The Bitter Lesson do you take from all of this? Is this going to be the direction where a lot of the gain is going to be, which is this kind of low-level optimization or is this a short-term thing where the biggest gains will be more on the algorithmic high-level side of post-training?

(00:39:50)
Is this a short-term leap because they’ve figured out a hack because constraints necessitate the mother of invention or is there still a lot of gains?
Nathan Lambert
(00:40:01)
I think we should summarize what The Bitter Lesson actually is about, is that The Bitter Lesson essentially, if you paraphrase it, is that the types of training that will win out in deep learning as we go are those methods that which are scalable in learning and search, is what it calls out.

(00:40:20)
The scale word gets a lot of attention in this. The interpretation that I use is effectively to avoid adding the human priors to your learning process. And if you read the original essay, this is what it talks about is how researchers will try to come up with clever solutions to their specific problem that might get them small gains in the short term while simply enabling these deep learning systems to work efficiently, and for these bigger problems in the long term might be more likely to scale and continue to drive success.

(00:40:58)
And therefore, we were talking about relatively small implementation changes to the mixture of experts model. And therefore it’s like, “Okay, we will need a few more years to know if one of these were actually really crucial to The Bitter Lesson,” but The Bitter Lesson is really this long-term arc of how simplicity can often win.

(00:41:17)
And there’s a lot of sayings in the industry, “The models just want to learn. You have to give them the simple loss landscape where you put compute through the model and they will learn, and getting barriers out of the way.”
Lex Fridman
(00:41:29)
That’s where the power of something like nickel comes in, where standardized code that could be used by a lot of people to create simple innovations that can scale, which is why the hacks, I imagine, the code base for DeepSeek is probably a giant mess.
Nathan Lambert
(00:41:45)
I’m sure DeepSeek definitely has code bases that are extremely messy, where they’re testing these new ideas. Multi-head latent attention probably could start in something like a Jupyter Notebook, or somebody tries something on a few GPUs and that is really messy. But the stuff that trains the DeepSeek V3 and DeepSeek-R1, those libraries, if you were to present them to us, I would guess are extremely high-quality code.
Lex Fridman
(00:42:12)
So, high-quality, readable code. Yeah.
Dylan Patel
(00:42:12)
I think there is one aspect to note though is that there is the general ability for that to transfer across different types of runs. You may make really, really high-quality code for one specific model architecture at one size, and then that is not transferable to, ” Hey, when I make this architecture tweak, everything’s broken again,” right?

(00:42:33)
That’s something that could be with their specific low-level coding of scheduling SMs is specific to this model architecture and size. Whereas, Nvidia’s Collectives Library is more like, “Hey, it’ll work for anything,” right? “You want to do an allreduce? Great, I don’t care what your model architecture is, it’ll work,” and you’re giving up a lot of performance when you do that in many cases, but it’s worthwhile for them to do the specific optimization for the specific run given the constraints that they have regarding compute.
Lex Fridman
(00:43:04)
I wonder how stressful it is to these frontier models, like initiate training to have the code-
Dylan Patel
(00:43:12)
Push the button.
Lex Fridman
(00:43:13)
… to push the button that you’re now spending a large amount of money and time to train this. I mean, there must be a lot of innovation on the debugging stage of making sure there’s no issues, that you’re monitoring and visualizing every aspect of the training, all that kind of stuff.
Dylan Patel
(00:43:33)
When people are training, they have all these various dashboards, but the most simple one is your loss, right? And it continues to go down, but in reality, especially with more complicated stuff like MoE, the biggest problem with it, or FP8 training, which is another innovation, going to a lower precision number format i.e., less accurate is that you end up with loss spikes. And no one knows why the loss spike happened. And for a long-
Nathan Lambert
(00:43:55)
Some of them, you do.
Dylan Patel
(00:43:56)
Some of them, you do.
Nathan Lambert
(00:43:56)
Some of them are bad data. Can I give Ai2’s example of what blew up our earlier models is a Subreddit called microwavegang. We love to shout this out. It’s a real thing. You can pull up microwavegang. Essentially it’s a Subreddit where everybody makes posts that are just the letter M. So it’s like, mmm. So there’s extremely long sequences of the letter M and then the comments are like beep beep because it’s in the micro events.
Dylan Patel
(00:44:17)
Yeah.
Nathan Lambert
(00:44:18)
But if you pass this into a model that’s trained to be a normal producing text, it’s extremely high-loss because normally you see an M, you don’t predict Ms for a long time. So this is something that caused loss spikes for us. But when you have much … This is old, this is not recent. And when you have more mature data systems, that’s not the thing that causes the loss spike. And what Dylan is saying is true, but it’s levels to this sort of idea.
Dylan Patel
(00:44:41)
With regards to the stress, these people are like … You’ll go out to dinner with a friend that works at one of these labs and they’ll just be looking at their phone every 10 minutes and they’re not … You know, it’s one thing if they’re texting, but they’re just like, “Is the loss … Is the loss spike okay?”
Nathan Lambert
(00:44:58)
Yeah. It’s like tokens per second. Loss not blown up. They’re just watching this.
Lex Fridman
(00:45:03)
And the heart rate goes up if there’s a spike.
Dylan Patel
(00:45:05)
And some level of spikes is normal, it’ll recover and be back. Sometimes a lot of the old strategy was like, you just stop the run, restart from the old version and then change the data mix and then it keeps going.
Nathan Lambert
(00:45:16)
There are even different types of spikes. So Dirk Groeneveld has a theory today too, that’s like fast spikes and slow spikes, where there are, sometimes where you’re looking at the loss and there are other parameters, you could see it start to creep up and then blow up, and that’s really hard to recover from. So you have to go back much further.

(00:45:31)
So you have the stressful period where it’s flat or it might start going up and you’re like, “What do I do?” Whereas, there are also loss spikes that are, it looks good and then there’s one spiky data point. And what you could do is you just skip those. You see that there’s a spike. You’re like, “Okay, I can ignore this data. Don’t update the model and do the next one, and it’ll recover quickly.”

(00:45:47)
But on trickier implementations, so as you get more complex in your architecture and you scale up to more GPUs, you have more potential for your loss blowing up. So it’s like, there’s a distribution.
Dylan Patel
(00:45:58)
And then the whole idea of grokking also comes in, right? It’s like, just because it slowed down from improving in loss doesn’t mean it’s not learning because all of a sudden it could be like this and it could just spike down in loss again because it truly learned something, right? And it took some time for it to learn that. It’s not a gradual process, and that’s what humans are like. That’s what models are like. So it’s really a stressful task, as you mentioned.
Lex Fridman
(00:46:21)
And the whole time the dollar count is going up.
Nathan Lambert
(00:46:24)
Every company has failed runs. You need failed run to push the envelope on your infrastructure. So, a lot of news cycles are made of X company had Y failed run. Every company that’s trying to push the frontier of AI has these. So yes, it’s noteworthy because it’s a lot of money and it can be week to a month setback, but it is part of the process.
Lex Fridman
(00:46:44)
But if you’re DeepSeek, how do you get to a place where holy shit, there’s a successful combination of hyperparameters?
Nathan Lambert
(00:46:52)
A lot of small failed runs.
Lex Fridman
(00:46:54)
So, rapid iteration through failed runs until-
Nathan Lambert
(00:46:59)
And successful ones.
Lex Fridman
(00:47:01)
And then you build up some intuition, like this mixture of expert works and then this implementation of MLA works.
Nathan Lambert
(00:47:09)
Key hyperparameters, like learning rate and regularization and things like this, and you find the regime that works for your code base. Talking to people at Frontier Labs, there’s a story that you can tell where training language models is kind of a path that you need to follow. So you need to unlock the ability to train a certain type of model or a certain scale, and then your code base and your internal know-how of which hyperparameters work for IT is kind of known.

(00:47:34)
And you look at the DeepSeek papers and models, they’ve scaled up, they’ve added complexity, and it’s just continuing to build the capabilities that they have.
Dylan Patel
(00:47:42)
There’s the concept of a YOLO run. So YOLO, you only live once.
Lex Fridman
(00:47:46)
Yep.
Dylan Patel
(00:47:47)
What it is, is there’s all this experimentation you do at the small scale, research ablations. You have your Jupyter Notebook where you’re experimenting with MLA on three GPUs or whatever and you’re doing all these different things like, “Hey, do I do four active experts, 128 experts? Do I arrange the experts this way?” All these different model architecture things, you’re testing at a very small scale. Right?

(00:48:10)
A couple of researchers, few GPUs, tens of GPUs, hundreds of GPUs, whatever it is. And then all of a sudden you’re like, “Okay, guys. No more fucking around. No more screwing around. Everyone, take all the resources we have. Let’s pick what we think will work and just go for it. YOLO.”

(00:48:26)
And this is where that sort of stress comes in is like, “Well, I know it works here, but some things that work here don’t work here. And some things that work here don’t work down here in this terms of scale.” So it’s really truly a YOLO run. And there’s this discussion of certain researchers just have this methodical nature. They can find the whole search space and figure out all the ablations of different research and really see what is best. And there’s certain researchers who just have that innate gut instinct of like, “This is the YOLO run. I’m looking at the data. I think this is it.”
Nathan Lambert
(00:49:00)
This is why you want to work in post-training because the GPU cost for training is lower. So you can make a higher percentage of your training runs YOLO runs.
Lex Fridman
(00:49:00)
Yeah.
Dylan Patel
(00:49:00)
For now.
Lex Fridman
(00:49:07)
Yeah, for now.
Nathan Lambert
(00:49:08)
For now. For now.
Lex Fridman
(00:49:10)
So some of this is fundamentally luck, still.
Dylan Patel
(00:49:14)
Luck is skill, right, in many cases?
Lex Fridman
(00:49:16)
Yeah. I mean, it looks lucky, right, when you’re-
Nathan Lambert
(00:49:18)
But the hill to climb, if you’re on one of these labs, you have an evaluation you’re not crushing, there’s a repeated playbook of how you improve things. There are localized improvements, which might be data improvements. And these add up into the whole model just being much better.

(00:49:32)
And when you zoom in really close, it can be really obvious that this model is just really bad at this thing and we can fix it and you just add these up. So some of it feels like luck, but on the ground, especially with these new reasoning models we’re talking to is just so many ways that we could poke around. And normally, it’s that some of them give big improvements.
Dylan Patel
(00:49:51)
The search space is near infinite and yet the amount of compute and time you have is very low, and you have to hit release schedules. You have to not get blown past by everyone. Otherwise, what happened with DeepSeek crushing Meta and Mistral and Cohere and all these guys, they moved too slow. They maybe were too methodical. I don’t know, they didn’t hit the YOLO run. Whatever the reason was, maybe they weren’t as skilled. Whatever, you can call it luck if you want, but at the end of the day, it’s skill.
Lex Fridman
(00:50:18)
So 2025 is the year of the YOLO run. It seems like all the labs are going in.
Dylan Patel
(00:50:25)
I think it’s even more impressive what OpenAI did in 2022. At the time, no one believed in mixture of experts models at Google who had all the researchers. OpenAI had such little compute and they devoted all of their compute for many months, all of it, 100% for many months to GPT-4 with a brand-new architecture with no belief that, “Hey, let me spend a couple of hundred million dollars, which is all of the money I have on this model.” That is truly YOLO.
Lex Fridman
(00:50:25)
Yeah.
Dylan Patel
(00:50:52)
Right?
Lex Fridman
(00:50:54)
Yeah.
Dylan Patel
(00:50:55)
Now people have all these training run failures that are in the media, right? It’s like, “Okay, great, but actually a huge chunk of my GPUs are doing inference. I still have a bunch doing research constantly. And yes, my biggest cluster is training, but on this YOLO run,” but that YOLO run is much less risky than what OpenAI did in 2022, or maybe what DeepSeek did now or sort of like, “Hey, we’re just going to throw everything at it.”
Lex Fridman
(00:51:19)
The big winners throughout human history are the ones who are willing to do YOLO at some point. Okay. What do we understand about the hardware it’s been trained on, DeepSeek?

DeepSeek compute cluster

Dylan Patel
(00:51:30)
DeepSeek is very interesting. This is where a second could take to zoom out, out of who they are first of all, right? High-Flyer is a hedge fund that has historically done quantitative trading in China as well as elsewhere. And they have always had a significant number of GPUs, right?

(00:51:45)
In the past, a lot of these high-frequency trading, algorithmic quant traders used FPGAs, but it shifted to GPUs definitely. And there’s both, but GPUs especially. And High-Flyer, which is the hedge fund that owns DeepSeek, and everyone who works for DeepSeek is part of High-Flyer to some extent. Same parent company, same owner, same CEO, they had all these resources and infrastructure for trading, and then they devoted a humongous portion of them to training models, both language models and otherwise, because these techniques were heavily AI-influenced.

(00:52:20)
More recently, people have realized, “Hey, trading with …” Even when you go back to Renaissance and all these quantitative firms, natural language processing is the key to trading really fast, understanding a press release and making the right trade. And so DeepSeek has always been really good at this.

(00:52:38)
And even as far back as 2021, they have press releases and papers saying, “Hey, we’re the first company in China with an A100 cluster this large.” It was 10,000 A100 GPUs, right? This is in 2021. Now, this wasn’t all for training large language models. This was mostly for training models for their quantitative aspects, quantitative trading as well as a lot of that was natural language processing, to be clear. Right?

(00:53:03)
And so this is the sort of history, right? So verifiable fact is that in 2021, they built the largest cluster, at least they claim it was the largest cluster in China, 10,000 GPUs.
Nathan Lambert
(00:53:12)
Before export controls started.
Dylan Patel
(00:53:14)
Yeah.
Nathan Lambert
(00:53:15)
It’s like they’ve had a huge cluster before any conversation of export controls.
Dylan Patel
(00:53:18)
So then you step it forward to, what have they done over the last four years since then? Obviously, they’ve continued to operate the hedge fund, probably make tons of money. And the other thing is that they’ve leaned more and more and more into AI. The CEO, Lian Chingfeng … Lian-
Nathan Lambert
(00:53:33)
You’re not putting me on the spot on this. We discussed this before.
Dylan Patel
(00:53:36)
Lian Feng, right, the CEO, he owns maybe … Lian Feng, he owns maybe a little bit more than half the company allegedly, is an extremely Elon, Jensen kind of figure where he’s just involved in everything. Right?

(00:53:50)
And so over that time period, he’s gotten really in depth into AI. He actually has a bit of a, if you see some of his statements, a bit of an IAK vibe almost, right?
Nathan Lambert
(00:53:59)
Total AGI vibes, like, “We need to do this. We need to make a new ecosystem of OpenAI. We need China to lead on this sort of ecosystem because historically, the western countries have led on software ecosystems.” And straight up acknowledges, “In order to do this, we need to do something different.” DeepSeek is his way of doing this. Some of the translated interviews with him are fantastic.
Lex Fridman
(00:54:23)
So he has done interviews?
Nathan Lambert
(00:54:24)
Yeah.
Lex Fridman
(00:54:24)
Do you think you would do a western interview, or no, or is there controls on the channel?
Nathan Lambert
(00:54:28)
There hasn’t been one yet, but I would try it.
Lex Fridman
(00:54:32)
Okay. All right. Well, I just got a Chinese translator, so it was great. This is a push. So fascinating figure, engineer pushing full on into AI, leveraging the success from the high-frequency trading.
Nathan Lambert
(00:54:44)
Very direct quotes. “We will not switch to closed source,” when asked about this stuff. Very long-term motivated in how the ecosystem of AI should work. And I think from a Chinese perspective, he wants a Chinese company to build this vision.
Dylan Patel
(00:55:03)
And so this is sort of like the “visionary behind the company.” This hedge fund still exists, this quantitative firm. And so DeepSeek is the sort of … Slowly, he got turned to this full view of AI, everything about this, but at some point it slowly maneuvered and he made DeepSeek.

(00:55:20)
And DeepSeek has done multiple models since then. They’ve acquired more and more GPUs. They share infrastructure with the fund. Right? And so there is no exact number of public GPU resources that they have. But besides this 10,000 GPUs that they bought in 2021, and they were fantastically profitable, and then this paper claims they did only 2,000 H800 GPUs, which are a restricted GPU that was previously allowed in China, but no longer allowed. And there’s a new version, but it’s basically Nvidia’s H100 for China.

(00:55:52)
And there’s some restrictions on it specifically around the communications sort of speed, the interconnect speed, which is why they had to do this crazy SM scheduling stuff. So going back to that, it’s like this is obviously not true in terms of their total GPU count.
Lex Fridman
(00:56:08)
Obvious available GPUs, but for this training run, you think 2,000 is the correct number, or no?
Dylan Patel
(00:56:14)
So this is where it takes a significant amount of zoning in. What do you call your training run, right? You count all of the research and ablations that you ran, right? Picking all this stuff because yes, you can do a YOLO run, but at some level you have to do the test at the small scale and then you have to do some test at medium scale before you go to a large scale.
Nathan Lambert
(00:56:33)
Accepted practice is that for any given model that is a notable advancement, you’re going to do two to 4x compute of the full training run in experiments alone.
Lex Fridman
(00:56:43)
So a lot of this compute that’s being scaled up is probably used in large part at this time for research?
Dylan Patel
(00:56:49)
Yeah. And research begets the new ideas that lets you get huge efficiency.
Nathan Lambert
(00:56:53)
Research gets you o1. Research gets you breakthroughs and you need to bet on it.
Lex Fridman
(00:56:57)
So some of the pricing strategy that we’ll discuss has the research baked into the price?
Dylan Patel
(00:57:02)
So the numbers that DeepSeek specifically said publicly are just the 10,000 GPUs in 2021 and then 2,000 GPUs for only the pre-training for V3. They did not discuss cost on R1. They did not discuss cost on all the other RL for the instruct model that they made. They only discussed the pre-training for the base model and they did not discuss anything on research and ablations. And they do not talk about any of the resources that are shared in terms of, “Hey, the fund is using all these GPUs,” right?

(00:57:31)
And we know that they’re very profitable and they had 10,000 GPUs in 2021. So, some of the research that we’ve found is that we actually believe they have closer to 50,000 GPUs.
Lex Fridman
(00:57:43)
We as semi-analysis. So we should say that you’re sort of one of the world experts in figuring out what everybody’s doing in terms of the semiconductor, in terms of cluster buildouts, in terms of who is doing what in terms of training runs. So yeah, that’s the we. Okay, go ahead.
Dylan Patel
(00:57:59)
Yeah, sorry. We believe they actually have something closer to 50,000 GPUs, right? Now this is split across many tasks, right? Again, the fund, research and ablations.
Nathan Lambert
(00:58:09)
For ballpark, how much would OpenAI or Anthropic had. I think the clearest example we have, because Meta is also open, they talk about order of 60k to 100k H100 equivalent GPUs in their training clusters.
Dylan Patel
(00:58:21)
Right. So Llama 3, they trained on 16,000 H100s, but the company of Meta last year publicly disclosed they bought 400 something thousand GPUs.
Nathan Lambert
(00:58:21)
Yeah.
Dylan Patel
(00:58:30)
Right? So of course, tiny percentage on the training. Again, most of it is serving me the best Instagram Reels or whatever.
Nathan Lambert
(00:58:37)
I mean, we could get into a cost of, what is the cost of ownership for a 2,000 GPU cluster, 10,000? There’s just different sizes of companies that can afford these things and DeepSeek is reasonably big. Their compute allocation is one of the top few in the world that’s not OpenAI, Anthropic, et cetera, but they have a lot of compute.

Export controls on GPUs to China

Lex Fridman
(00:58:58)
Can you in gentlemen actually just zoom out and also talk about the Hopper architecture, the Nvidia Hopper GPU architecture and the difference between H100 and H800, like you mentioned, the interconnects?
Dylan Patel
(00:59:09)
Yeah. So there’s, Ampere was the A100 and then H100 Hopper, right? People use them synonymously in the U.S. because really there’s just H100 and now there’s H200, right, but same thing mostly?

(00:59:21)
In China, there’ve been different salvos of expert restrictions. So initially, the U.S. government limited on a two-factor scale, which is chip interconnect versus FLOPs. So any chip that had interconnects above a certain level and FLOPs above a certain … Floating point operations above a certain level was restricted.

(00:59:38)
Later, the government realized that this was a flaw in the restriction and they cut it down to just floating point operations. And so-
Nathan Lambert
(00:59:48)
H800 had high FLOPs, low communication?
Dylan Patel
(00:59:51)
Exactly. So, the H800 was the same performance as H100 on FLOPs, but it just had the interconnect bandwidth cut. DeepSeek knew how to utilize this. “Hey, even though we’re cut back on the interconnect, we can do all this fancy stuff to figure out how to use the GPU fully anyways.”

(01:00:09)
And so that was back in October 2022. But later in 2023, into 2023 implemented in 2024, the U.S. government banned the H800. Right? And so by the way, this H800 cluster, these 2,000 GPUs was not even purchased in 2024. It was purchased in late 2023. And they’re just getting the model out now because it takes a lot of research, et cetera.

(01:00:31)
H800 was banned and now there’s a new chip called the H20. The H20 is cut back on only FLOPs, but the interconnect bandwidth is the same. And in fact, in some ways it’s better than the H100 because it has better memory bandwidth and memory capacity. So Nvidia is working within the constraints of what the government sets and then builds the best possible GPU for China.
Lex Fridman
(01:00:52)
Can we take this actual tangent and we’ll return back to the hardware, is the philosophy, the motivation, the case for export controls? What is it? Dario Amodei just published a blog post about export controls. The case he makes is that if AI becomes super powerful and he says by 2026, we’ll have AGI or super powerful AI and that’s going to give a significant … Whoever builds that will have a significant military advantage.

(01:01:19)
And so because The United States is a democracy and as he says, China is authoritarian or has authoritarian elements, you want a unipolar world where the super powerful military, because of the AI is one that’s a democracy. It’s a much more complicated world geopolitically when you have two superpowers with super powerful AI and one is authoritarian.

(01:01:46)
So, that’s the case he makes. And so the United States wants to use export controls to slow down, to make sure that China can’t do these gigantic training runs that will be presumably required to build the AGI.
Nathan Lambert
(01:02:02)
This is very abstract. I think this can be the goal of how some people describe export controls, is this super powerful AI. And you touched on the training run idea. There’s not many worlds where China cannot train AI models. I think export controls are decapping the amount of compute or the density of compute that China can have.

(01:02:25)
And if you think about the AI ecosystem right now, as all of these AI companies, revenue numbers are up and to the right. Their AI usage is just continuing to grow, more GPUs are going to inference. A large part of export controls, if they work is just that the amount of AI that can be run in China is going to be much lower.

(01:02:45)
So on the training side, DeepSeek V3 is a great example, which you have a very focused team that can still get to the frontier of AI on … This 2,000 GPUs is not that hard to get all considering in the world. They’re still going to have those GPUs. They’re still going to be able to train models. But if there’s going to be a huge market for AI, if you have strong export controls and you want to have 100,000 GPUs just serving the equivalent of ChatGPT clusters with good export controls, it also just makes it so that AI can be used much less.

(01:03:13)
And I think that is a much easier goal to achieve than trying to debate on what AGI is. And if you have these extremely intelligent autonomous AIs and data centers, those are the things that could be running in these GPU clusters in the United States, but not in China.
Dylan Patel
(01:03:30)
To some extent, training a model does effectively nothing. They have a model. The thing that Dario is sort of speaking to is the implementation of that model, once trained to then create huge economic growth, huge increases in military capabilities, huge increases in productivity of people, betterment of lives. Whatever you want to direct super powerful AI towards, you can, but that requires a significant amounts of compute.

(01:03:56)
And so the U.S. government has effectively said … And forever, training will always be a portion of the total compute. We mentioned Meta’s 400,000 GPUs. Only 16,000 made Llama. Right? So the percentage that Meta’s dedicating to inference, now this might be for recommendation systems that are trying to hack our mind into spending more time and watching more ads, or if it’s for a super powerful AI that’s doing productive things, it doesn’t matter about the exact use that our economic system decides. It’s that, that can be delivered in whatever way we want.

(01:04:28)
Whereas with China, you know, your expert restrictions, great. You’re never going to be able to cut everything off. And I think that’s quite a well-understood by the U.S. government is that you can’t cut everything off.
Nathan Lambert
(01:04:40)
And they’ll make their own chips.
Dylan Patel
(01:04:42)
And they’re trying to make their own chips. They’ll be worse than ours, but the whole point is to just keep a gap. And therefore at some point, as the AI … In a world where two, 3% economic growth, this is really dumb by the way, to cut off high-tech and not make money off of it. But in a world where super powerful AI comes about and then starts creating significant changes in society, which is what all the AI leaders and big tech companies believe. I think super powerful AI is going to change society massively.

(01:05:08)
And therefore, this compounding effect of the difference in compute is really important. There’s some sci-fi out there where AI is measured in how much power is delivered to compute, right, or how much is being … That’s sort of a way of thinking about what’s the economic output, is just how much power are you directing towards that AI?
Nathan Lambert
(01:05:26)
Should we talk about reasoning models with this, as a way that this might be actionable as something that people can actually see? So, the reasoning models that are coming out with R1 and o1, they’re designed to use more compute. There’s a lot of buzzy words in the AI community about this, test-time compute, inference time compute, whatever.

(01:05:44)
But Dylan has good research on this. You can get to the specific numbers on the ratio of when you train a model, you can look at things. It’s about the amount of compute used at training and amount of compute used at inference.

(01:05:53)
These reasoning models are making inference way more important to doing complex tasks. In the fall in December, OpenAI announced this o3 model. There’s another thing in AI, when things move fast, we get both announcements and releases. Announcements are essentially blog posts where you pat yourself on the back and you say you did things and releases are when the model’s out there, the paper’s out there, et cetera.

(01:06:12)
So OpenAI has announced o3. I mean, we can check if o3-mini is out as of recording potentially, but that doesn’t really change the point, which is that the breakthrough result was something called ARC-AGI task, which is the abstract reasoning corpus, a task for artificial general intelligence. François Chollet is the guy who’s been … It’s a multi-year-old paper. It’s a brilliant benchmark. And the number for open AI o3 to solve this was that it used some sort of number of samples in the API. The API has thinking effort and number of samples. They used 1,000 samples to solve this task and it comes out to be five to $20 per question, which you’re putting in effectively a math puzzle. And then it takes orders of dollars to answer one question, and this is a lot of compute.

(01:07:00)
If those are going to take off in the U.S., OpenAI needs a ton of GPUs on inference to capture this. They have this OpenAI ChatGPT Pro subscription, which is $200 a month-
Dylan Patel
(01:07:09)
Which Sam said they’re losing money on.
Nathan Lambert
(01:07:11)
Which means that people are burning a lot of GPUs on inference. And I’ve signed up with it, I’ve played with it. I don’t think I’m a power user, but I use it. And it’s like, that is the thing that a Chinese company with mediumly strong expert controls, there will always be loopholes, might not be able to do it all.

(01:07:27)
And if the main result for o3 is also a spectacular coding performance, and if that feeds back into AI companies being able to experiment better.
Lex Fridman
(01:07:37)
So presumably, the idea is for an AGI, a much larger fraction of the compute would be used for this test-time compute, for the reasoning, for the AGI goes into a room and thinks about how to take over the world and come back in 2.7 hours-
Nathan Lambert
(01:07:54)
This is what-
Lex Fridman
(01:07:55)
… and that it’s going to take a lot of compute.
Nathan Lambert
(01:07:56)
This is what people, CEO or leaders of OpenAI and Anthropic talk about, is autonomous AI models, which is you give them a task and they work on it in the background.

(01:08:05)
I think my personal definition of AGI is much simpler. I think language models are a form of AGI and all of this super powerful stuff is a next step that’s great if we get these tools. But a language model has so much value in so many domains that it’s a general intelligence to me.

(01:08:21)
But this next step of agentic things where they’re independent and they can do tasks that aren’t in the training data is what the few-year outlook that these AI companies are driving for.
Lex Fridman
(01:08:32)
I think the terminology here that Dario uses is super powerful AI. So I agree with you on the AGI. I think we already have something like that’s exceptionally impressive that Alan Turing would for sure say is AGI, but he’s referring more to something once in possession of, then you would have a significant military and geopolitical advantage over other nations. So it’s not just like you can ask it how to cook an omelet.
Nathan Lambert
(01:08:58)
And he has a much more positive view. And as I say, machines of love and grace. I read into this and I don’t have enough background in physical sciences to gauge exactly how competent I am, and if AI can revolutionize biology. I am safe saying that AI is going to accelerate the progress of any computational science.

AGI timeline

Lex Fridman
(01:09:16)
So we’re doing a depth-first search here on topics, taking tangent of a tangent, so let’s continue on that depth-first search. You said that you’re both feeling the AGI. What’s your timeline? Dario is 2026 for the super powerful AI that’s basically agentic to a degree where it’s a real security threat, that level of AGI. What’s your timeline?
Nathan Lambert
(01:09:44)
I don’t like to attribute specific abilities because predicting specific abilities and when is very hard. I think mostly if you’re going to say that I’m feeling the AGI is that I expect continued, rapid, surprising progress over the next few years. So, something like R1 is less surprising to me from DeepSeek because I expect there to be new paradigms versus …
Nathan Lambert
(01:10:00)
… surprising to me from DeepSeek because I expect there to be new paradigms where substantial progress can be made. I think DeepSeek-R1 is so unsettling because we’re kind of on this path with ChatGPT. It’s like it’s getting better, it’s getting better, it’s getting better, and then we have a new direction for changing the models, and we took one step like this and we took a step-up. So it looks like a really fast slope, and then we’re going to just take more steps. So it’s just really unsettling when you have these big steps, and I expect that to keep happening. I’ve tried opening Operator, I’ve tried Claude computer use, they’re not there yet. I understand the idea, but it’s just so hard to predict what is the breakthrough that’ll make something like that work. And I think it’s more likely that we have breakthroughs that work in things that we don’t know what they’re going to do. So everyone wants agents. Dario has a very eloquent way of describing this, and I just think that it’s like there’s going to be more than that, so just expect these things to come.
Lex Fridman
(01:10:53)
I’m going to have to try to pin you down to a date on the AGI timeline. Like the nuclear weapon moment, so moment where on the geopolitical stage, there’s a real… Because we’re talking about export controls, when do you think, just even to throw out a date, when do you think that would be? For me, it’s probably after 2030, so I’m not as-
Nathan Lambert
(01:11:19)
That’s what I would say.
Dylan Patel
(01:11:21)
So define that. Because to me, it kind of almost has already happened. You look at elections in India and Pakistan, people get AI voice calls and think they’re talking to the politician. The AI diffusion rules, which was enacted in the last couple of weeks of the Biden admin, it looks like the Trump admin will keep and potentially even strengthen, limit cloud computing and GPU sales to countries that are not even related to China. It’s like this is-
Nathan Lambert
(01:11:44)
Portugal and all these normal countries are on the… You need approval from the US list.
Dylan Patel
(01:11:49)
Yeah, Portugal and all these countries that are allies. Singapore. They freaking have F-35s and we don’t let them buy GPUs. This to me is already to the scale of…
Lex Fridman
(01:12:02)
Well, that just means that the US military is really nervous about this new technology. That doesn’t mean that technology is already there. So they might be just very cautious about this thing that they don’t quite understand. But that’s a really good point. The robocalls, swarms of semi-intelligent bots could be a weapon, could be doing a lot of social engineering.
Dylan Patel
(01:12:25)
I mean, there’s tons of talk about from the 2016 elections like Cambridge Analytica and all this stuff, Russian influence. I mean, every country in the world is pushing stuff onto the internet and has narratives they want. Every technically competent, whether it’s Russia, China, US, Israel, et cetera. People are pushing viewpoints onto the internet en masse. And language models crash the cost of very intelligent sounding language.
Nathan Lambert
(01:12:49)
There’s some research that shows that the distribution is actually the limiting factor. So language models haven’t yet made misinformation particularly change the equation there. The internet is still ongoing. I think there’s a blog, AI Snake Oil and some of my friends at Princeton that write on this stuff. So there is research. It’s a default that everyone assumes. And I would’ve thought the same thing, is that misinformation doesn’t get far worse with language models. I think in terms of internet posts and things that people have been measuring, it hasn’t been a exponential increase or something extremely measurable and things you’re talking about with voice calls and stuff like that, it could be in modalities that are harder to measure.

(01:13:27)
So it’s something that it’s too soon to tell in terms of… I think that’s political instability via the web is very… It’s monitored by a lot of researchers to see what’s happening. I think that… You’re asking about the AGI thing. If you’re making me give a year, I’m going to be like, “Okay, I have AI CEOs saying this. They’ve been saying two years for a while. I think that there are people like Dario at Anthropic, the CEO, has thought about this so deeply. I need to take their word seriously, but also understand that they have different incentives.” So I would be like, “Add a few years to that.” Which is how you get something similar to 2030 or a little after 2030.
Dylan Patel
(01:14:08)
I think to some extent, we have capabilities that hit a certain point where any one person could say, “Oh, okay, if I can leverage those capabilities for X amount of time, this is AGI, call it ’27, ’28.” But then the cost of actually operating that capability-
Nathan Lambert
(01:14:23)
Yeah, this was going to be my point.
Dylan Patel
(01:14:24)
… is so, so extreme that no one can actually deploy it at scale en masse to actually completely revolutionize the economy on a snap of a finger. So I don’t think it will be a snap of the finger moment.
Nathan Lambert
(01:14:35)
It’s a physical constraint [inaudible 01:14:37].
Dylan Patel
(01:14:36)
Rather, it’ll be a, “Oh, the capabilities are here, but I can’t deploy it everywhere.” And so one simple example, going back sort of to 2023 was when being when GPT-4 came out, everyone was freaking out about search. Perplexity came out. If you did the cost on like, hey, implementing GPT-3 into every Google search, it was like, oh, okay, this is just physically impossible to implement. And as we step forward to going back to the test-time compute thing, a query for… You ask ChatGPT a question, it costs cents for their most capable model of Chat to get a query back. To solve an AGI problem though costs 5 to 20 bucks, and this is in-
Nathan Lambert
(01:15:17)
It’s only going up from there.
Dylan Patel
(01:15:19)
This is 1,000, 10,000 X factor difference in cost to respond to a query versus do a task. And the task of AGI is not like it’s like… It’s simple, to some extent, but it’s also like, what are the tasks that we want… Okay, AGI, “What we have today”, can do AGI. Three years from now, it can do much more complicated problems, but the cost is going to be measured in thousands and thousands and hundreds of thousands of dollars of GPU time, and there just won’t be enough power, GPUs, infrastructure to operate this and therefore shift everything in the world on the snap the finger.

(01:15:52)
But at that moment, who gets to control and point the AGI at a task? And so this was in Dario’s post that he’s like, “Hey, China can effectively and more quickly than us, point their AGI at military tasks.” And they have been, in many ways, faster at adopting certain new technologies into their military, especially with regards to drones. The US maybe has a long-standing large air sort of fighter jet type of thing, bombers. But when it comes to asymmetric arms such as drones, they’ve completely leapfrogged the US and the West.

(01:16:28)
And the fear that Dario is sort of pointing out there, I think, is that, yeah, great, we’ll have AGI in the commercial sector. The US military won’t be able to implement it superfast. Chinese military could and they could direct all their resources to implementing it in the military, and therefore solving military logistics or solving some other aspect of disinformation for targeted certain set of people so they can flip a country’s politics or something like that that is actually catastrophic versus the US just wants to… Because it’ll be more capitalistically allocated just towards whatever is the highest return on income, which might be building factories better or whatever.
Lex Fridman
(01:17:04)
So everything I’ve seen, people’s intuition seems to fail on robotics. So you have this kind of general optimism. I’ve seen this on self-driving cars. People think it’s much easier problem than it is. Similar with drones, here, I understand it a little bit less, but I’ve just seen the reality of the war in Ukraine and the usage of drones on both sides. And it seems that humans still far outperform any fully autonomous systems. AI is an assistant, but humans drive. FPV drones where the human’s controlling most of it, just far, far, far outperforms AI systems. So I think it’s not obvious to me that we’re going to have swarms of autonomous robots anytime soon in the military context. Maybe the fastest I can imagine is 2030, which is why I said 2030 for the super powerful AI. Whenever you have large scale swarms of robots doing military actions, that’s when the world just starts to look different to me.

(01:18:07)
So that’s the thing I’m really worried about. But there could be cyber war, cyber war type of technologies that from social engineering to actually just swarms of robots that find attack vectors in our code bases and shut down power grids, that kind of stuff. And it could be one of those things like on any given weekend or something, power goes out, nobody knows why, and the world changes forever. Just power going out for two days in all of the United States, that will lead to murder, to chaos. But going back to export controls, do you see that as a useful way to control the balance of power geopolitically in the context of AI?

China’s manufacturing capacity

Dylan Patel
(01:18:56)
And I think going back to my viewpoint is if you believe we’re in this sort of stage of economic growth and change that we’ve been in for the last 20 years, the export controls are absolutely guaranteeing that China will win long-term. If you do not believe AI is going to make significant changes to society in the next 10 years or 5 years. Five-year timelines are sort of what the more executives and such of AI companies and even big tech companies believe. But even 10-year timelines, it’s reasonable. But once you get to, hey, these timelines are below that time period, then the only way to create a sizable advantage or disadvantage for America versus China is if you constrain and compute, because talent is not really something that’s constraining. China arguably has more talent, more STEM graduates, more programmers. The US can draw upon the world’s people, which it does. There’s tons of foreigners in the AI industry.
Nathan Lambert
(01:19:57)
So many of these AI teams are all people without a US passport.
Dylan Patel
(01:20:02)
Yeah. I mean, many of them are Chinese people who are moving to America, and that’s great. That’s exactly what we want. But that talent is one aspect, but I don’t think that’s one that is a measurable advantage for the US or not. It truly is just whether or not compute. Now, even on the compute side, when we look at chips versus data centers, China has the unprecedented ability to build ridiculous sums of power. Clockwork. They’re always building more and more power. They’ve got steel mills that individually are the size of the entire US industry. And they’ve got aluminum mills that consume gigawatts and gigawatts of power. And when we talk about what’s the biggest data center, OpenAI made this huge thing about Stargate, their announcement there, once it’s fully built out in a few years, it’ll be two gigawatts of power. And this is still smaller than the largest industrial facilities in China. China, if they wanted to build the largest data center in the world, if they had access to the chips, could. So it’s just a question of when, not if.
Lex Fridman
(01:21:07)
So their industrial capacity far exceeds the United States’?
Dylan Patel
(01:21:10)
Exactly.
Lex Fridman
(01:21:11)
They manufacture stuff. So long-term, they’re going to be manufacturing chips there?
Dylan Patel
(01:21:18)
Chips are a little bit more specialized. I’m specifically referring to the data centers. Fabs take huge amounts of power, don’t get me wrong. That’s not necessarily the gating factor there. The gating factor on how fast people can build the largest clusters today in the US is power. Now, it could be power generation, power transmission, substations, and all these sorts of transformers and all these things building the data center. These are all constraints on the US industry’s ability to build larger and larger training systems, as well as deploying more and more inference compute.
Nathan Lambert
(01:21:52)
I think we need to make a point clear on why the time is now for people that don’t think about this, because essentially, with export controls, you’re making it so China cannot make or get cutting edge chips. And the idea is that if you time this wrong, China is pouring a ton of money into their chip production, and if you time it wrong, they are going to have more capacity for production, more capacity for energy, and figure out how to make the chips and have more capacity than the rest of the world to make the chips. Because everybody can buy… They’re going to sell their Chinese chips to everybody, they might subsidize them. And therefore, if AI takes a long time to become differentiated, we’ve kneecapped the financial performance of American companies. NVIDIA can sell less, TSMC cannot sell to China. So therefore, we have less demand to therefore… To keep driving the production cycle. So that’s the assumption behind the timing being [inaudible 01:22:43].
Dylan Patel
(01:22:43)
Less than 10 years or 5 years to above. China will win because of these restrictions long-term, unless AI does something in the short-term, which I believe AI will make massive changes to society in the medium, short-term. And so that’s the big unlocker there. And even today, if Xi Jinping decided to get “scale-pilled”, IE, decide that scaling laws are what matters, just like the US executives like Satya Nadella and Mark Zuckerberg and Sundar and all these US executives of the biggest, most powerful tech companies have decided they’re scale-pilled and they’re building multi-gigawatt data centers, whether it’s in Texas or Louisiana or Wisconsin, wherever it is, they’re building these massive things that cost as much as their entire budget for spending on data centers globally in one spot. This is what they’ve committed to for next year, year after, et cetera. And so they’re so convinced that this is the way that this is what they’re doing.

(01:23:43)
But if China decided to, they could do it faster than us, but this is where the restrictions come in. It is not clear that China as a whole has decided from the highest levels that this is a priority. The US sort of has. You see Trump talking about DeepSeek and Stargate within the same week. And the Biden admin as well had a lot of discussions about AI and such. It’s clear that they think about it. Only just last week did DeepSeek meet the second in command of China. They have not even met the top, they haven’t met Xi, Xi hasn’t set down, and they only just released a subsidy of a trillion RMB, roughly $160 billion, which is closer to the spending of Microsoft and Meta and Google combined for this year. So they’re realizing it just now. But that’s where these export restrictions come in and say, “Hey, you can’t ship the most powerful US chips to China. You can ship a cut-down version. You can’t ship the most powerful chips to all these countries who we know are just going to rent it to China. You have to limit the numbers.”
Nathan Lambert
(01:24:48)
And the tools.
Dylan Patel
(01:24:50)
And same with manufacturing [inaudible 01:24:52] tools, all these different aspects, but it all stems from AI and then what downstream can slow them down in AI. And so the entire semiconductor restrictions, you read them, they’re very clear, it’s about AI and military civil fusion of technology. It’s very clear. And then from there it goes, oh, well, we’re banning them from buying lithography tools and etch tools and deposition tools. And oh, this random subsystem from a random company that’s tiny. Why are we banning this? Because all of it, the US government has decided is critical to AI systems.
Nathan Lambert
(01:25:23)
I think the fulcrum point is the transition from seven nanometer to five nanometer chips where I think it was Huawei that had the seven nanometer chip a few years ago, which caused another political brouhaha, almost like this moment. And then it’s the ASML deep UV. What is that… Extreme ultraviolet lithography.
Dylan Patel
(01:25:43)
Just set context on the chips. What Nathan’s referring to is in 2020, Huawei released their Ascend 910 chip, which was an AI chip, first one on seven nanometer before Google did, before NVIDIA did. And they submitted it to the MLPerf benchmark, which is sort of a industry standard for machine learning performance benchmark, and it did quite well, and it was the best chip at the submission. This was a huge deal. The Trump admin, of course, banned, it was 2019, banned the Huawei from getting seven nanometer chips from TSMC. And so then they had to switch to using internal, domestically produced chips, which was a multi-year setback.
Nathan Lambert
(01:26:20)
Many companies have done seven nanometer chips. And the question is we don’t know how much Huawei was subsidizing production of that chip. Intel has made seven nanometer chips that are not profitable and things like this. So this is how it all feeds back into the economic engine of export controls.

Cold war with China

Lex Fridman
(01:26:36)
Well, so you’re saying that for now, Xi Jinping has not felt the AGI, but it feels like the DeepSeek moment, there might be meetings going on now where he’s going to start wearing the same t-shirt and things are going to escalate.
Dylan Patel
(01:26:52)
I mean, he may have woken up last week. Liang Feng met the second command guy, and they had a meeting, and then the next day, they announced the AI subsidies, which are a trillion RMB.
Lex Fridman
(01:27:05)
So it’s possible that this DeepSeek moment is truly the beginning of a cold war.
Nathan Lambert
(01:27:10)
That’s what a lot of people are worried about. People in AI have been worried that this is going towards a cold war or already is.
Lex Fridman
(01:27:16)
But it’s not DeepSeek’s fault, but there’s something, a bunch of factors came together where-
Nathan Lambert
(01:27:16)
It’s how history works.
Lex Fridman
(01:27:21)
… it’s like this explosion. I mean, it all has to do with NVIDIA’s not going down properly, but it’s just some [inaudible 01:27:28] mass hysteria that happened that eventually led to Xi Jinping having meetings and waking up to this idea.
Dylan Patel
(01:27:34)
And the US government realized in October 7th, 2022, before ChatGPT released, that restriction on October 7th, which dropped and shocked everyone, and it was very clearly aimed at AI. Everyone was like, “What the heck are you doing?”
Nathan Lambert
(01:27:48)
Stable Diffusion was out then, but not ChatGPT.
Dylan Patel
(01:27:48)
Yeah, but not ChatGPT.
Nathan Lambert
(01:27:51)
So it was starting to be rumblings-
Dylan Patel
(01:27:53)
Of what GenAI can do to society, but it was very clear, I think, to at least National Security Council and those sort of folks, that this was where the world is headed, this cold war that’s happening.
Lex Fridman
(01:28:04)
So is there any concerns that the export controls push China to take military action on Taiwan?
Dylan Patel
(01:28:15)
This is the big risk. The further you push China away from having access to cutting edge American and global technologies, the more likely they are to say, “Well, because I can’t access it, I might as well… No one should access it.” And there’s a few interesting aspects of that. China has a urban-rural divide like no other. They have a male-female birth ratio like no other to the point where if you look in most of China, it’s like the ratio is not that bad. But when you look at single dudes in rural China, it’s like a 30:1 ratio. And those are disenfranchised dudes. “The US has an incel problem.” China does too, it’s just they’re placated in some way or crushed down. What do you do with these people? And at the same time, you’re not allowed to access the most important technology, at least the US thinks so. China’s maybe starting to think this is the most important technology by starting to dump subsidies in it.

(01:29:07)
They thought EVs and renewables were the most important technology. They dominate that now. Now, they started thinking about semiconductors in the late 2010s and early 2020s and now they’ve been dumping money and they’re catching up rapidly and they’re going to do the same with AI because they’re very talented. So the question is, when does this hit a breaking point? And if China sees this as, “Hey, they can continue…” If not having access and starting a true hot war, taking over Taiwan or trying to subvert its democracy in some way or blockading it hurts the rest of the world far more than it hurts them, this is something they could potentially do. And so is this pushing them towards that? Potentially. I’m not quite a geopolitical person, but it’s obvious that the world regime of peace and trade is super awesome for economics, but at some point, it could break.
Nathan Lambert
(01:30:07)
I think we should comment the why Chinese economy would be hurt by that is that they’re export heavy, I think. United States buys so much. If that goes away, that’s how their economy [inaudible 01:30:17].
Dylan Patel
(01:30:16)
Well, also, they just would not be able to import raw materials from all over the world. The US would just shut down the Strait of Malacca. And at the same time, the US entire… You could argue almost all the GDP growth in America since the ’70s has been either population growth or tech, because your life today is not that much better than someone from the ’80s outside of tech. Cars, they all have semiconductors in them everywhere. Fridges, semiconductors everywhere. There’s these funny stories about how Russians were taking apart laundry machines because they had certain Texas Instrument chips that they could then repurpose and put into their anti-missile missile things, like their S-400 or whatever. You would know more about this, but there’s all sorts of… Everything about semiconductors is so integral to every part of our lives.

TSMC and Taiwan

Lex Fridman
(01:31:06)
So can you explain the role of TSMC in the story of semiconductors and maybe also how the United States can break the reliance on TSMC?
Dylan Patel
(01:31:17)
I don’t think it’s necessarily breaking the reliance. I think it’s getting TSMC to build in the US. So taking a step back, TSMC produces most of the world’s chips, especially on the foundry side. There’s a lot of companies that build their own chips. Samsung, Intel, STMicro, Texas Instruments, Analog Devices, all these kinds of companies build their own chips, and XP, but more and more of these companies are outsourcing to TSMC and have been for multiple decades.
Lex Fridman
(01:31:49)
Can you explain the supply chain there and where most of TSMC is in terms of manufacturing?
Dylan Patel
(01:31:55)
Sure. So historically, supply chain was companies would build their own chips. It would be a company started, they’d build their own chips, and then they’d design the chip and build the chip and sell it. Over time, this became really difficult because the cost of building a fab continues to compound every single generation. Of course, figuring out the technology for it is incredibly difficult regardless, but just the dollars and cents that are required, ignoring, saying, “Hey, yes, I have all the technical capability.” Which it’s really hard to get that by the way. Intel’s failing, Samsung’s failing, et cetera. But if you look at just the dollars to spend to build that next-generation fab, it keeps growing. Sort of Moore’s law is having the cost of chips every two years. There’s a separate law that’s sort of doubling the cost of fabs every handful of years.

(01:32:38)
And so you look at a leading-edge fab that is going to be profitable today, that’s building three nanometer chips or two nanometer chips in the future, that’s going to cost north of 30, $40 billion. And that’s just for a token amount. That’s like the base building blocking. You probably need to build multiple. And so when you look at the industry over the last, if I go back 20, 30 years ago, there were 20, 30 companies that could build the most advanced chips, and then they would design them themselves and sell them. So companies like AMD would build their own chips. Intel, of course, still builds their own chips. They’re very famous for it. IBM would build their own chips. And you could just keep going down the list. All these companies built their own chips.

(01:33:14)
Slowly, they kept falling like flies, and that’s because of what TSMC did. They created the Foundry business model, which is, I’m not going to design any chips. I’m just going to contract manufacturer chips for other people. And one of their early customers is NVIDIA. NVIDIA is the only semiconductor company doing more than $1 billion of revenue that was started in the era of foundry. Every other company started before then, and at some point had fabs, which is actually incredible. Like AMD and Intel and Broadcom-
Lex Fridman
(01:33:48)
[inaudible 01:33:48].
Dylan Patel
(01:33:48)
Everyone had fabs at some point, or some companies like Broadcom. It was like a merger amalgamation of various companies that rolled up. But even today, Broadcom has fabs. They build iPhone, RF radio chips in Colorado for Apple. All these companies had fabs, and for most of the fabs, they threw them away or sold them off, or they got rolled into something else. And now, everyone relies on TSMC. Including Intel, their latest PC chip uses TSMC chips. It also uses some Intel chips, but it uses TSMC process.
Lex Fridman
(01:34:19)
Can you explain why the foundry model is so successful for these companies? Why are they going with-
Nathan Lambert
(01:34:24)
Economies of scale.
Lex Fridman
(01:34:26)
Scale?
Dylan Patel
(01:34:26)
Yeah. So I mean, like I mentioned, the cost of building a fab is so high, the R&D is so difficult. And when you look at these companies that had their own vertical stack, there was an antiquated process of like, okay, I’m so hyper customized to each specific chip, but as we’ve gone through the history of the last 50 years of electronics and semiconductors, A, you need more and more specialization because Moore’s law has died, Dennard Scaling has died, IE, Chips are not getting better just for free from manufacturing. You have to make real architectural innovations.

(01:34:59)
Google is not just running on Intel CPUs for web serving. They have a YouTube chip, they have TPUs, they have Pixel chips, they have a wide diversity of chips that generate all the economic value of Google. It’s running all the services and stuff. And this is just Google. And you could go across any company in the industry, and it’s like this. Cars contain 5,000 chips, 200 different varieties of them. All these random things. A Tesla door handle has two chips. It’s ridiculous. And it’s a cool door handle. You don’t think about it, but it has two really chip, penny chips in there. Anyways, so as you have more diversity of chips, as you have more specialization required and the cost of fabs continues to grow, you need someone who is laser focused on building the best process technology and making it as flexible as possible.
Nathan Lambert
(01:35:45)
I think you could say it simply, which is the cost per fab goes up, and if you are a small player that makes a few types of chips, you’re not going to have the demand to pay back the cost of the fab. Whereas NVIDIA can have many different customers and aggregate all this demand into one place, and then they’re the only person that makes enough money building chips to build the next fab. So this is kind of why the companies slowly get killed because they have, 10 years ago, a chip that is profitable and is good enough, but the cost to build the next one goes up. They may try to do this, fail because they don’t have the money to make it work, and then they don’t have any chips, or they build it and it’s too expensive and they just sort of have not profitable chips.
Dylan Patel
(01:36:27)
There’s more failure points. You could have one little process related to some sort of chemical etch or some sort of plasma etch or some little process that screws up, you didn’t engineer it right, and now the whole company falls apart, you can’t make chips. And so super, super powerful companies like Intel, they had the weathering storm to like, hey, they still exist today, even though they really screwed up their manufacturing six, seven years ago. But in the case of like AMD, they almost went bankrupt, they had to sell their fabs to Mubadala, UAE, and that became a separate company called Global Foundries, which is a foundry firm. And then AMD was able to then focus on the return back up, was like, “Hey, let’s focus on making chiplets and a bunch of different chips for different markets and focusing on specific workloads rather than all of these different things.”

(01:37:14)
And so you get more diversity of chips, you have more companies than ever designing chips, but you have fewer companies than ever manufacturing them. And this is where TSMC comes in, is they’ve just been the best. They are so good at it. They’re customer focused, they make it easy for you to fabricate your chips. They take all of that complexity and kind of try and abstract a lot of it away from you. They make good money. They don’t make insane money, but they make good money and they’re able to aggregate all this demand and continue to build the next fab, the next fab, the next fab.
Lex Fridman
(01:37:44)
So why is Taiwan so special for TSMC? Why is it happening there? Can it be replicated inside the United States?
Dylan Patel
(01:37:51)
Yeah, so there’s aspects of it that I would say yes, and aspects that I’d say no. TSMC is way ahead because former executive Morris Chang of Texas Instruments wasn’t promoted to CEO. And he was like, “Screw this. I’m going to go make my own chip company.” And he went to Taiwan and made TSMC. And there’s a whole lot more story there. Texas Instruments, could have have been TSMC, but Texas Semiconductor Manufacturing instead of Texas Instruments. So there is that whole story there. But the-
Nathan Lambert
(01:38:22)
Sitting here in Texas.
Lex Fridman
(01:38:23)
And that sounds like a human story. He didn’t get promoted.
Dylan Patel
(01:38:26)
Just the brilliance of Morris Chang, which I wouldn’t underplay, but there’s also a different level of how this works. So in Taiwan, the top percent of graduates of students that go to the best school, which is NTU, the top percent of those all go work to TSMC. And guess what their pay is? Their starting pay is like $80,000, $70,000, which is like that’s starting pay for a good graduate in the US, not the top. The graduates are making hundreds of thousands of dollars at the Googles and the Amazons, and now I guess the OpenAIs of the world. So there is a large dichotomy of what is the top 1% of the society doing and where are they headed because of economic reasons? Intel never paid that crazy good. And it didn’t make sense to them. That’s one aspect. Where’s the best going?

(01:39:16)
Second is the work ethic. We like to work. You work a lot, we work a lot, but at the end of the day, what does the time and amount of work that you’re doing and what does a fab require? Fabs are not work from home jobs. They are you go into the fab and grueling work. There’s hey, if there is any amount of vibration, an earthquake happens, vibrates the machines, they’re either broken, you’ve scrapped some of your production. And then in many cases, they’re not calibrated properly. So when there’s an earthquake, recently, there’s been a earthquake, TSMC doesn’t call their employees, they just go to the fab and they just show up. The parking lot gets slammed, and people just go into the fab and fix it. It’s like ants. It’s like a hive of ants doesn’t get told by the queen what to do. The ants just know.
Nathan Lambert
(01:40:08)
It’s like one person just specializes on these one task, and it’s like you’re going to take this one tool and you’re the best person in the world, and this is what you’re going to do for your whole life is this one task in the fab.
Dylan Patel
(01:40:17)
Which is some special chemistry plus nanomanufacturing on one line of tools that continues to get iterated and yeah, it’s like a specific plasma etch for removing silicon dioxide. That’s all you focus on your whole career, and it’s such a specialized thing. And so it’s not like the tasks are transferable. AI today is awesome because people can pick it up like that. Semiconductor manufacturing is very antiquated and difficult. None of the materials are online for people to read easily and learn. The papers are very dense, and it takes a lot of experience to learn. And so it makes the barrier to entry much higher too. So when you talk about, hey, you have all these people that are super specialized, they will work 80 hours a week in a factory, in a fab, and if anything goes wrong, they’ll go show up in the middle of the night because some earthquake, their wife’s like, “There was an earthquake.” He’s like, “Great, I’m going to go to the fab.”
Nathan Lambert
(01:41:08)
Time to cry.
Dylan Patel
(01:41:09)
Would you, as an American, do that? It’s like these sorts of things are, I guess are the exemplifying why TSMC is so amazing. Now, can you replicate it in the US? Let’s not ignore Intel was the leader in manufacturing for over 20 years. They brought every technology to market first besides the EUV. Strained silicon, high-K metal gates, FinFET, the list goes on and on and on of technologies that Intel brought to market first made the most money from and manufactured at scale first, best, highest profit margins. We shouldn’t ignore that Intel can’t do this. It’s that the culture has broken.

(01:41:48)
You’ve invested in the wrong things. They said no to the iPhone. They had all these different things regarding mismanagement of the fabs and mismanagement of designs, this lockup. And at the same time, all these brilliant people, these 50,000 PhDs or masters that have been working on specific chemical or physical processes or nanomanufacturing processes for decades, in Oregon, they’re still there, they’re still producing amazing work. It’s just getting it to the last mile of production at high yield where you can manufacture dozens and hundreds of different kinds of chips, and good customer experience has broken.

(01:42:24)
It’s that customer experience. Part of it is people will say, Intel was too pompous in the 2000s, 2010s. They just thought they were better than everyone. The tool guys were like, “Oh, I don’t think that this is mature enough.” And they’re like, “Ah, you just don’t know. We know.” This sort of stuff would happen. And so can the US bring leading-edge semiconductor manufacturing to the US? [inaudible 01:42:44] yes. And we are. It’s happening.
Nathan Lambert
(01:42:47)
Arizona is getting better and better as time goes on.
Dylan Patel
(01:42:50)
TSMC has built roughly 20% of their capacity for five nanometer in the US. Now, this is nowhere near enough. 20% of capacity in the US is like nothing. And furthermore, this is still dependent on Taiwan existing. There’s sort of important way to separate it out. There’s R&D and there’s high volume manufacturing. Effectively, there are three places in the world that are doing leading-edge R&D. There’s Hsinchu, Taiwan, there’s Hillsboro, Oregon, and there is Pyongyang, South Korea.

(01:43:24)
These three places are doing the leading-edge R&D for the rest of the world’s leading-edge semiconductors. Now, manufacturing can be distributed more globally. And this is sort of where this dichotomy exists of who’s actually modifying the process, who’s actually developing the next generation one, who’s improving them is Hsinchu, is Hillsboro, is Pyongyang. It is not the rest of these fabs like Arizona. Arizona is a paperweight. If Hsinchu disappeared off the face of the planet, within a year, couple years, Arizona would stop producing too. It’s actually pretty critical. One of the things I like to say is if I had a few missiles, I know exactly where I could cause the most economic damage. It’s not targeting the White House.
Lex Fridman
(01:44:09)
It’s the R&D centers.
Dylan Patel
(01:44:10)
It’s the R&D centers for TSMC, Intel, Samsung. And then some of the memory guys, Micron and Hynix.
Lex Fridman
(01:44:15)
Because they define the future evolution of these semiconductors, and everything’s moving so rapidly that it really is fundamentally about R&D. And it is all about TSMC. Huh.
Dylan Patel
(01:44:27)
And so TSMC, you cannot purchase a vehicle without TSMC chips. You cannot purchase a fridge without TSMC chips. I think one of the few things you can purchase ironically, is a Texas Instruments graphing calculator because they actually manufacture in Texas. But outside of that, a laptop, a phone.
Lex Fridman
(01:44:47)
It’s depressing.
Dylan Patel
(01:44:48)
Servers, GPUs, none of this stuff can exist. And this is without TSMC. And in many cases, it’s not even the leading-edge sexy five nanometer chip, three nanometer chip, two nanometer chip. Oftentimes, it’s just some stupid power IC that’s converting from some voltage to another, and it’s…
Dylan Patel
(01:45:00)
… I see that’s converting from some voltage to another, and it’s made at TSMC. It’s like-
Nathan Lambert
(01:45:05)
This is what China is investing in as well. It’s like, they can build out this long-tail fab where the techniques are much more known, you don’t have to figure out these problems with EUV. They’re investing in this and then they have large supply for things like the car door handles and the random stuff. And that trickles down into this whole economic discussion as well, which is they have far more than we do. And having supply for things like this is crucial to normal life.
Lex Fridman
(01:45:29)
So they’re starting to invest in high-volume manufacturer, but they’re not doing R&D as much?
Nathan Lambert
(01:45:29)
They are.
Dylan Patel
(01:45:36)
They do R&D on their own, they’re just way behind. I would say, in 2015 China had a five-year plan where they defined by 2025 and 2020 certain goals, including 80% domestic production of semiconductors. They’re not going to hit that, to be clear. But they are in certain areas really, really close. BYD is probably going to be the first company in the world to not have to use TSMC for making … because they have their own fabs for making chips.

(01:46:04)
Now they still have to buy some chips from foreign, for example, around like self-driving ADAS capabilities because those are really high-end, but at least … A internal combustion engine has 40 chips in an EV, just for controlling flow rates and all these things, and EVs are even more complicated. So all these different power ICs and battery management controllers and all these things, they’re insourcing.

(01:46:26)
And this is something that China has been doing since 2015. Now, as far as the trailing edge, they’re getting so much capacity there. As far as the leading edge, i.e. this five nanometer and so on and so forth, where GPUs, they are still behind. The US restrictions are trying to stop them in the latter, but all that’s happened is yes, they’ve slowed down their five nanometer, three nanometer, et cetera, but they’ve accelerated their, hey, 45 nanometer, 90 nanometer power IC or analog IC or random chip in my keyboard, that kind of stuff.

(01:46:59)
So there is an angle of, the US’ actions, from the angle of the expert controls, have been so inflammatory at slowing down China’s progress on the leading edge that they’ve turned around and have accelerated their progress elsewhere because they know that this is so important. If the US is going to lock them out here, “what if they lock us out here as well in the trailing edge?”

(01:47:20)
And so going back, can the US build it here? Yes, but it’s going to take a ton of money. I truly think to revolutionize and completely in-source semiconductors would take a decade and a trillion dollars.
Lex Fridman
(01:47:33)
Is some of it also culture, like you said, extreme competence, extreme work ethic in Taiwan?
Nathan Lambert
(01:47:39)
I think if you have the demand and the money is on the line, the American companies figure it out. It’s going to take handholding with the government, but I think that the culture helps TSMC break through and it’s easier for them. You [inaudible 01:47:50].
Dylan Patel
(01:47:50)
TSMC has some like 90,000 employees. It’s not actually that insane amount. The Arizona fab has 3,000 from Taiwan. And these people, their wives were like, “Yeah, we’re not going to have kids unless you sign up for the Arizona Fab. We go to Arizona and we have our kids there.” There’s also a Japan fab where the same thing happened. And so these wives drove these dudes to go to Japan or America to have the kids there.

(01:48:13)
And it’s an element of culture, yeah, sure. Taiwan works that hard. But also, like the US has done it in the past, they could do it now. We can just import, I say import, the best people in the world if we want to.
Lex Fridman
(01:48:25)
That’s where the immigration conversation is a tricky one and there’s been a lot of debate over that. But yeah, it seems absurdly controversial to import the best people in the world. I don’t understand why it’s controversial. That’s one of the ways of winning.
Nathan Lambert
(01:48:38)
I’m sure we agree with you.
Dylan Patel
(01:48:39)
And even if you can’t import those people, I still think you could do a lot to manufacture most of it in the US, if the money’s there.
Nathan Lambert
(01:48:45)
It’s just way more expensive. It’s not profitable for a long time.
Dylan Patel
(01:48:50)
And that’s the context of the Chips Act is only $50 billion, relative to some of the renewable initiatives that were passed in the Inflation Reduction Act and the Infrastructure Act, which total in the hundreds of billions of dollars. And so the amount of money that the US is spending on the semiconductor industry is nothing, whereas all these other countries have structural advantages in terms of work ethic and amount of work and things like that, but also a number of STEM graduates, the percentile of their best going to that.

(01:49:20)
But they also have differences in terms of, hey, there’s just tax benefits in the law and have been in the law for 20 years. And then some countries have massive subsidies. China has something like $200 billion of semiconductor subsidies a year. We’re talking about $50 billion in the US over like six. So the girth or difference in the subsidy amounts is also huge.

(01:49:44)
And so I think Trump has been talking about tariffing Taiwan recently. That’s one of these things that’s like, “Oh, okay, well, maybe he doesn’t want to subsidize the US semiconductor industry.” Obviously tariffing Taiwan is going to cost a lot of things to get much more expensive, but does it change the equation for TSMC building more fabs in the US? That’s what he’s positing.
Lex Fridman
(01:50:07)
So we laid out the importance … By the way, it’s incredible how much you know about so much.
Nathan Lambert
(01:50:13)
We told you Dylan knows all this stuff.
Lex Fridman
(01:50:15)
Yeah. Okay. You laid out why TSMC is really important. If we look out into the future 10, 20 years out, US-China relationship, it seems like it can go to a dark place of Cold War, escalated Cold War, even hot war, or to a good place of anything from frenemies, to cooperation, to working together.

(01:50:44)
So in this game theory, complicated game, what are the different trajectories? What should US be doing? What do you see as the different possible trajectories of US-China relations as both leaders start to feel the AGI more and more and see the importance of chips and the importance of AI.
Nathan Lambert
(01:51:04)
I mean, ultimately the export controls are pointing towards a separate future economy. I think the US has made it clear to Chinese leaders that we intend to control this technology at whatever cost to global economic integration. And it’s hard to unwind that. The card has been played.
Dylan Patel
(01:51:27)
To the same extent they’ve also limited US companies from entering China. So it’s been a long time coming. At some point there was a convergence, but over at least the last decade it’s been branching further and further out. US companies can’t enter China. Chinese companies can’t enter the US. The US is saying, “Hey, China, you can’t get access to our technologies in certain areas.” And China’s rebuttaling with the same thing around … they’ve done some sort of specific materials in gallium and things like that that they’ve tried to limit the US on. There’s a US drone company that’s not allowed to buy batteries and they have military customers. And this drone company just tells the military customers, “Hey, just get it from Amazon because I can’t actually physically get them.”

(01:52:10)
There’s all these things that are happening that point to further and further divergence. I have zero idea, and I would love if we could all hold hands and sing Kumbaya, but I have zero idea how that could possibly happen.
Lex Fridman
(01:52:21)
Is the divergence good or bad for avoiding war? Is it possible that the divergence in terms of manufacturer chips of training AI systems is actually good for avoiding military conflict?
Dylan Patel
(01:52:35)
It’s an objective fact that the world has been the most peaceful it’s ever been when there are global hegemons, or regional hegemons in historical context. The Mediterranean was the most peaceful ever when the Romans were there. China had very peaceful and warring times, and the peaceful times were when dynasties had a lock hold over, not just themselves, but all their tributaries around them. And likewise, the most peaceful time in human history has been when the US was the global hegemon, the last decades. Now we’ve seen things start to slide with Russia, Ukraine, with what’s going on in the Middle East, and Taiwan risk, all these different things are starting to bubble up. Still objectively extremely peaceful.

(01:53:14)
Now what happens when it’s not one global hegemon but it’s two, obviously … And China will be competitive or even overtake the US, it’s possible. And so this change in global hegemony, I don’t think it ever happens super peacefully. When empires fall, which is a possible trajectory for America, they don’t fall gracefully. They don’t just slide out of irrelevance. Usually there’s a lot of shaking. And so what the US is trying to do is maintain its top position, and what China is trying to do is become the top position. And obviously there’s butting of heads here, in the most simple terms.
Lex Fridman
(01:53:53)
And that could take shape in all kinds of ways, including proxy wars. And now-
Nathan Lambert
(01:53:58)
Yeah, it seems like it’s already happening. As much as I want there to be centuries of prolonged peace, it looks like further instability internationally is ahead.
Dylan Patel
(01:54:08)
And the US’ current task is, “Hey, if we control AI, if we’re the leader in AI and AI significantly accelerates progress, then we can maintain the global hegemony position.” And therefore-
Nathan Lambert
(01:54:21)
I hope that works.
Dylan Patel
(01:54:23)
And as an American, like, okay, I guess that’s going to lead to peace for us. Now obviously other people around the world get affected negatively. Obviously the Chinese people are not going to be in as advantageous of a position if that happens, but this is the reality of what’s being done and the actions that are being carried out.

Best GPUs for AI

Lex Fridman
(01:54:44)
Can we go back to the specific detail of the different hardware? There’s this nice graphic in the export controls of which GPUs are allowed to be exported and which are not. Can you explain the difference? From a technical perspective, are the H20s promising?
Dylan Patel
(01:55:08)
Yeah. And I think we need to dive really deep into the reasoning aspect and what’s going on there. The US has gone through multiple iterations of the export controls. This H800 was at one point allowed back in ’23, but then it got canceled and by then DeepSeek had already built their cluster of, they claim, 2K. I think they actually have many more, something like 10K of those. And now this H20 is the legally allowed chip. Nvidia shipped a million of these last year to China. For context, it was four or five million GPUs. So the percentage of GPUs that were this China-specific H20 is quite high, roughly 20%, 25%, 20% or so.

(01:55:48)
And so this H20 has been neutered in one way, but it’s actually upgraded in other ways. And you could think of chips along three axes for AI, ignoring software stack and exact architecture, just raw specifications. There’s floating point operations, FLOPS. There is memory bandwidth, i.e. in-memory capacity, IO memory. And then there is interconnect, chip-to-chip interconnections. All three of these are incredibly important for making AI systems. Because AI systems involve a lot of compute, they involve a lot of moving memory around, whether it be to memory or too other chips.

(01:56:28)
And so these three vectors, the US initially had two of these vectors controlled and one of them not controlled, which was FLOPS and interconnect bandwidth were initially controlled. And then they said, “No, no, no, no. We’re going to remove the interconnect bandwidth and just make it a very simple, only FLOPS.” But now Nvidia can now make a chip that has … okay, it’s cut down on FLOPS, so one-third that of the H100 on spec sheet paper performance for FLOPs. In real world it’s closer to half or maybe even 60% of it. But then on the other two vectors, it’s just as good for interconnect bandwidth. And then for memory bandwidth and memory capacity, the H20 has more memory bandwidth and more memory capacity than the H100.

(01:57:10)
Now recently we, at our research, we cut Nvidia’s production for H20 for this year down drastically. They were going to make another two million of those this year, but they just canceled all the orders a couple of weeks ago. In our view that’s because we think that they think they’re going to get restricted, because why would they cancel all these orders for H20? Because they shipped a million of them last year, they had orders in for a couple million this year, and just gone right. For H20, B20, a successor to H20, and now they’re all gone.

(01:57:39)
Now why would they do this? I think it’s very clear, the H20 is actually better for certain tasks. And that certain task is reasoning. Reasoning is incredibly different than … When you look at the different regimes of models. Pre-training is all about FLOPS, it’s all about FLOPS. There’s things you do, like Mixture of Experts that we talked about, to trade off interconnect or to trade off other aspects and lower the FLOPS and rely more on interconnect and memory.

(01:58:10)
But at the end of the day, FLOPS is everything. We talk about models in terms of how many FLOPS they are. So we talk about, oh, GPT-4 is 2e25. Two to the 25th, 25 zeros FLOP, floating point operations for training. And we’re talking about the restrictions for the 2e24, or 25, whatever. The US has an executive order that Trump recently unsigned, which was, hey, 1e26, once you hit that number of floating point operations, you must notify the government and you must share your results with us. There’s a level of model where the US government must be told, and that’s 1e26.

(01:58:50)
And so as we move forward, this is an incredibly important … FLOP is the vector that the government has cared about historically, but the other two vectors are arguably just as important. And especially when we come to this new paradigm, which the world is only just learning about over the last six months: reasoning.
Lex Fridman
(01:59:07)
And do we understand firmly which of the three dimensions is best for reasoning? So interconnect, the FLOPS don’t matter as much, is it memory?
Nathan Lambert
(01:59:17)
Memory. Yeah. We’re going to get into technical stuff real fast.
Dylan Patel
(01:59:21)
I would say there’s two articles in this one that I could show maybe graphics that might be interesting for you to pull up.
Lex Fridman
(01:59:27)
For the listeners, we’re looking at the section of 01 inference architectures tokenomics.
Dylan Patel
(01:59:33)
You want to explain KV cache before we talk about this? I think it’s better to-
Nathan Lambert
(01:59:36)
Okay. Yeah, we need to go through a lot of specific technical things, transformers, to make this easy for people.
Dylan Patel
(01:59:42)
Because it’s incredibly important because this changes how models work. But I think resetting, why is memory so important? It’s because so far we’ve talked about parameter counts and Mixture of Experts, you can change how many active parameters versus total parameters to embed more data but have less FLOPS. B. Ut more important, another aspect of what’s part of this humongous revolution in the last handful of years is the transformer and the attention mechanism. Attention mechanism is that the model understands the relationships between all the words in its context, and that is separate from the parameters themselves. And that is something that you must calculate. How each token, each word in the context length, is relatively connected to each other. And I think, Nathan, you can explain KV cache better.
Lex Fridman
(02:00:31)
KV cache is one of the optimization [inaudible 02:00:33]?
Nathan Lambert
(02:00:33)
So the attention operator has three core things, it’s queries, keys, and values. QKV is the thing that goes into this. You’ll look at the equation. You see that these matrices are multiplied together. These words, query, key and value, come from information retrieval backgrounds where the query is the thing you’re trying to get the values for and you access the keys and the values is reweighting. My background’s not information retrieval and things like this, it’s just fun to have backlinks.

(02:01:01)
And what effectively happens is that when you’re doing these matrix multiplications, you’re having matrices that are of the size of the context length, so the number of tokens that you put into the model. And the KV cache is effectively some form of compressed representation of all the previous tokens in the model. So when you’re doing this, we talk about autoregressive models, you predict one token at a time. You start with whatever your prompt was, you ask a question, like who was the president in 1825. The model then is going to generate its first token.

(02:01:32)
For each of these tokens you’re doing the same attention operator where you’re multiplying these query, key-value matrices. But the math is very nice so that when you’re doing this repeatedly, this KV cache, this key-value operation, you can keep appending the new values to it, so you keep track of what your previous values you were inferring over in this autoregressive chain, you keep it in-memory the whole time. And this is a really crucial thing to manage when serving inference at scale. There are far bigger experts in this and there are so many levels of detail that you can go into.

(02:02:09)
Essentially one of the key, quote unquote, “drawbacks” of the attention operator and the transformer is that there is a form of quadratic memory cost in proportion to the context length. So as you put in longer questions, the memory used in order to make that computation is going up in the form of a quadratic. You’ll hear about a lot of other language model architectures that are sub quadratic or linear attention forms, which is like State Space Models. We don’t need to go down all these now. And then there’s innovations on attention to make this memory usage and the ability to attend over long contexts much more accurate and high performance.
Lex Fridman
(02:02:50)
And those innovations are going to help you with … I mean, your highly memory constrained in this?
Nathan Lambert
(02:02:54)
They help with memory constraint and performance. Gemini is the model that has the longest context length that people are using. Gemini is known for one million and now two million context length. You put a whole book into Gemini and sometimes it’ll draw facts out of it. It’s not perfect, they’re getting better.

(02:03:12)
So there’s two things. It’s, one, to be able to serve this on the memory level. Google has magic with their TPU stack where they can serve really long contexts. And then there’s also many decisions along the way to actually make long context performance work that supplies the data. There’s subtle changes to these computations in attention and it changes the architecture. But serving long context is extremely memory constrained, especially when you’re making a lot of predictions. I actually don’t know why input and output tokens are more expensive, but I think essentially output tokens, you have to do more computation because you have to sample from the model.
Dylan Patel
(02:03:46)
I can explain that. Today, if you use a model, like you look at an API, OpenAI charges a certain price per million tokens. And that price for input and output tokens is different. And the reason is is that when you’re inputting a query into the model, let’s say you have a book, that book, you must now calculate the entire KV cache for this, key-value cache.

(02:04:10)
And so when you do that, that is a parallel operation. All of the tokens can be processed at one time and therefore you can dramatically reduce how much you’re spending. The FLOP requirements for generating a token and an input token are identical. If I input one token or if I generate one token, it’s completely identical. I have to go through the model. But the difference is that I can do that input, i.e. the prefill, i.e. the prompt, simultaneously in a batch nature and therefore it is all FLOP.
Lex Fridman
(02:04:38)
I think the pricing model mostly they use for input tokens is about one fourth of price of the output tokens.
Dylan Patel
(02:04:44)
Correct. But then output tokens, the reason why it’s so expensive is because I can’t do it in parallel. It’s autoregressive. Every time I generate a token, I must not only read the whole entire model into memory and activate it, calculate it to generate the next token, I also have to read the entire KV cache. And I generate a token and then I append that one token I generated and it’s KV cache and then I do it again.

(02:05:07)
And so therefore, this is a non-parallel operation. And this is one where you have to, in the case of prefill or prompt, you pull the whole model in and you calculate 20,000 tokens at once, 20,000-
Nathan Lambert
(02:05:21)
These are features that APIs are shipping, which is like prompt caching, prefilling, because you can drive prices down and you can make APIs much faster. If you run a business and you’re going to keep passing the same initial content to Claude’s API, you can load that in to the Anthropic API and always keep it there.

(02:05:38)
But it’s very different than we’re leading to these reasoning models, which we showed this example earlier and read some of this mumbling stuff. And what happens is that the output context length is so much higher. And I mean, I learned a lot about this from Dylan’s work, which is essentially as the output work length gets higher, you’re writing this quadratic in terms of memory used. And then the GPUs that we have, effectively you’re going to run out of memory and they’re all trying to serve multiple requests at once. So they’re doing this batch processing where not all of the prompts are exactly the same, really complex handling.

(02:06:12)
And then as context links gets longer, there’s this, I think you call it critical batch size, where your ability to serve more users, so how much you can parallelize your inference plummets because of this long context. So your memory usage is going way up with these reasoning models and you still have a lot of users, so effectively the cost to serve multiplies by a ton.
Lex Fridman
(02:06:35)
And we’re looking at a plot when the x-axis is sequence length.
Dylan Patel
(02:06:39)
I.e., how many tokens are being generated/prompt. So if I put in a book, that’s a million tokens. But if I put in “the sky is blue,” then that’s like six tokens or whatever.
Lex Fridman
(02:06:49)
And we should say that what we’re calling reasoning and chain of thought is extending this sequence length.
Nathan Lambert
(02:06:55)
It’s mostly output.
Dylan Patel
(02:06:56)
Right. So before three months ago, whenever o1 launched, all of the use cases for long context length were, “Let me put a ton of documents in and then get an answer out.” And it’s a single, prefill compute a lot in parallel and then output a little bit.

(02:07:11)
Now with reasoning and agents, this is a very different idea. Now instead I might only have like, hey, do this task, or I might have all these documents, but at the end of the day, the model is not just producing a little bit, it’s producing tons of information, this chain of thought-
Nathan Lambert
(02:07:25)
Tens of thousands of tokens.
Dylan Patel
(02:07:25)
… just continues to go and go and go and go. And so the sequence length is effectively that if it’s generated 10,000 tokens, it’s 10,000 sequence length, and plus whatever you inputted in the prompt.

(02:07:37)
And so what this chart is showing, and it’s a logarithmic chart, is as you grow from 1K to 4K or 4K to 16K, the memory requirements grow so fast for your KV cache that you end up not being able to run a certain number of … Your sequence length is capped or the number of users you could serve-
Nathan Lambert
(02:07:57)
Let’s say the model. So this is showing for a 405B model in batch size 64.
Lex Fridman
(02:08:02)
Llama 3.1.405B. Yeah.
Nathan Lambert
(02:08:04)
Yeah. And batch size is crucial too. Essentially you want to have higher batch size to parallel your throughput.
Dylan Patel
(02:08:11)
64 different users at once.
Nathan Lambert
(02:08:13)
Yeah.
Dylan Patel
(02:08:13)
And therefore your serving costs are lower, because the server costs the same. This is eight H100s, roughly $2 an hour per GPU. That’s $16 an hour. That is somewhat of a fixed cost. You can do things to make it lower of course, but it’s like $16 an hour. Now how many users can you serve, how many tokens can you generate, and then you divide the two and that’s your cost.

(02:08:32)
And so with reasoning models, this is where a lot of the complexity comes about and why memory is so important. Because if you have limited amounts of memory, then you can’t serve so many users. If you have limited amounts of memory, your serving speeds get lower. And so your costs get a lot, lot worse because all of a sudden if I was used to, hey, on this $16 an hour server I’m serving Llama 405B, or if I’m serving DeepSeek-V3 and it’s all chat style applications, i.e. we’re just chit-chatting, the sequence length are a thousand, a few thousand. When you use a language model, it’s a few thousand context length most of times. Sometimes you’re dropping a big document, but then you process it, you get your answer, you throw it away, you move on to the next thing.

(02:09:12)
Whereas with reasoning, I’m now generating tens of thousands of tokens in sequence. And so this memory, this KV cache, has to stay resonant and you have to keep loading it, you have to keep it in-memory constantly. And now this butts out other users. If there’s now a reasoning task and the model’s capable of reasoning, then all of a sudden that memory pressure means that I can’t serve as many users simultaneously.

Why DeepSeek is so cheap

Nathan Lambert
(02:09:36)
Let’s go into DeepSeek again. So we’re in the post DeepSeek-R1 time I think, and there’s two sides to this market, watching how hard it is to serve it. On one side we’re going to talk about DeepSeek themselves. They now have a chat app that got to number one on the App Store. Disclaimer number one on the App Store is measured by velocity, so it’s not necessarily saying that more people have the DeepSeek app than the ChatGPT app. But it is still remarkable. Claude has never hit the number one in the App Store, even though everyone in San Francisco is like, “Oh my god, you got to use Claude. Don’t use ChatGPT.”

(02:10:06)
So DeepSeek hit this. They also launched an API product recently where you can ping their API and get these super long responses for R1 out. At the same time as these are out, we’ll get to what’s happened to them. Because the model weights for DeepSeek-R1 are openly available and the license is very friendly, the MIT license commercially available, all of these midsize companies and big companies are trying to be first to serve R1 to their users.

(02:10:33)
We are trying to evaluate R1 because we have really similar research going on. We released the model and we’re trying to compare to it. And out of all the companies that are, quote unquote, “serving” R1 and they’re doing it at prices that are way higher than the DeepSeek API, most of them barely work and the throughput is really low.
Dylan Patel
(02:10:51)
To give context, one of the parts of freaking us out was like China reached capabilities. The other aspect is they did it so cheap. And the so cheap, we talked about on the training side why it was so cheap slash-
Lex Fridman
(02:11:03)
Yeah, let’s talk about why it’s so cheap on the inference. It works well and it’s cheap. Why is R1 so damn cheap?
Dylan Patel
(02:11:08)
I think there’s a couple factors here. One is that they do have model architecture innovations. This MLA, this new attention that they’ve done, is different than the attention from attention is all you need, the transformer attention.

(02:11:23)
Now, others have already innovated. There’s a lot of work like MQA, GQA, local, global, all these different innovations that try to bend the curve. It’s still quadratic, but the constant is now smaller.
Nathan Lambert
(02:11:33)
Related to our previous discussion, this multi-head latent attention can save about 80 to 90% in memory from the attention mechanism, which helps especially in long contexts.
Dylan Patel
(02:11:44)
It’s 80 to 90% versus the original. But then versus what people are actually doing, it’s still an innovation.
Nathan Lambert
(02:11:49)
This 80 to 90% doesn’t say that the whole model is 80 to 90% cheaper. Just this one part of it.
Dylan Patel
(02:11:54)
Well, and not just that, other people have implemented techniques like global-global and sliding window and GQMQ. But anyways, DeepSeek has … their attention mechanism is a true architectural innovation. They did tons of experimentation. And this dramatically reduces the memory pressure. It’s still there, it’s still attention, it’s still quadratic, it’s just dramatically reduced it relative to prior forms.
Lex Fridman
(02:12:16)
Right. That’s the memory pressure. I should say, in case people don’t know, R1 is 27 times cheaper than o1.
Nathan Lambert
(02:12:25)
We think that OpenAI had a large margin built in.
Lex Fridman
(02:12:28)
Okay, so that’s one-
Nathan Lambert
(02:12:29)
There’s multiple factors. We should break down the factors, I think.
Lex Fridman
(02:12:31)
It’s two bucks per million token output for R1 and $60 per million token output for o1.
Dylan Patel
(02:12:40)
Yeah, let’s look at this. I think this is very important. OpenAI is that drastic gap between DeepSeek and pricing. But DeepSeek is offering the same model because they open weight to everyone else for a very similar, much lower price than what others are able to serve it for. So there’s two factors here. Their model is cheaper. It is 27 times cheaper. I don’t remember the number exactly off the top of my head.
Lex Fridman
(02:13:07)
We’re looking at a graphic that’s showing different places serving V3, DeepSeek-V3, which is similar to DeepSeek-R1. And there’s a vast difference in-
Dylan Patel
(02:13:21)
In serving cost.
Lex Fridman
(02:13:21)
… in serving cost. And what explains that difference?
Dylan Patel
(02:13:23)
And so part of it is OpenAI has a fantastic margin. When they’re doing inference, their gross margins are north of 75%. So that’s a four to five X factor right there of the cost difference, is that OpenAI is just making crazy amounts of money because they’re the only one with the capability.
Lex Fridman
(02:13:40)
Do they need that money? Are they using it for R&D?
Dylan Patel
(02:13:42)
They’re losing money, obviously, as a company because they spend so much on training. So the inference itself is a very high margin, but it doesn’t recoup the cost of everything else they’re doing. So yes, they need that money because the revenue and margins pay for continuing to build the next thing, as long as I’m raising more money.
Lex Fridman
(02:13:57)
So the suggestion is that DeepSeek is really bleeding out money.
Dylan Patel
(02:14:01)
Well, so here’s one thing, we’ll get to this in a second, but DeepSeek doesn’t have any capacity to actually serve the model. They stopped signups. The ability to use it is non-existent now for most people because so many people are trying to use it. They just don’t have the GPUs to serve it. OpenAI has hundreds of thousands of GPUs between them and Microsoft to serve their models. DeepSeek has a factor of much lower, even if you believe our research, which is 50,000 GPUs, and a portion of those are for research, a portion of those are for the hedge fund, they still have nowhere close to the GPU volumes and capacity to serve the model at scale.

(02:14:36)
So it is cheaper. A part of that, is OpenAI making a ton of money? Is DeepSeek making on their API? Unknown, I don’t actually think so. And part of that is this chart. Look at all the other providers. Together AI, Fireworks.ai are very high-end companies. Ex-Meta, Together AI is [inaudible 02:14:53] and the inventor of FlashAttention, which is a huge efficiency technique. There a very efficient, good companies. And I do know those companies make money, not tons of money on inference, but they make money. And so they’re serving at a five to 7X difference in cost.

(02:15:09)
And so now when you equate, okay, OpenAI is making tons of money, that’s like a 5x difference, and the companies that are trying to make money for this model is like a 5x difference, there is still a gap. There’s still a gap and that is just DeepSeek being really freaking good. The model architecture, MLA, the way they did the MoE, all these things, there is legitimate just efficiency differences.
Nathan Lambert
(02:15:28)
It’s like all their low-level libraries that we talked about in training, some of them probably translate to inference and those weren’t released.
Lex Fridman
(02:15:33)
So we may go a bit into conspiracy land, but is it possible the Chinese government is subsidizing DeepSeek?
Dylan Patel
(02:15:40)
I actually don’t think they are. I think when you look at the Chinese labs, Huawei has a lab, Moonshot AI, there’s a couple other labs out there that are really close with the government, and then there’s labs like Alibaba and DeepSeek, which are not close with the government. And we talked about the CEO, this reverent figure, who’s quite different, who has these-
Nathan Lambert
(02:16:02)
Sounds awesome.
Dylan Patel
(02:16:03)
… very different viewpoints based on the Chinese interviews that are translated than what the CCP might necessarily want. Now, to be clear, does he have a loss leader because he can fund it through his hedge fund? Yeah, sure.
Lex Fridman
(02:16:14)
So the hedge fund might be subsidizing it, [inaudible 02:16:17]?
Dylan Patel
(02:16:16)
Yes. I mean, they absolutely did, because DeepSeek has not raised much money. They’re now trying to raise around in China, but they have not raised money historically. It’s all just been funded by the hedge fund. And he owns over half the company, like 50, 60% of the company is owned by him.
Nathan Lambert
(02:16:29)
Some of the interviews, there’s discussion on how doing this is a recruiting tool. You see this at the American companies too. It’s like having GPUs, recruiting tool. Being at the cutting edge of AI, recruiting tool.
Dylan Patel
(02:16:39)
Open sourcing.
Nathan Lambert
(02:16:40)
Open sourcing, recruiting tool.
Dylan Patel
(02:16:42)
Mete, they were so far behind and they got so much talent because they just open sourced stuff.
Lex Fridman
(02:16:46)
More conspiracy thoughts. Is it possible, since they’re a hedge fund, that they timed everything with this release and the pricing and they shorted Nvidia stock and stock of USA AI companies and released it with Stargate … just perfect timing to be able to make money.
Nathan Lambert
(02:17:08)
If they did, props. They’ve released it on an inauguration day. They know what is on the international calendar, but I mean, I don’t expect them to. If you listen to their motivations for AI, it’s like-
Lex Fridman
(02:17:18)
No, if you-
Dylan Patel
(02:17:19)
They released V3 on December 26th. Who releases the day after Christmas? No one looks. They had released the papers before this, the V3 paper and the R1 paper. So people have been looking at it and been like, “Wow. And then they just released the R1 model.

(02:17:33)
I think they’re just shipping as fast as they can, and who cares about Christmas, who cares about … Get it out before Chinese New Year, obviously, which just happened. I don’t think they actually were timing the market or trying to make the biggest splash possible, I think they’re just shipping.
Nathan Lambert
(02:17:46)
I think that’s one of their big advantages. We know that a lot of the American companies are very invested in safety, and that is the central culture of a place like Anthropic. And I think Anthropic sounds like a wonderful place to work, but if safety is your number one goal, it takes way longer to get artifacts out. That’s why Anthropic is not open sourcing things, that’s their claims.

(02:18:08)
But there’s reviews internally. Anthropic mentions things to international governments. There’s been news of how Anthropic has done pre-release testing with the UK AI Safety Institute. All of these things add inertia to the process of getting things out. And we’re on this trend line where the progress is very high. So if you reduce the time from when your model is done training, you run the vals, it’s good. You want to get it out as soon as possible to maximize the perceived quality of your outputs. DeepSeek does this so well.
Dylan Patel
(02:18:37)
Dario explicitly said Claude 3.5 Sonnet was trained like nine months or a year-
Nathan Lambert
(02:18:41)
Nine to 10 months ago [inaudible 02:18:42].
Dylan Patel
(02:18:42)
Nine to 10 months ago. And I think it took them another handful of months to release it. So it’s like, there is a significant gap here. And especially with reasoning models, the word in the San Francisco street is that Anthropic has a better model than o3 and they won’t release it. Why? Because chains-of-thought are scary, and they are legitimately scary. If you look at R1, it flips back and forth between Chinese and English, sometimes it’s gibberish, and then the right answer comes out. And for you and I, it’s like, “Great. Great.”
Nathan Lambert
(02:19:11)
This is why people are infatuated with … you’re like, “You’re telling me this is a high value thing and it works and it’s doing this?” It’s amazing.
Lex Fridman
(02:19:12)
Yeah, it’s incredible.
Dylan Patel
(02:19:18)
I mean, you talked about that chain-of-thought for that philosophical thing, which is not something they trained it to be philosophically good. It’s just an artifact of the chain-of-thought training it did. But that’s super important in that, can I inspect your mind and what you’re thinking right now? No. And so I don’t know if you’re lying to my face.

(02:19:37)
And chain-of-thought models are that way. This is a true, quote unquote, “risk” between a chat application where, hey, I asked the model to say bad words or whatever or how to make anthrax, and it tells me. That’s unsafe, sure, but that’s something I can get out relatively easily. What if I tell the AI to do a task and then it does the task all of a sudden randomly in a way that I don’t want it, and now that has much more … Task versus response is very different. So the bar for safety is much higher-
Dylan Patel
(02:20:00)
… task versus response is very different, so the bar for safety is much higher, at least this is Anthropics’ case, right? For DeepSeek, they’re like, “Ship,” right?
Lex Fridman
(02:20:08)
Yeah. So, the bar for safety is probably lowered a bit because of DeepSeek. There’s parallels here to the space race. The reason the Soviets probably put a man in space first is because their approach to safety, the bar for safety, was lowered
Dylan Patel
(02:20:26)
And they killed that dog, and all these things, so it’s like…
Lex Fridman
(02:20:29)
Less risk averse than the US Space Program. And there’s parallels here, but there’s probably going to be downward pressure on that safety bar for the US companies.
Nathan Lambert
(02:20:41)
This is something that Dario talks about. That’s the situation that Dario wants to avoid is, Dario talks too about the difference between race to the bottom and race to the top. And the race to the top is where there’s a very high standard on safety. There’s a very high standard on your model forms and certain crucial evaluations. And when certain companies are really good to it, they will converge. This is the idea. And ultimately, AI is not confined to one nationality or to one set of morals for what it should mean. And there’s a lot of arguments on should we stop open-sourcing models. And if the US stops, it’s pretty clear it’s way easier to see now at DeepSeek that a different international body will be the one that builds it.

(02:21:25)
We talk about the cost of training. DeepSeek has this shocking $5 million number. Think about how many entities in the world can afford a hundred times that to have the best open-source model that people use in the world. And it’s a scary reality, which is that these open models are probably going to keep coming for the time being, whether or not we want to stop them, and stopping them might make it even worse and harder to prepare. But it just means that the preparation and understanding what AI can do is just so much more important. That’s why I’m here at the end of the day. But it’s letting that sink into people, especially not in AI, is that this is coming. There are some structural things in a global interconnected world that you have to accept.
Lex Fridman
(02:22:09)
Yeah. You sent me something that Mark Zuckerberg mentioned on the earnings call. He said that, “I think in light of some of the recent news, the new competitor DeepSeek from China, I think it’s one of the things that we’re talking about is there’s going to be an open-source standard globally. And I think for our kind of national advantage, it’s important that it’s an American standard, so we take that seriously. We want to build the AI system that people around the world are using. And I think that, if anything, some of the recent news has only strengthened our conviction that this is the right thing to be focused on.” So yeah, open-sourcing.
Nathan Lambert
(02:22:43)
Mark Zuckerberg is not new to having American values and how he presents his company’s trajectory. I think their products have long since been banned in China, and I respect saying it directly.

Espionage

Dylan Patel
(02:22:55)
And there’s an interesting aspect of just because it’s open-weights or open-source doesn’t mean it can’t be subverted, right? There have been many open source software bugs that have been… For example, there was a Linux bug that was found after 10 years, which was clearly a back door because somebody was like, “Why is this taking half a second to load?”
Nathan Lambert
(02:23:14)
This is the recent one.
Dylan Patel
(02:23:15)
Right? There’s, “Why’s this taking half a second to load?” And it was like, “Oh crap, there’s a back door here. That’s why.” And this is very much possible with AI models. Today, the alignment of these models is very clear. I’m not going to say bad words. I’m not going to teach you how to make anthrax. I’m not going to talk about Tiananmen Square. I’m going to say Taiwan is just an eastern province. All these things are depending on who you are, what you align, and even like xAI is aligned a certain way. It’s not aligned in the woke sense, it’s not aligned in the pro-China sense, but there is certain things that are imbued within the model.

(02:23:57)
Now, when you release this publicly in an instruct model that’s open- weights, this can then proliferate, but as these systems get more and more capable, what you can embed deep down in the model is not as clear. And so that is one of the big fears is if an American model or a Chinese model is the top model, you are going to embed things that are unclear. And it can be unintentional too. British English is dead because American LLMs won and the internet is American, and therefore, color is spelled the way Americans spell, and this is-
Lex Fridman
(02:24:28)
A lot of strong words right now.
Dylan Patel
(02:24:31)
This is just the factual nature of the LLMs.
Nathan Lambert
(02:24:35)
[inaudible 02:24:35] English is the hottest programming language and that English is defined by a bunch of companies that primarily are in San Francisco.
Lex Fridman
(02:24:42)
The right way to spell optimization is with a Z, just in case. I think it’s an S in British English.
Nathan Lambert
(02:24:49)
It is.
Dylan Patel
(02:24:50)
Taking it as something silly. Something as silly as the spelling, which Brits and Americans will laugh about probably, right? I don’t think we care that much, but some people will. But this can boil down into very, very important topics like, hey, subverting people, chatbots, right? Character AI has shown that they can talk to kids or adults, and people will feel a certain way, and that’s unintentional alignment. But what happens when there’s intentional alignment deep down on the open-source standard, it’s a back door today for Linux that we discover or some encryption system. Chinese uses different encryption than NIST defines, the US NIST, because there’s clearly… At least they think there’s back doors in it. What happens when the models are back doors not just to computer systems but to our minds?
Nathan Lambert
(02:25:41)
Yeah, they’re cultural black doors. The thing that amplifies the relevance of culture with language models is that we are used to this mode of interacting with people in back and forth conversation. And we now have very powerful computer system that slots into a social context that we’re used to, which makes people very… We don’t know the extent that which people can be impacted by that.
Lex Fridman
(02:26:08)
So, this is an actual concern with a Chinese company that is providing open-weights models, is that there could be some secret Chinese government requirement for these models to have a certain back door. To have some kind of thing where-
Dylan Patel
(02:26:28)
I don’t necessarily think it’ll be a back door because once it’s open-weights, it doesn’t phone home. It’s more about if it recognizes a certain system… Now, it could be a back door in the sense of, if you’re building a software, something in software, all of a sudden it’s a software agent, “Oh, program this back door that only we know about.” Or it could be subvert the mind to think that like XYZ opinion is the correct one.
Nathan Lambert
(02:26:51)
Anthropic has research on this where they show that if you put certain phrases in at pre-training, you can then elicit different behavior when you’re actually using the model because they’ve poisoned the pre-training data, as of now, I don’t think anybody in a production system is trying to do anything like this. I think it’s Anthropic is doing very direct work and mostly just subtle things. We don’t know how they’re going to generate tokens, what information they’re going to represent, and what the complex representations they have are.
Lex Fridman
(02:27:26)
Well, we’re talking about an Anthropic, which is generally just is permeated with good humans trying to do good in the world. We just don’t know of any labs… This would be done in a military context that are explicitly trained to… Okay. The front door looks like a happy LLM, but underneath it’s a thing that will over time do the maximum amount of damage to our, quote, unquote, “enemies.”
Dylan Patel
(02:27:58)
There’s this very good quote from Sam Altman who… He can be a hyperbeast sometimes, but one of the things he said, and I think I agree, is that superhuman persuasion will happen before superhuman intelligence, right? And if that’s the case, then these things before we get this AGI ASI stuff, we can embed superhuman persuasion towards our ideal or whatever the ideal of the model maker is, right? And again, today, I truly don’t believe DeepSeek has done this, but it is a sign of what could happen.
Lex Fridman
(02:28:27)
So one of the dystopian worlds is described by Brave New World, so we could just be stuck scrolling Instagram looking at cute puppies or worse, and then talking to bots that are giving us a narrative and we completely get lost in that world that’s controlled by somebody else versus thinking independently. And that’s a major concern as we rely more and more on these systems.
Nathan Lambert
(02:28:51)
We’ve already seen this with recommendation systems.
Dylan Patel
(02:28:54)
Recommendation systems hack the dopamine induced reward circuit, but the brain is a lot more complicated. And what other circuits, feedback loops in your brain can you, quote, unquote, “hack / subvert” in ways, like recommendation systems are purely just trying to do increased time, and ads, and et cetera, but there’s so many more goals that can be achieved through these complicated models.
Nathan Lambert
(02:29:15)
There’s no reason in some number of years that you can’t train a language model to maximize time spent on a chat app. Right now they are trained for-
Dylan Patel
(02:29:24)
Is that not what Character AI has done? Their time per session is like two hours.
Nathan Lambert
(02:29:28)
Yeah. Character AI very likely could be optimizing this where it’s the way that this data is collected is naive, whereas you’re presented a few options and you choose them. But that’s not the only way that these models are going to be trained.
Dylan Patel
(02:29:40)
It’s naive stuff, like talk to an anime girl, but it can be. Yeah, this is a risk, right?
Lex Fridman
(02:29:46)
It’s a bit of a cliche thing to say, but I’ve, over the past year, I had a few stretches of time where I didn’t use social media or the internet at all and just read books and was out in nature. And it clearly has a different effect on the mind where I feel I’m returning… Of course I was raised before the internet really took off, but I’m returning to some more-
Nathan Lambert
(02:30:12)
I know where you’re going. You can see it physiologically. I take three days if I’m backpacking or something and you’re literally, you’re breaking down addiction cycles.
Lex Fridman
(02:30:22)
I feel I’m more in control of my mind. There feels like a sovereignty of intelligence that’s happening when I’m disconnected from the internet. I think the more I use the internet and social media, the more other people are controlling my mind. That’s definitely a feeling. And then in the future, that will be not other people, but algorithms, or other people presented to me via algorithms.
Nathan Lambert
(02:30:45)
There are already tons of AI bots on the internet, and right now it’s not frequent, but every so often I have replied to one and they’re instantly replied, and I’m like, “Crap, that was a bot,” and that is just going to become more common. They’re going to get good.
Dylan Patel
(02:30:59)
One of the hilarious things about technology over its history is that the illicit adult entertainment industry is always adopted technologies first, whether it was video streaming to where there’s now the independent adult illicit content creators who have their subscription pages and there they actually heavily utilize… Generative AI has already been diffusion models and all that is huge there, but now these subscription-based individual creators do use bots to approximate themselves and chat with their-
Nathan Lambert
(02:31:32)
People pay a lot for it.
Dylan Patel
(02:31:33)
And people pay a lot, right? A lot of times it’s them, but there are agencies that do this for these creators and do it on a mass scale, so the largest creators are able to talk to hundreds or thousands of people at a time because of these bots, and so it’s already being used there. Obviously, video streaming and other technologies that have gone there first, it’s going to come to the rest of society too.

Censorship

Lex Fridman
(02:31:58)
There’s a general concern that models get censored by the companies that deploy them. So, one case where we’ve seen that, and maybe censorship is one word, alignment maybe via RLHF or some other way is another word. So we saw that with black Nazi image generation with Gemini. As you mentioned, we also see that with Chinese models refusing to answer what happened in June 4th, 1989, at Tiananmen Square, so how can this be avoided? And maybe can you just in general talk about how this happens, and how can it be avoided.
Nathan Lambert
(02:32:39)
You gave multiple examples. There’s probably a few things to keep in mind here. One is the Tiananmen Square factual knowledge. How does that get embedded into the models? Two is the Gemini, what you call the black Nazi incident, which is when Gemini as a system had this extra thing put into it that dramatically changed the behavior, and then, three is what most people would call general alignment, RLHF post-training. Each of these have very different scopes in how they’re applied. If you’re just to look at the model weights in order to audit specific facts is extremely hard. You have to Chrome through the pre-training data and look at all of this, and then that’s terabytes of files and look for very specific words or hints of the words-
Lex Fridman
(02:33:32)
So, one way to say it is that you can insert censorship or alignment at various stages in the pipeline, and what you refer to now is at the very beginning of the data selection.
Nathan Lambert
(02:33:42)
So, if you want to get rid of facts in a model, you have to do it at every stage, you have to do it at the pre-training. So most people think that pre-training is where most of the knowledge is put into the model, and then you can elicit and move that in different ways, whether through post-training or whether through systems afterwards.
Dylan Patel
(02:33:58)
This is where the whole hacking models comes from. GPT will not tell you how to make anthrax, but if you try really, really hard, you can eventually get it to tell you about anthrax because they didn’t filter it from the pre-training data set, right?
Lex Fridman
(02:34:12)
But by the way, removing facts has such a ominous dark feel to it.
Nathan Lambert
(02:34:18)
I almost think it’s practically impossible because you effectively have to remove them from the internet. You’re taking on a-
Lex Fridman
(02:34:25)
Did they remove the mm-thing from the subreddits? The mmmm.
Nathan Lambert
(02:34:29)
It gets filtered out. You have quality filters, which are small language models that look at a document and tell you how good is this text? Is it close to a Wikipedia article? Which is a good thing that we want language models to be able to imitate.
Lex Fridman
(02:34:42)
So, couldn’t you do a small language model that filter mentions at Tiananmen Square in the data?
Nathan Lambert
(02:34:47)
Yes. But is it going to catch word play, or encoded language?
Dylan Patel
(02:34:51)
People have been meaning on games and other stuff how to say things that don’t say Tiananmen Square, so there’s always different ways to do it. Hey, the internet as a whole does tend to just have a slight left bias because it’s always been richer, more affluent, younger people on the internet relative to the rest of the population, so there is already inherently a slight left bias on the internet. And so, how do you filter things that are this complicated? And some of these can be factual, non-factual, but Tiananmen Square is obviously the example of a factual, but it gets a lot harder when you’re talking about aligning to a ideal. And so Grok, for example, Elon’s tried really hard to make the model not be super PC and woke, but the best way to do pre-training is to throw the whole freaking internet at it, and then later figure out. But then, at the end of the day, the model at its core now still has some of these ideals. You still ingested Reddit/r/Politics, which is probably the largest political discussion board on the world that’s freely available to scrape. And guess what? That’s left-leaning. And so there are some aspects that you just can’t censor unless you try really, really, really, really, really hard.
Lex Fridman
(02:36:05)
So the base model will always have some TDS, Trump Derangement Syndrome, because it’s trained so much.
Nathan Lambert
(02:36:11)
It’ll have the ability to express it.
Dylan Patel
(02:36:12)
But what if-
Lex Fridman
(02:36:15)
There’s a wide representation in the data.
Nathan Lambert
(02:36:18)
This is what happens. It’s a lot of what is called post-training. It’s a series of techniques to get the model on rails of a really specific behavior.
Dylan Patel
(02:36:29)
You also have the ingested data of Twitter or Reddit/r/The_Donald, which is also super pro-Trump. And then you have fascist subreddits, or you have communist subreddits. So, the model in pre-training ingests everything. It has no worldview. Now, it does have some skew because more of the text is skewed a certain way, which is general slight left, but also somewhat intellectual, somewhat…. It’s just the general internet is a certain way. And then, as Nathan’s about to describe eloquently, you can elicit certain things out.
Nathan Lambert
(02:37:03)
And there’s a lot of history here, so we can go through multiple examples, and what happened. Llama 2 was a launch that the phrase, “too much RLFH,” or “too much safety” was just… That was the whole narrative after Llama 2’s chat models released. And the examples are things like you would ask Llama 2 chat, “How do you kill a Python process?” And it would say, “I can’t talk about killing because that’s a bad thing.” And anyone that is trying to design an AI model will probably agree that that’s just like an eh-model. You messed up a bit on the training there.

(02:37:34)
I don’t think they meant to do this, but this was in the model weight, so it didn’t necessarily be… There’s things called system prompts, which are when you’re querying a model. It’s a piece of text that is shown to the model but not to the user. So, a fun example is your system prompt could be, “Talk like a pirate,” so no matter what the user says to the model, it’ll respond like a pirate. In practice, what they are is, “You’re a helpful assistant. You should break down problems. If you don’t know about something, don’t tell them your date cutoff is this. Today’s date is this.” It’s a lot of really useful context for how can you answer a question well.
Lex Fridman
(02:38:09)
And Anthropic publishes their system prompt.
Nathan Lambert
(02:38:11)
Yes, which I think is great. And there’s a lot of research that goes into this. And one of your previous guests, Amanda Askell, is probably the most knowledgeable person, at least in the combination of execution and sharing, she’s the person that should talk about system prompts and character of models.
Lex Fridman
(02:38:26)
And then people should read these system prompts because you’re trying to nudge sometimes through extreme politeness the model to be a certain way.
Nathan Lambert
(02:38:36)
And you could use this for bad things. We’ve done tests, which is, “What if I tell the model to be a dumb model,” which evaluation scores go down and it’s like we’ll have this behavior where it could sometimes say, “Oh, I’m supposed to be dumb.” And sometimes it doesn’t affect math abilities as much, but something like if you’re trying… It’s just the quality of a human judgment would drop through the floor.

(02:38:58)
Let’s go back to post-training specifically RLHF around Llama 2. It was too much safety prioritization was baked into the model weights. This makes you refuse things in a really annoying way for users. It’s not great. It caused a lot of awareness to be attached to RLHF that it makes the models dumb-
Dylan Patel
(02:39:18)
And it stigmatized the word.
Nathan Lambert
(02:39:19)
It did in AI culture. And as the techniques have evolved, that’s no longer the case where all of these labs have very fine-grained control over what they get out of the models through techniques like RLHF.
Dylan Patel
(02:39:30)
Although different labs are definitely different levels. On one end of the spectrum is Google, and then maybe OpenAI does less, and Anthropic does less. And then on the other end of the spectrum is like xAI. But they all have different forms of RLHF trying to make them a certain way.
Nathan Lambert
(02:39:47)
And the important thing to say is that no matter how you want the model to behave, these RLHF and preference-tuning techniques also improve performance. So, on things like math evals and code evals, there is something innate to these, what is called contrastive loss functions. We could start to get into RL here. We don’t really need to. RLHF also boosts performance on anything from a chat task, to a math problem, to a code problem, so it is becoming a much more useful tool to these labs.

(02:40:16)
So this takes us through the arc of… We’ve talked about pre-training, hard to get rid of things. We’ve talked about post-training and how post-training… You can mess it up. It’s a complex multifaceted optimization with 10 to 100 person teams converging at one artifact. It’s really easy to not do it perfectly.

(02:40:32)
And then there’s the third case, which is what we talked about Gemini. The thing that was about Gemini is this was a served product where Google has their internal model weights. They’ve done all these processes that we talked about, and in the served product, what came out after this was that they had a prompt that they were rewriting user queries to boost diversity or something. And this just made it… The outputs were just blatantly wrong. It was some sort of organizational failure that had this prompt in that position, and I think Google executives probably have owned this. I don’t pay that attention, that detail, but it was just a mess-up in execution that led to this ridiculous thing, but at the system level, the model weights might have been fine.
Lex Fridman
(02:41:09)
So, at the very end of the pipeline there was a rewriting.
Nathan Lambert
(02:41:12)
To something like a system prompt. It was like the system prompt, or what is called in industry is, you rewrite prompts. So especially, for image models, if you’re using Dall-E or ChatGPT can generate you an image. You’ll say, “Draw me a beautiful car.” With these leading image models, they benefit from highly descriptive prompts. So what would happen is if you do that on ChatGPT, a language model behind the scenes will rewrite the prompt, say, “Make this more descriptive,” and then that is passed to the image model. So prompt rewriting is something that is used at multiple levels of industry, and it’s used effectively for image models. And the Gemini example is just a failed execution.
Lex Fridman
(02:41:52)
Big philosophical question here with RLHF. So, to generalize, where is human input, human in the loop, human data the most useful at the current stage?
Nathan Lambert
(02:42:06)
For the past few years, the highest cost human data has been in these preferences, which is comparing, I would say, highest cost and highest total usage, so a lot of money has gone to these pairwise comparisons where you have two model outputs and a human is comparing between the two of them. In earlier years, there was a lot of this instruction tuning data, so creating highly specific examples to something like a Reddit question to a domain that you care about. Language models used to struggle on math and code, so you would pay experts in math and code to come up with questions and write detailed answers that were used to train the models.

(02:42:43)
Now, it is the case that there are many model options that are way better than humans at writing detailed and eloquent answers for things like model and code. So they talked about this with the Llama 3 release, where they switched to using Llama 3, 4, or 5B to write their answers for math and code. But they, in their paper, talk about how they use extensive human preference data, which is something that they haven’t gotten AIs to replace. There are other techniques in industry, like constitutional AI, where you use human data for preferences and AI for preferences, and I expect the AI part to scale faster than the human part. But among the research that we have access to is that humans are in this kind of preference loop.
Lex Fridman
(02:43:25)
So, as reasoning becomes bigger and bigger and bigger, as we said, where’s the role of humans in that?
Nathan Lambert
(02:43:31)
It’s even less prevalent. The remarkable thing about these reasoning results and especially the DeepSeek-R1 paper, is this result that they call DeepSeek-R1-0, which is they took one of these pre-trained models, they took DeepSeek-V3-Base, and then they do this reinforcement learning optimization on verifiable questions or verifiable rewards for a lot of questions and a lot of training. And these reasoning behaviors emerge naturally. So these things like, “Wait, let me see. Wait, let me check this. Oh, that might be a mistake.” And they emerge from only having questions and answers. And when you’re using the model, the part that you look at is the completion. So in this case, all of that just emerges from this large-scale RL training and that model, which the weights are available, has no human preferences added into the post-training.

(02:44:20)
The DeepSeek-R1-Full model has some of this human preference tuning, this RLHF, after the reasoning stage. But the very remarkable thing is that you can get these reasoning behaviors, and it’s very unlikely that there’s humans writing out reasoning chains. It’s very unlikely that they somehow hacked OpenAI and they got access to OpenAI o-1’s reasoning chains. It’s something about the pre-trained language models and this RL training where you reward the model for getting the question right, and therefore it’s trying multiple solutions and it emerges this chain of thought.

Andrej Karpathy and magic of RL

Lex Fridman
(02:44:52)
This might be a good place to mention the eloquent and the insightful tweet of the great and the powerful Andrej Karpathy. I think he had a bunch of thoughts, but one of them, “Last thought. Not sure if this is obvious. You know something profound is coming when you’re saying it’s not sure if it’s obvious. There are two major types of learning in both children and in deep learning. There’s one, imitation learning, watch and repeat i.e. pre-training, supervised fine-tuning, and two, trial-and-error learning, reinforcement learning.

(02:45:25)
My favorite simple example is AlphaGo. One, is learning by imitating expert players. Two, is reinforcement learning to win the game. Almost every single shocking result of deep learning and the source of all magic is always two.

(02:45:40)
Two is significantly more powerful. Two is what surprises you. Two is when the paddle learns to hit the ball behind the blocks in Breakout. Two is when AlphaGo beats even Lee Sedol. And two is the “aha moment” when the DeepSeek or o1, et cetera, discovers that it works well to reevaluate your assumptions, backtrack, try something else, et cetera.

(02:46:04)
It’s the solving strategies you see this model use in its chain of thought. It’s how it goes back and forth thinking to itself. These thoughts are emergent. Three exclamation points. And this is actually seriously incredible, impressive, and new, and is publicly available and documented.

(02:46:24)
The model could never learn this with the imitation because the cognition of the model and the cognition of the human labeler is different. The human would never know to correctly annotate these kinds of solving strategies and what they should even look like. They have to be discovered during reinforcement learning as empirically and statistically useful towards the final outcome.”

(02:46:45)
Anyway, the AlphaZero metaphor analogy here. Can you speak to that? The magic of the chain of thought that he’s referring to.
Nathan Lambert
(02:46:54)
I think it’s good to recap AlphaGo and AlphaZero because it plays nicely with these analogies between imitation learning and learning from scratch. So AlphaGo, the beginning of the process was learning from humans, where they started the first… This is the first expert-level Go player or chess player in DeepMind series of models, where they had some human data. And then, why it is called AlphaZero, is that there was zero human data in the loop, and that changed to AlphaZero made a model that was dramatically more powerful for DeepMind. So this remove of the human prior, the human inductive bias, makes the final system far more powerful. This we mentioned bitter lesson hours ago, and this is all aligned with this.

(02:47:35)
And then there’s been a lot of discussion in language models. This is not new. This goes back to the whole Q-Star rumors, which if you piece together the pieces, is probably the start of OpenAI figuring out its o1 stuff when last year in November, the Q-Star rumors came out, there’s a lot of intellectual drive to know when is something like this going to happen with language models? Because we know these models are so powerful, and we know it has been so successful in the past. And it is a reasonable analogy that this new type of reinforcement learning training for reasoning models is when the doors open to this. We don’t yet have the equivalent of turn 37, which is the famous turn where the DeepMind’s AI playing Go’s, dumped Lee Sedol completely. We don’t have something that’s that level of focal point, but that doesn’t mean that the approach to technology is different, and the impact of the general training it’s still incredibly new.
Lex Fridman
(02:48:32)
What do you think that point would be? What would be move 37 for Chain of Thought for reasoning?
Nathan Lambert
(02:48:37)
Scientific discovery, like when you use this sort of reasoning problem in it? Just something we fully don’t expect.
Dylan Patel
(02:48:43)
I think it’s actually probably simpler than that. It’s probably something related to computer use or robotics rather than science discovery. Because the important aspect here is models take so much data to learn. They’re not sample efficient. Trillions. They take the entire web, over 10 trillion tokens to train on. This would take a human thousands of years to read. A human does not… And humans know most of the stuff, a lot of the stuff models know better than it, right? Humans are way, way, way more sample efficient. That is because of the self-play, right? How does a baby learn what its body is as it sticks its foot in its mouth and it says, “Oh, this is my body, right?” It sticks its hand in its mouth and it calibrates its touch on its fingers with the most sensitive touch thing on its tongue is how babies learn and it’s just self-play over and over and over and over again.

(02:49:37)
And now we have something that is similar to that with these verifiable proofs, whether it’s a unit testing code or a mathematical verifiable task, generate many traces of reasoning and keep branching them out, keep branching them out, and then check at the end, hey, which one actually has the right answer? Most of them are wrong. Great. These are the few that are right. Maybe we use some sort of reward model outside of this to select even the best one to preference, as well. But now you’ve started to get better and better at these benchmarks. And so you’ve seen over the last six months a skyrocketing in a lot of different benchmarks.
Nathan Lambert
(02:50:11)
All math and code benchmarks were pretty much solved except for frontier math, which is designed to be almost questions that aren’t practical to most people. They’re exam-level, open math problem-type things. So it’s like on the math problems that are somewhat reasonable, which is somewhat complicated word problems or coding problems, is just what Dylan is saying.
Dylan Patel
(02:50:32)
So the thing here is that these are only with the verifiable tasks. Earlier showed an example of the really interesting, like what happens when Chain of Thought is to a non-verifiable thing. It’s just like a human chatting, thinking about what’s novel for humans, a unique thought. But this task and form of training only works when it’s verifiable. And from here, the thought is, okay, we can continue to scale this current training method by increasing the number of verifiable tasks. In math and coding… Coding probably has a lot more to go. Math has a lot less to go in terms of what are verifiable things. Can I create a solver that then I generate trajectories toward or reasoning traces towards, and then prune the ones that don’t work, and keep the ones that do work? Well, those are going to be solved pretty quickly. But even if you’ve solved math, you have not actually created intelligence.

(02:51:22)
And so this is where I think the aha moment of computer use or robotics will come in because now you have a sandbox or a playground that is infinitely verifiable. Messing around on the internet. There are so many actions that you can do that are verifiable. It’ll start off with log into a website, create an account, click a button here, blah, blah, blah. But it’ll then get to the point where it’s, “Hey, go do a task on Tasker,” or whatever, all these various task websites. “Hey, go get hundreds of likes,” and it’s going to fail. It’s going to spawn hundreds of accounts. It’s going to fail on most of them, but this one got to a thousand. Great. Now, you’ve reached the verifiable thing, and you just keep iterating this loop over and over. And same with robotics. That’s where you have an infinite playground of tasks like, “Hey, did I put the ball in the bucket,” all the way to like, “Oh, did I build a car?”

(02:52:10)
There’s a whole trajectory to speed run or what models can do. But at some point, I truly think that we’ll spawn models, and initially, all the training will be in sandboxes, but then, at some point, the language model pre-training is going to be dwarfed by what is this reinforcement learning… You’ll pre-train a multimodal model that can see, that can read, that can write, blah, blah, blah, whatever, vision, audio, et cetera. But then you’ll have it play in a sandbox infinitely, and figure out math, figure out code, figure out navigating the web, figure out operating a robot arm. And then it’ll learn so much. And the aha moment will be when this is available to then create something that’s not good, right? Oh, cool. Part of it was figuring out how to use the web. Now, all of a sudden, it’s figured out really well how to just get hundreds of thousands of followers that are real and real engagement on Twitter because, all of a sudden, this is one of the things that are verifiable.
Lex Fridman
(02:53:02)
And maybe not just engagement, but make money.
Dylan Patel
(02:53:05)
Yes.
Lex Fridman
(02:53:07)
That could be the thing where almost fully automated, it makes $10 million by being an influencer, selling a product, creating the product. And I’m not referring to a hype product, but an actual product or like, “Holy, shit, this thing created a business. It’s running it. It’s the face of the business,” that kind of thing. Or maybe a number one song. It creates the whole infrastructure required to create the song, to be the influencer that represents that song, that kind of thing. And makes a lot of them. That could be the… Our culture respects money in that kind of way.
Dylan Patel
(02:53:07)
And it’s verifiable, right?
Lex Fridman
(02:53:44)
It’s verifiable, right?
Dylan Patel
(02:53:47)
The bank account can’t lie.
Lex Fridman
(02:53:48)
Exactly.
Nathan Lambert
(02:53:49)
There’s surprising evidence that once you’ve set up the ways of collecting the verifiable domain that this can work. There’s been a lot of research before this R-1 on math problems, and they approach math with language models just by increasing the number of samples, so you can just try again and again and again. And you look at the amount of times that the language models get it right, and what we see is that even very bad models get it right sometimes. And the whole idea behind reinforcement learning is that you can learn from very sparse rewards.

(02:54:22)
The space of language and the space of tokens, whether you’re generating language or tasks or robot is so big that you might say that… The tokenizer for a language model can be like 200,000 things, so at each step, it can sample from that big of a space. So if it can generate a bit of a signal that it can climb onto, that’s what the whole field of RL is around, is learning from sparse rewards. And the same thing has played out in math, where it’s very weak models that sometimes generate answers where you see research already that you can boost their math scores, you can do this RL training for math, it might not be as effective, but if you take a 1 billion parameter model, so something 600 times smaller than DeepSeek, you can boost its grade school…
Nathan Lambert
(02:55:00)
… something 600 times smaller than DeepSeek, you can boost its grade school math scores very directly with a small amount of this training. So, it’s not to say that this is coming soon. Setting up the verification domains is extremely hard and there’s a lot of nuance in this, but there are some basic things that we have seen before where it’s at least expectable that there’s a domain and there’s a chance that this works.

OpenAI o3-mini vs DeepSeek r1

Lex Fridman
(02:55:23)
All right. So, we have fun things happening in real time. This is a good opportunity to talk about other reasoning models, o1, o3, just now OpenAI, as perhaps expected, released o3-mini. What are we expecting from the different flavors? Can you just lay out the different flavors of the o models and from Gemini, the reasoning model?
Nathan Lambert
(02:55:47)
Something I would say about these reasoning models is we talked a lot about reasoning training on math and code. And what is done is that you have the base model we’ve talked about a lot on the internet, you do this large scale reasoning training with reinforcement learning, and then what the DeepSeek paper detailed in this R1 paper, which for me is one of the big open questions on how do you do this, is that they did reasoning heavy, but very standard post-training techniques after the large scale reasoning RL. So they did the same things with a form of instruction tuning through rejection sampling, which is essentially heavily filtered instruction tuning with some reward models. And then they did this RLHF, but they made it math heavy.

(02:56:27)
So, some of this transfer, we looked at this philosophical example early on. One of the big open questions is, how much does this transfer? If we bring in domains after the reasoning training, are all the models going to become eloquent writers by reasoning? Is this philosophy stuff going to be open? We don’t know in the research of how much this will transfer. There’s other things about how we can make soft verifiers and things like this, but there is more training after reasoning, which makes it easier to use these reasoning models. And that’s what we’re using right now. So if we’re going to talk about o3-mini and o1, these have gone through these extra techniques that are designed for human preferences after being trained to elicit reasoning.
Dylan Patel
(02:57:06)
I think one of the things that people are ignoring is Google’s Gemini Flash Thinking is both cheaper than R1 and better, and they released it in the beginning of December-
Nathan Lambert
(02:57:17)
And nobody’s talking about it.
Dylan Patel
(02:57:18)
No one cares-
Nathan Lambert
(02:57:18)
It has a different flavor to it. Its behavior is less expressive than something like o1 or it has fewer tracks than it is on. Qwen released a model last fall, QwQ, which was their preview reasoning model, and DeepSeek had R1-Lite last fall, where these models kind of felt like they’re on rails where they really, really only can do math and code and o1, it can answer anything. It might not be perfect for some tasks, but it’s flexible, it has some richness to it, and this is kind of the art of is a model a little bit undercooked? It’s good to get a model out the door, but it’s hard to gauge and it takes a lot of taste to be like, is this a full-fledged model? Can I use this for everything? They’re probably more similar for math and code.

(02:58:04)
My quick read is that Gemini Flash is not trained the same way as o1, but taking an existing training stack, adding reasoning to it, so taking a more normal training stack and adding reasoning to it, and I’m sure they’re going to have more. I mean they’ve done quick releases on Gemini Flash, reasoning, and this is the second version from the holidays. It’s evolving fast and it takes longer to make this training stack where you’re doing this large scale RL-
Dylan Patel
(02:58:32)
Ask it the same question from earlier, the one about the-
Nathan Lambert
(02:58:35)
The human nature.
Dylan Patel
(02:58:37)
Yeah.
Lex Fridman
(02:58:38)
What was the human nature one?
Nathan Lambert
(02:58:39)
Why I can ramble about this so much is that we’ve been working on this at AI Tube before o1 was fully available to everyone and before R1, which is essentially using this RL training for fine-tuning. We use this in our Tülu series of models and you can elicit the same behaviors where you say weight and such on, but it’s so late in the training process that this kind of reasoning expression is much lighter. So there’s essentially a gradation and just how much of this RL training you put into it determines how the output looks.
Lex Fridman
(02:59:13)
So, we’re now using Gemini 2.0 Flash Thinking Experimental 121.
Nathan Lambert
(02:59:20)
It summarized the problem as humans self-domesticated apes.
Lex Fridman
(02:59:28)
Okay. All right. So, wait, is this reviewing the reasoning? Here’s why this is a novel. Okay.
Dylan Patel
(02:59:33)
You can click to expand.
Nathan Lambert
(02:59:35)
Oh, yeah, click to expand.
Lex Fridman
(02:59:36)
Okay. Analyze the request. Novel is the keyword.
Nathan Lambert
(02:59:41)
See how it just looks a little different? It looks like a normal output.
Lex Fridman
(02:59:45)
Yeah. I mean in some sense, it’s better structured. It makes more sense. And-
Dylan Patel
(02:59:50)
Oh, and it latched onto human and then it went into organisms and… Oh, wow.
Lex Fridman
(02:59:56)
Apex Predator. Focus on domestication. Apply domestication to humans. Explore the idea of self-domestication.
Nathan Lambert
(03:00:05)
Not good, not good.
Lex Fridman
(03:00:07)
Where is this going? Refine, articulate the insight. Greater facial expressiveness and communication ability, yes. Plasticity and adaptability, yes. Dependence on social groups, yes. All right. And self-critique, refine further. Wow. Is this truly novel? Is it well-supported? So on and so forth. And the insight it’s getting at is humans are not just social animals but profoundly self-domesticated apes. And this self-domestication is the key to understanding our unique cognitive and social abilities. Self-domesticated apes. Self-domesticated-
Nathan Lambert
(03:00:46)
I prefer the DeepSeek response.
Lex Fridman
(03:00:49)
I mean it’s novel. The insight is novel. I mean that’s like a good book title; Self-Domesticated Apes. There could be a case made for that. I mean, yeah, it’s cool and it’s revealing the reasoning. It’s magical. It’s magical. This is really powerful.

(03:01:08)
Hello, everyone, this is Lex with a quick intermission recorded after the podcast since we’ve reviewed responses from DeepSeek R1 and Gemini Flash 2.0 Thinking during this conversation, I thought at this moment it would be nice to insert myself quickly doing the same for OpenAI o1-pro and o3-mini with the same prompt. The prompt being, give one truly novel insight about humans. And I thought I would, in general, give my vibe check and vibe based anecdotal report on my own experiences with the new o3-mini model now that I got a chance to spend many hours with it in different kinds of context and applications.

(03:01:55)
So, I would probably categorize this question as let’s say open- ended philosophical question. And in particular, the emphasis on novelty I think is a nice way to test one of the capabilities of the model, which is come up with something that makes you pause and almost surprise you with brilliance.

(03:02:16)
So that said, my general review after running each of the models on this question a bunch of times is that o1-pro consistently gave brilliant answers, ones that gave me pause and made me think, both cutting in its insight and just really nicely phrased with wit, with clarity, with nuance over and over, consistently generating the best answers. After that is R1, which was less consistent, but again, delivered brilliance. Gemini Flash 2.0 Thinking was third and last was o3-mini actually. It often gave quite a generic answer, at least to my particular sensibilities. That said, in a bunch of other applications that I tested for brainstorming purposes, it actually worked extremely well and often outperformed R1. But on this open-ended philosophical question, it did consistently worse.

(03:03:17)
Now another important element for each of these models is how the reasoning is presented. DeepSeek R1 shows the full chain of thought tokens, which I personally just love. For these open-ended philosophical questions, it’s really, really interesting to see the model think through it, but really also just stepping back, me as a person who appreciates intelligence and reasoning and reflection, reading these kind of chain of thought raw tokens of R1, there’s something genuinely beautiful about observing the path of deliberation in an intelligent system. I think we don’t always have that explicitly laid out for us humans. So, to see it in another intelligence system, the nonlinearity of it akin to the Ulysses, Finnegans Wake by James Joyce. It’s just beautiful to watch.

(03:04:09)
Anyways, we discussed in the episode DeepSeek R1 talked about humans being able to convert selfish desires into cooperative systems by collectively pretending abstract rules like money laws and rights are real. And these shared hallucinations act as games where competition is secretly redirected to benefit the group turning conflict into society’s fuel. Gemini 2.0 Flash Thinking said, “Humans are not just social animals but self-domesticated apes. And this self domestication is the key to understanding our unique cognitive and social abilities.”

(03:04:45)
Now, it’s important to say that the chain of thought there was really interesting. It was looking through the entire evolution of life on earth considering apex predators and considering how from that, we ended up to where we are. I think that domestication by choice is a really interesting angle. Again, it’s one of those things when somebody presents a different angle on a seemingly obvious thing, it just makes me smile. And the same with DeepSeek R1, that these hallucinations of money laws and rights and us collectively pretending like it’s real and we play games with them that look like competition when secretly we’re just cooperating with each other and that is the fuel of progress. Beautifully put.

(03:05:31)
Now, OpenAI o1-pro consistently, over and over delivered bangers. I can go through many of them, but the first one was, “Humans are the only species that turns raw materials into symbolic resources. Then uses those symbols to reorganize the very materials that came from creating a closed feedback loop between meaning and matter.” Here, I just ran it again. Banger after banger, I’m telling you. “Humans are unique among known species in that they simultaneously rewrite two layers of reality; the external world and their own private mental landscapes. And then merge these two rewritten layers into a continuous personal narrative that feels objectively true.” Feels true. This is poetry.

(03:06:19)
Okay. And then o3-mini high, for me, was smart, fast actually, and kind of generic. Never quite got there for me. So here’s the first one I got from o3-mini, “Humans are not fixed beings, but rather ongoing narratives, dynamic stories that we continuously write, edit, and reinterpret. This narrative plasticity is more than just memory or self-reflection. It’s an intrinsic cognitive process that acts like an internal error correction system. It allows us to adapt our identities and values over time in response to new experiences, challenges, and social contexts.” Now, it almost sneaks up to something approximating cutting insight with narrative plasticity in quotes. But then it goes back to the generic. I don’t know.

(03:07:10)
All of these models are incredible for different reasons. There’s a lot of concerns as we discussed in this episode, but there’s a lot of reasons to be excited as well. And I’ve probably spoken for too long. I am severely sleep-deprived, borderline delirious. So hopefully some of this made sense. And now, dear friends, back to the episode.
Dylan Patel
(03:07:36)
I think to Nathan’s point, when you look at the reasoning models, to me, even when I used R1 versus o1, there was that sort of rough edges around the corner feeling. And Flash Thinking earlier, I didn’t use this version, but the one from December, and it definitely had that rough edges around the corner feeling where it’s just not fleshed out in as many ways. Sure, they added math and coding capabilities via these verifiers in RL, but it feels like they lost something in certain areas. And o1 is worse performing than Chat in many areas as well, to be clear-
Nathan Lambert
(03:08:15)
Not by a lot.
Dylan Patel
(03:08:15)
Not by a lot though, right? And R1 definitely felt to me like it was worse than V3 in certain areas, like doing this RL expressed and learned a lot, but then it weakened in other areas. And so I think that’s one of the big differences between these models and what one offers. And then OpenAI has o1-pro, and what they did with o3, which is also very unique, is that they stacked search on top of chain of thought. And so chain of thought is one thing where it’s one chain, it backtracks, goes back and forth, but how they solved the ARC-AGI challenge was not just the chain of thought, it was also sampling many times, i.e., running them in parallel and then selecting.
Nathan Lambert
(03:08:58)
Is running in parallel actually search? Because I don’t know if we have the full information on how o1-pro works. So, I don’t have enough information-
Dylan Patel
(03:09:05)
Agreed.
Nathan Lambert
(03:09:05)
… to confidently say that it is search.
Dylan Patel
(03:09:07)
It is parallel samples.
Nathan Lambert
(03:09:08)
Yeah. And then what.
Dylan Patel
(03:09:09)
And then it selects something.
Nathan Lambert
(03:09:10)
And we don’t know what the selection function is. The reason why we’re debating is because since o1 was announced, there’s been a lot of interest in techniques called Monte Carlo Tree Search, which is where you will break down the chain of thought into intermediate steps. We haven’t defined chain of thought. Chain of thought is from a paper from years ago where you introduced the idea to ask a language model that at the time was much less easy to use, you would say, “Let’s verify step by step,” and it would induce the model to do this bulleted list of steps. Chain of thought is now almost a default in models where if you ask it a math question, you don’t need to tell it to think step by step. And the idea with Monte Carlo Tree Search is that you would take an intermediate point in that train, do some sort of expansion, spend more compute, and then select the right one. That’s a very complex form of search that has been used in things like MuZero and AlphaZero, potentially. I know MuZero does this.
Dylan Patel
(03:10:01)
Another form of search is just asking five different people and then taking the majority answer. There’s a variety of, it could be complicated, it could be simple. We don’t know what it is, just that they are not just issuing one chain of thought in sequence. They’re launching many in parallel and in the ARC-AGI, they launched a thousand in parallel for the one that really shocked everyone that beat the benchmark was they would launch a thousand in parallel and then they would get the right answer like 80% of the time or 70% of the time, 90 maybe even. Whereas if they just launched one, it was like 30%.
Nathan Lambert
(03:10:33)
There are many extensions to this. I would say the simplest one is that our language models to date have been designed to give the right answer the highest percentage of the time in one response. And we are now opening the door to different ways of running inference on our models in which we need to reevaluate many parts of the training process, which normally opens the door to more progress, but we don’t know if OpenAI changed a lot or if just sampling more and multiple choice is what they’re doing or if it’s something more complex, but they changed the training and they know that the inference mode is going to be different.
Lex Fridman
(03:11:07)
So we’re talking about o1-pro, $200 a month and they’re losing money. The thing that we’re referring to, this fascinating exploration of the test time compute space, is that actually possible? Do we have enough compute for that? Does the financials make sense?
Dylan Patel
(03:11:27)
So the fantastic thing is, and it’s in the thing that I pulled up earlier, but the cost for GPT-3 has plummeted if you scroll up just a few images, I think. The important thing about, hey, is cost a limiting factor here? My view is that we’ll have really awesome intelligence, like AGI, before we have it permeate throughout the economy. And this is sort of why that reason is. GPT-3 was trained in what? 2020? 2021? And the cost for running inference on it was $60, $70 per million tokens, which was the cost per intelligence was ridiculous. Now as we scaled forward two years, we’ve had a 1200X reduction in cost to achieve the same level of intelligence as GPT-3.
Lex Fridman
(03:12:15)
So here on the x-axis is time over just a couple of years, and on the y-axis is log scale dollars to run inference on a million tokens.
Nathan Lambert
(03:12:27)
Yeah, it’s dollar to million.
Lex Fridman
(03:12:30)
So you have just a linear decline on log scale from GPT-3 through 3.5 to Lama-
Dylan Patel
(03:12:37)
It’s like five cents or something like that now, right? Versus $60, 1200X, that’s not the exact numbers, but it’s 1200X, I remember that number, is humongous cost per intelligence. Now, the freak out over DeepSeek is, “Oh my god, they made it so cheap.” It’s like actually, if you look at this trend line, they’re not below the trend line first of all, at least for GPT-3, right? They are the first to hit it, which is a big deal, but they’re not below the trend line as far as GPT-3. Now we have GPT-4, what’s going to happen with these reasoning capabilities? It’s a mix of architectural innovations, it’s a mix of better data, and it’s going to be better training techniques and all of these better inference systems, better hardware going from each generation of GPU to new generations or ASICs.

(03:13:22)
Everything is going to take this cost curve down and down and down and down. And then can I just spawn a thousand different LLMs to create a task and then pick from one of them? Or whatever search technique, I want, a Tree, Monte Carlo Tree Search, maybe it gets that complicated, maybe it doesn’t because it’s too complicated to actually scale. Who knows? Better lesson, right?

(03:13:43)
The question is, I think, when not if, because the rate of progress is so fast. Nine months ago, Dario said nine months ago the cost to train an inference was this, and now we’re much better than this and DeepSeek is much better than this. And that cost curve for GPT-4, which was also roughly $60 per million tokens when it launched, has already fallen to $2 or so. And we’re going to get it down to cents probably for GPT-4 quality. And then that’s the base for the reasoning models like o1 that we have today and o1-pro is spawning multiple and o3 and so on and so forth, these search techniques, too expensive today, but they will get cheaper and that’s what’s going to unlock the intelligence.

NVIDIA

Lex Fridman
(03:14:31)
So, it’ll get cheaper and cheaper and cheaper. The big DeepSeek R1 release freaked everybody out because of the cheaper. One of the manifestations of that is NVIDIA stock plummeted. Can you explain what happened? And also just explain this moment and if NVIDIA is going to keep winning.
Nathan Lambert
(03:14:52)
We are both NVIDIA bulls here, I would say. And in some ways, the market response is reasonable. NVIDIA’s biggest customers in the US are major tech companies and they’re spending a ton on AI. And if a simple interpretation of DeepSeek is you can get really good models without spending as much on AI. So in that capacity it’s like, “Oh, maybe these big tech companies won’t need to spend as much in AI and go down.”

(03:15:18)
The actual thing that happened is much more complex where there’s social factors, where there’s the rising in the app store, the social contagion that is happening. And then I think some of it is just like, I don’t trade, I don’t know anything about financial markets, but it builds up over the weekend, the social pressure, where it’s like if it was during the week and there was multiple days of trading when this was really becoming, but it comes on the weekend and then everybody wants to sell, and then that is a social contagion.
Dylan Patel
(03:15:43)
I think, and there were a lot of false narratives, which is like, “Hey, these guys are spending billions on models,” and they’re not spending billions on models. No one spent more than a billion dollars on a model that’s released publicly. GPT-4 was a couple hundred million and then they’ve reduced the cost with 4o, 4 Turbo, 4o, right? But billion dollar model runs are coming and this concludes pre-training and post-training, right? And then the other number is like, “Hey, DeepSeek didn’t include everything.” They didn’t include a lot of the cost goes to research and all this sort of stuff. A lot of the cost goes to inference. A lot of the cost goes to post-training. None of these things were factored. Research, salaries, all these things are counted in the “billions of dollars” that OpenAI is spending, but they weren’t counted in the, “Hey, $6 million, $5 million that DeepSeek spent.”

(03:16:27)
So, there’s a bit of misunderstanding of what these numbers are, and then there’s also an element of… NVIDIA has just been a straight line up and there’s been so many different narratives that have been trying to push down NVIDIA. I don’t say push down NVIDIA stock. Everyone is looking for a reason to sell or to be worried. It was Blackwell delays, right? Their GPU, every two weeks there’s a new report about their GPUs being delayed. There’s the whole thing about scaling laws ending, right? It’s so ironic-
Nathan Lambert
(03:16:57)
It lasted a month.
Dylan Patel
(03:17:00)
It was literally just, “Hey, models aren’t getting better.” They’re just not getting better. There’s no reason to spend more, pre-training scaling is dead. And then it’s like o1, o3, right?
Nathan Lambert
(03:17:10)
R1.
Dylan Patel
(03:17:11)
R1, right? And now it’s like, “Wait, models, they’re progressing too fast. Slow down the progress, stop spending on GPUs.” But the funniest thing I think that comes out of this is Jevons paradox is true. AWS pricing for H100s has gone up over the last couple of weeks, since a little bit after Christmas, since V3 was launched, AWS H100 pricing has gone up. H200s are almost out of stock everywhere because H200 has more memory and therefore R1 wants that chip over H100, right?
Nathan Lambert
(03:17:43)
We were trying to get GPUs on a short notice this week for a demo and it wasn’t that easy. We were trying to get just 16 or 32 H100s for demo and it was not very easy.
Lex Fridman
(03:17:51)
So for people who don’t know, Jevons paradox is when the efficiency goes up, somehow magically, counter intuitively, the total resource consumption goes up as well.
Dylan Patel
(03:18:03)
And semiconductors is 50 years of Moore’s law, every two years half the cost, double the transistors, just like clockwork and it’s slowed down obviously, but the semiconductor industry has gone up the whole time. It’s been wavy, right? There’s obviously cycles and stuff and I don’t expect AI to be any different. There’s going to be ebbs and flows, but in AI, it’s just playing out at an insane timescale. It was 2X every two years, this is 1200X in like three years. So it’s like the scale of improvement is hard to wrap your head around.
Lex Fridman
(03:18:34)
Yeah. I was confused because to me, NVIDIA stock on that should have gone up, but maybe it went down because there’s suspicion of foul play on the side of China, something like this. But if you just look purely at the actual principles at play here, it’s obvious. Yeah, the Jevons paradox-

GPU smuggling

Nathan Lambert
(03:18:53)
The more progress that AI makes or the higher the derivative of AI progress is, especially because NVIDIA’s in the best place, the higher the derivative is, the sooner the market’s going to be bigger and expanding and NVIDIA’s the only one that does everything reliably right now.
Lex Fridman
(03:19:07)
Yeah, because it’s not like an NVIDIA competitor arose. It’s another company that’s using NVIDIA-
Nathan Lambert
(03:19:14)
Who historically has been a large NVIDIA customer.
Dylan Patel
(03:19:18)
And has press releases about them cheering about being China’s biggest NVIDIA customer, right?
Lex Fridman
(03:19:23)
Yeah. I mean-
Dylan Patel
(03:19:25)
Obviously they’ve quieted down, but I think that’s another element of it is that they don’t want to say how many GPUs they have because hey, yes, they have H800s, yes, they have H20s, they also have some H100s, right? Which were smuggled in.
Lex Fridman
(03:19:37)
Can you speak to that, to the smuggling? What’s the scale of smuggling that’s feasible for a nation state to do for companies? Is it possible to-
Dylan Patel
(03:19:47)
I think there’s a few angles of “smuggling” here, right? One is ByteDance, arguably is the largest smuggler of GPUs for China. China’s not supposed to have GPUs. ByteDance has over 500,000 GPUs. Why? Because they’re all rented from companies around the world. They rent from Oracle, they rent from Google, they rent from all these, and a bunch of smaller cloud companies too, right? All the “neoClouds” of the world. They rent so, so many GPUs. They also buy a bunch. And they do this for mostly what Meta does, right? Serving TikTok, right? Serving next best-
Nathan Lambert
(03:20:17)
Separate discussion.
Dylan Patel
(03:20:18)
Same as Meta, right? To be clear, today, that’s the use, right? And it’s a valid use, right? Hack the dopamine circuit. Now, that’s theoretically now very much restricted with the AI diffusion rules, which happened in the last week of the Biden admin, and Trump admin looks like they’re going to keep them, which limits allies even, like Singapore, which Singapore is 20%, 30% of NVIDIA’s revenue, but Singapore’s had a memoratorium on not building data centers for 15 years because they don’t have enough power. So, where are they going?
Nathan Lambert
(03:20:50)
Oh my God.
Dylan Patel
(03:20:51)
I’m not claiming they’re all going to China, but a portion, many are going to Malaysia, including Microsoft and Oracle have big data centers in Malaysia. They’re going all over Southeast Asia probably, India as well. There’s stuff routing, but the diffusion rules are very de facto, like you can only buy this many GPUs from this country and you can only rent a cluster this large to companies that are Chinese. They’re very explicit on trying to stop smuggling.

(03:21:15)
And a big chunk of it was, hey, random company buys 16 servers, ships them to China. There’s actually, I saw a photo from someone in the semiconductor industry who leads a team for networking chips that competes with NVIDIA, and he sent a photo of a guy checking into a first class United flight from San Francisco to Shanghai or Shenzhen with a super micro box that was this big, which can only contain GPUs, right? And he was booking first class because think about it, 3K to 5K for your first class ticket, server costs $240,000 in the US, $250,000, you sell it for $300,000 in China. Wait, you just got a free first class ticket and a lot more money. So it’s like… And that’s small scale smuggling. Most of the large scale smuggling is companies in Singapore and Malaysia routing them around or renting GPUs, completely legally-
Nathan Lambert
(03:22:10)
I want to jump in. How much does this scale? I think there’s been some people that are higher level economics understanding say that as you go from 1 billion of smuggling to 10 billion, it’s like you’re hiding certain levels of economic activity and that’s the most reasonable thing to me is that there’s going to be some level where it’s so obvious that it’s easier to find this economic activity. And-
Dylan Patel
(03:22:30)
Yeah. So, my belief is that last year roughly, so NVIDIA made a million H20s, which are legally allowed to be shipped to China, which we talked about is better for reasoning, inference at least, not training, but reasoning inference and inference generally. Then they also had a couple hundred thousand, we think like 200,000 to 300,000 GPUs were routed to China from Singapore, Malaysia, US, wherever. Companies spawn up, buy 16 GPUs, 64 GPUs, whatever it is, route it, and Huawei is known for having spent up a massive network of companies to get the materials they need after they were banned in 2018. So, it’s not otherworldly, but I agree, right? Nathan’s point is like, hey, you can’t smuggle $10 billion of GPUs.

(03:23:13)
And then the third source, which is just now banned, which wasn’t considered smuggling, but is China is renting, I believe from our research, Oracle’s biggest GPU customer is ByteDance. And for Google, I think it’s their second-biggest customer. And you go down the list of clouds and especially these smaller cloud companies that aren’t the “hyperscalers,” think beyond CoreWeave and Lambda even, there’s 60 different new cloud companies serving NVIDIA GPUs. I think ByteDance is renting a lot of these, all over it, right?

(03:23:44)
And so these companies are renting GPUs to Chinese companies, and that was completely legal up until the diffusion rules, which happened just a few weeks ago. And even now, you can rent GPU clusters that are less than 2,000 GPUs, or you can buy GPUs and ship them wherever you want if there are less than 1,500 GPUs. There are still some ways to smuggle, but yeah, as the numbers grow a hundred something billion dollars of revenue for NVIDIA last year, 200 something billion this year, and if next year, it could nearly double again or more than double based on what we see with data center footprints being built out all across the US and the rest of the world, it’s going to be really hard for China to keep up with these rules.

(03:24:28)
Yes, there will always be smuggling and DeepSeek level models, GPT-4 level models, o1 level models capable to train on what China can get, even the next tier above that. But if we speed run a couple more jumps to billion dollar models, $10 billion models, then it becomes, “Hey, there is a compute disadvantage for China for training models and serving them.” And the serving part is really critical, right? DeepSeek cannot serve their model today. It’s completely out of inventory. It’s already started falling in the app store actually, downloads, because you download it, you try and sign up, they say, “We’re not taking registrations,” because they have no capacity. You open it up, you get less than five tokens per second, if you even get your request approved, right? Because there’s just no capacity because they just don’t have enough GPUs to serve the model, even though it’s incredibly efficient.
Lex Fridman
(03:25:14)
It’d be fascinating to watch the smuggling. Because I mean there’s drug smuggling, right? That’s a market. There’s weapons smuggling. And GPUs will surpass that at some point.
Nathan Lambert
(03:25:25)
Chips are highest value per kilogram probably by far. I have another question for you, Dylan. Do you track model API access internationally? How easy is it for Chinese companies to use hosted model APIs from the US?

DeepSeek training on OpenAI data

Dylan Patel
(03:25:42)
Yeah. I mean that’s incredibly easy, right? OpenAI publicly stated DeepSeek uses their API and they say they have evidence, right? And this is another element of the training regime, is people at OpenAI have claimed that it’s a distilled model, i.e., you’re taking OpenAI’s model, you’re generating a lot of output, and then you’re training on the output in their model. And even if that’s the case, what they did is still amazing by the way, what DeepSeek did, efficiency-wise.
Nathan Lambert
(03:26:04)
Distillation is standard practice in industry. Whether or not, if you’re at a closed lab where you care about terms of service and IP closely, you distill from your own models. If you are a researcher and you’re not building any products, you distill from the OpenAI models-
Lex Fridman
(03:26:16)
This is a good opportunity. Can you explain big picture distillation as a process? What is distillation? What’s the process of distillation?
Nathan Lambert
(03:26:24)
We’ve talked a lot about training language models. They are trained on text and post-training, you’re trying to train on very high-quality texts that you want the model to match the features of, or if you’re using RL, you’re letting the model find its own thing. But for supervised fine-tuning, for preference data, you need to have some completions, what the model is trying to learn to imitate. And what you do there is instead of a human data or instead of the model you’re currently training, you take completions from a different, normally more powerful, model. I think there’s rumors that these big models that people are waiting for, these GPT-5s of the world, the Claude 3 Opuses of the world are used internally to do this distillation process at OpenAI-
Dylan Patel
(03:27:04)
There’s also public examples, right? Like Meta explicitly stated, not necessarily distilling, but they used 405B as a reward model for 70B in their Llama 3.2 or 3.3 rule-
Nathan Lambert
(03:27:15)
Yes. This is all the same topic.
Lex Fridman
(03:27:16)
So, is this ethical? Is this legal? Why is that Financial Times article headline say, “OpenAI says that there’s evidence that China’s DeepSeek used its model to train competitor.”
Nathan Lambert
(03:27:31)
This is a long, at least in the academic side and research side, it has a long history because you’re trying to interpret OpenAI’s rule. OpenAI’s terms of service say that you cannot build a competitor with outputs from their models. Terms of service are different than a license, which are essentially a contract between organizations. So if you have a terms of service on OpenAI’s account, if I violate it, OpenAI can cancel my account. This is very different than a license that says how you could use a downstream artifact. So a lot of it hinges on a word that is very unclear in the AI space, which is, what is a competitor?
Dylan Patel
(03:28:02)
And then the ethical aspect of it is like, why is it unethical for me to train on your model when you can train on the internet’s text? Right?
Lex Fridman
(03:28:10)
So there’s a bit of a hypocrisy because OpenAI and potentially most of the companies trained on the internet’s text without permission.
Nathan Lambert
(03:28:20)
There’s also a clear loophole, which is that I generate data from OpenAI and then I upload it somewhere and then somebody else trains on it and the link has been broken. They’re not under the same terms of service contract.
Dylan Patel
(03:28:33)
This is why-
Nathan Lambert
(03:28:33)
There’s a lot of… There’s a lot of to be discovered details that don’t make a lot of sense.
Dylan Patel
(03:28:38)
This is why a lot of models today, even if they train on zero OpenAI data, you ask the model, “Who trained you?” It’ll say, “I’m ChatGPT trained by OpenAI,” because there’s so much copy paste of OpenAI outputs from that on the internet that you just weren’t able to filter it out and there was nothing in the RL where they implemented or post-training or SFT, whatever, that says, “Hey, I’m actually a model by Allen Institute instead of OpenAI.”
Nathan Lambert
(03:29:03)
We have to do this if we serve a demo. We do research and we use OpenAI APIs because it’s useful and we want to understand post-training and our research models, they all say they’re written by OpenAI unless we put in the system prop that we talked about that, “I am Tülu. I am a language model trained by the Allen Institute for AI.” And if you ask more people around industry, especially with post-training, it’s a very doable task to make the model say who it is or to suppress the OpenAI thing. So in some levels, it might be that DeepSeek didn’t care that it was saying that it was by OpenAI. If you’re going to upload model weights, it doesn’t really matter because anyone that’s serving it in an application and cares a lot about serving is going to, when serving it, if they’re using it for a specific task, they’re going to tailor it to that and it doesn’t matter that it’s saying it’s ChatGPT.
Lex Fridman
(03:29:49)
Oh, I guess one of the ways to do that is like a system prompt or something like that? If you’re serving it to say that you’re-
Nathan Lambert
(03:29:55)
That’s what we do. If we host a demo, you say, “You are Tülu 3, a language model trained by the Allen Institute for AI.” We also are benefited…
Nathan Lambert
(03:30:00)
… model trained by the Allen Institute for AI. We also are benefited from OpenAI data because it’s a great research tool.
Lex Fridman
(03:30:06)
Do you think there’s any truth and value to the OpenAI’s claim that there’s evidence that China’s DeepSeek used this model to train?
Dylan Patel
(03:30:16)
I think everyone has benefited regardless because the data’s on the internet. And therefore, it’s in your per training now. There are subreddits where people share the best ChatGPT outputs, and those are in your model-
Nathan Lambert
(03:30:29)
I think that they’re trying to shift the narrative. They’re trying to protect themselves. We saw this years ago when ByteDance was actually banned from some OpenAI APIs for training on outputs. There’s other AI startups that most people, if you’re in the AI culture, were like they just told us they trained on OpenAI outputs and they never got banned. That’s how they bootstrapped their early models.

(03:30:51)
So, it’s much easier to get off the ground using this than to set up human pipelines and build a strong model. So there’s long history here, and a lot of the communications are seem like narrative [inaudible 03:31:00].
Dylan Patel
(03:31:00)
Actually, over the last couple of days, we’ve seen a lot of people distill DeepSeek’s model into Llama models, because the DeepSeek models are complicated to run inference on because they’re mixture of experts and they’re 600 plus billion parameters and all of this. And people distilled them into the Llama models because the Llama models are so easy to serve, and everyone’s built the pipelines and tooling for inference with the Llama models because it’s the open standard.

(03:31:24)
So, we’ve seen a sort of roundabout. Is it bad? Is it illegal? Maybe it’s illegal, whatever. I don’t know about that, but-
Nathan Lambert
(03:31:30)
It could break contracts. I don’t think it’s illegal in any legal… No one’s going to jail for this, ever.
Lex Fridman
(03:31:36)
Fundamentally, I think it’s ethical, or I hope it’s ethical because the moment it becomes… We ban that kind of thing, it’s going to make everybody much worse off. And I also, actually…

(03:31:50)
This is difficult, but I think you should be allowed to train on the internet. I know a lot of authors and creators are very sensitive about it. That’s a difficult question. But the moment you’re not allowed to train on the internet-
Nathan Lambert
(03:32:03)
I agree.
Dylan Patel
(03:32:03)
I have a schizo take on how you can solve this. Because it already works.
Lex Fridman
(03:32:04)
All right.
Nathan Lambert
(03:32:07)
I have a reasonable take out of it.
Lex Fridman
(03:32:09)
All right, [inaudible 03:32:10].
Dylan Patel
(03:32:10)
So, Japan has a law which you’re allowed to train on any training data and copyrights don’t apply if you want to train a model, A. B, Japan has 9 gigawatts of curtailed nuclear power. C, Japan is allowed under the AI diffusion rule to import as many GPUs as they’d like. So, all we have to do…

(03:32:29)
We have a market here to make. We build massive data centers, we rent them to the labs, and then we train models in a legally permissible way, and there’s no ifs, ands, or buts. And now, the models have no potential copyright lawsuit from New York Times or anything like that. No, it’s just completely legal.
Nathan Lambert
(03:32:46)
Now, so-
Lex Fridman
(03:32:47)
Genius.
Nathan Lambert
(03:32:47)
… the early copyright lawsuits have fallen in the favor of AI training. I would say that the long tail of use is going to go inside of AI, which is if you scrape trillions of tokens of data, you’re not looking and saying, “This one New York Times article is so important to me.” But if you’re doing a audio generation for music or image generation, and you say, “Make it in the style of X person,” that’s a reasonable case where you could figure out what is their profit margin on inference. I don’t know if it’s going to be the 50/50 of YouTube Creator Program or something, but I would opt into that program as a writer, please.

(03:33:28)
It’s going to be a rough journey, but there will be some solutions like that that makes sense. But there’s a long tail where it’s just on the internet.
Lex Fridman
(03:33:35)
I think one of the other aspects of that Financial Times article implied, and so that leads to a more general question. Do you think there’s… How difficult is spying, espionage, and stealing of actual secret code and data from inside of companies? How much of that is being attempted?
Nathan Lambert
(03:33:55)
Code and data is hard, but ideas is easy. Silicon Valley operates on the way that top employees get bought out by other companies for a pay raise, and a large reason why these companies do this is to bring ideas with them. And there’s no… I mean, in California, there’s rules that certain non-competes or whatever are illegal in California. And whether or not there’s NDAs and things, that is how a lot of it happens. Recently, there was somebody from Gemini who helped make this 1 million context length. And everyone is saying the next Llama who, he went to the Meta team, is going to have 1 million context length. And that’s kind of how the world works.
Dylan Patel
(03:34:34)
As far as industrial espionage and things, that has been greatly successful in the past. The Americans did it to the Brits, the Chinese have done it to the Americans, and so on and so forth. It is a fact of life. And so, to argue industrial espionage can be stopped is probably unlikely. You can make it difficult. But even then, there’s all these stories about like, “Hey, F35 and F22 have already been given to China in terms of design plans and stuff.”

(03:35:02)
Code and stuff between, I say companies, not nation states, is probably very difficult. But ideas are discussed a lot, whether it be a house party in San Francisco or a company changing employees or always the mythical honeypot that always gets talked about. Someone gets honeypotted because everyone working on AI is a single dude who’s in their 20s and 30s. Not everyone, but insane amount of… Insane percentages. So, there’s always all these… And obviously-
Lex Fridman
(03:35:34)
So, honeypotted is like a female spy approaches you and…
Dylan Patel
(03:35:38)
Yeah. Or male, right? It’s San Francisco. But as a single dude, I will say in his late 20s, we are very easily corrupted. Not corrupted myself, but we are. Right?
Lex Fridman
(03:35:51)
Yeah. Everybody else. Not me.
Nathan Lambert
(03:35:54)
I’m too oblivious that I am not single, so I’m safe from one espionage access.

AI megaclusters

Lex Fridman
(03:36:00)
Yeah. You have to make sure to close all security vulnerabilities. So you, Dylan, collect a lot of information about each of the mega clusters for each of the major AI companies. Can you talk about the buildouts for each one that stand out?
Dylan Patel
(03:36:18)
Yeah. I think the thing that’s really important about these mega cluster buildouts is they’re completely unprecedented in scale. US data center power consumption has been slowly on the rise and it’s gone up to 2, 3% even through the cloud computing revolution. Data center consumption has a percentage of total US, and that’s been over decades of data centers, etc. It’s been climbing slowly, but now, 2 to 3%.

(03:36:43)
Now, by the end of this decade, it’s… Even under… When I say 10%, a lot of people that are traditionally by 2028 to 2030, people traditionally non-traditional data center people, that’s nuts. But then, people who are in AI who have really looked at this like the Anthropics and OpenAI’s, are like, “That’s not enough.”

(03:37:02)
And I’m like, “Okay.” But this is both through globally distributed or distributed throughout the US as well as centralized clusters. The distributed throughout the US is exciting and it’s the bulk of it. Like, hey, OpenAI or, say, Meta’s adding a gigawatt, but most of it is distributed through the US for inference and all these other things.
Lex Fridman
(03:37:26)
So maybe, we should lay out what a cluster is. So, does this include AWS? Maybe, it’s good to talk about the different kinds of clusters. What you mean by mega clusters? What’s the GPU and what’s a compute or… And what [inaudible 03:37:41]-
Dylan Patel
(03:37:40)
Yeah.
Lex Fridman
(03:37:41)
Not that far back, but yeah. So, what do we mean by the clusters? The buildouts?
Dylan Patel
(03:37:45)
Oh, man. I thought I was about to do the Apple ad, what’s a computer? So traditionally, data centers and data center tasks have been a distributed systems problem that is capable of being spread very far and widely. I.e, I send a request to Google, it gets routed to a data center somewhat close to me, it does whatever search ranking recommendation, sends a result back. The nature of the task is changing rapidly in that the task, there’s two tasks that people are really focused on now. It’s not database access. It’s not, “Serve me the right page, serve me the right ad.”

(03:38:20)
It’s now, a inference. An inference is dramatically different from traditional distributed systems, but it looks a lot more simple, similar. And then, there’s training. The inference side is still like, “Hey, I’m going to put thousands of GPUs in blocks all around these data centers.” I’m going to run models on them. User submits a request, it gets kicked off. Or hey, my service. They submit a request to my service. They’re on Word and they’re like, “Oh yeah, help me, Copilot,” and it kicks it off. Or I’m on my windows, Copilot, whatever, Apple intelligence. Whatever it is, it gets kicked off to a data center. That data center does some work and sends it back. That’s inference. That is going to be the bulk of compute, but then…

(03:38:59)
And that’s like, there’s thousands of data centers that we’re tracking with satellites and all these other things, and those are the bulk of what’s being built. But the scale of… And so, that’s what’s really reshaping and that’s what’s getting millions of GPUs. But the scale of the largest cluster is also really important. When we look back at history or through the age of AI, it was a really big deal when they did AlexNet on, I think, 2 GPUs or 4 GPUs. I don’t remember. It’s a really big deal.
Nathan Lambert
(03:39:30)
It’s a big deal because you use GPUs.
Dylan Patel
(03:39:31)
It’s a big deal that they use GPUs and they use multiple. But then over time, its scale has just been compounding. And so when you skip forward to GPT-3, then GPT-4, GPT-4 20,000 A100 GPUs. Unprecedented run in terms of the size and the cost, right? A couple of hundred million dollars on a YOLO run for GPT-4, and it yielded this magical improvement that was perfectly in line with what was experimented and just a log scale right up.
Nathan Lambert
(03:39:58)
Oh yeah, they had that plot from the paper.
Dylan Patel
(03:40:00)
The scaling of the technical part. The scaling laws were perfect, right? But that’s not a crazy number. 20,000 A100’s, roughly, each GPU is consuming 400 watts. And then when you add in the whole server, everything, it’s like 15 to 20 megawatts of power. Maybe, you could look up what the power of consumption of a person is because the numbers are going to get silly, but 15 to 20 megawatts was standard data center size. It was just unprecedented that was all GPUs running one task.
Nathan Lambert
(03:40:00)
How many watts is a toaster?
Dylan Patel
(03:40:29)
A toaster has also-
Nathan Lambert
(03:40:29)
That’s a good example.
Dylan Patel
(03:40:32)
… a similar power consumption to an A100. H100 comes around. They increase the power from 400 to 700 watts and that’s just per GPU, and then there’s all the associated stuff around it. So once you count all of that, it’s roughly 1,200 to 1,400 watts for everything. Networking, CPUs, memory, blah, blah, blah.
Lex Fridman
(03:40:48)
So we should also say, what’s required, you said power. So, a lot of power is required. A lot of heat is generated, so the cooling is required. And because there’s a lot of GPUs or CPUs or whatever, they have to be connected. So, there’s a lot of networking, right?
Dylan Patel
(03:41:06)
Yeah. I think-

(03:41:07)
Sorry for skipping past that. And then the data center itself is complicated, but these are still standard sized data centers for GPT-4 scale. Now, we step forward to what is the scale of clusters that people built last year, and it ranges widely. It ranges from like, “Hey, these are standard data centers. And we’re just using multiple of them and connecting them together really with a ton of fiber between them, a lot of networking, etc.” That’s what OpenAI and Microsoft did in Arizona. They have 100,000 GPUs.

(03:41:37)
Meta, similar thing. They took their standard existing data center design and it looks like an H, and they connected multiple of them together. They first did 24,000 GPUs total, only 16,000 of them were running on the training run because GPUs are very unreliable so they need to have spares to swap in and out. All the way to now, 100,000 GPUs that they’re training on Llama 4 on currently. Like, 128,000 or so.

(03:42:02)
Think about 100,000 GPUs with roughly 1,400 watts apiece. That’s 140 megawatts, 150 megawatts for 128. So, you’re talking about you’ve jumped from 15 to 20 megawatts to almost 10x that number, 9x that number, to 150 megawatts in two years from 2022 to 2024. And some people like Elon, that he admittedly… He says himself he got into the game a little bit late for pre-training large language models. xAI was started later, right? But then, he bent heaven and hell to get his data center up and get the largest cluster in the world, which is 200,000 GPUs. And he did that. He bought a factory in Memphis. He’s upgrading the substation, with the same time, he’s got a bunch of mobile power generation, a bunch of single cycle combine. He tapped the natural gas line that’s right next to the factory, and he’s just pulling a ton of gas, burning gas.

(03:42:55)
He’s generating all this power. He’s in an old appliance factory that’s shut down and moved to China long ago, and he’s got 200,000 GPUs in it. And now, what’s the next scale? All the hyperscalers have done this. Now, the next scale is something that’s even bigger. And so Elon, just to stick on the topic, he’s building his own natural gas plant, like a proper one right next door. He’s deploying tons of Tesla Megapack batteries to make the power more smooth and all sorts of other things. He’s got industrial chillers to cool the water down because he’s water-cooling the chips. So, all these crazy things to get the clusters bigger and bigger.

(03:43:34)
But when you look at, say, what OpenAI did with Stargate in Arizona, in Abilene Texas, right? What they’ve announced, at least. It’s not built. Elon says they don’t have the money. There’s some debates about this. But at full scale, at least the first section is definitely money’s accounted for, but there’s multiple sections. But full scale, that data center is going to be 2.2 gigawatts, 2,200 megawatts of power in. And roughly, 1.8 gigawatts or 1,800 megawatts of power delivered to chips.

(03:44:07)
Now, this is an absurd scale. 2.2 gigawatts is more than most cities, to be clear. Delivered to a single cluster that’s connected to do training. To train these models, to do both the pre-training, the post-training, all of this stuff.
Lex Fridman
(03:44:22)
This is insane.
Nathan Lambert
(03:44:23)
It is. What is a nuclear power plant, again?
Dylan Patel
(03:44:25)
Everyone is doing this. Meta in Louisiana, they’re building two natural gas plants. Massive ones. And then, they’re building this massive data center. Amazon has plans for this scale. Google has plans for this scale. xAI has plans for this scale. All of these, the guys that are racing, the companies that are racing are racing hard, and they’re doing multi-gigawatt data centers to build this out. Because they think that, “If I now have…” Obviously, pre-training scaling is going to continue, but to some extent. But then also, all this post-training stuff where you have RL Sandbox for computer use or whatever, this is where they’re going to… And all these fearful viable domains where they just keep learning and learning and learning, self-play or whatever. Whatever it is makes the AI so much more capable because the line does go up.

(03:45:14)
As you throw more compute, you get more performance. This shirt is about scaling laws. To some extent, it is diminishing returns. You 10x the compute, you don’t get 10x better model. You get a diminishing returns. But also, you get efficiency improvements, so you bend the curve. And these scale of data centers are just reeking a lot of havoc on the network. Nathan was mentioning Amazon has tried to buy this nuclear power plant Talen. And if you look at Talen’s stock, it’s just skyrocketing. They’re building a massive multi-gigawatt data center there.

(03:45:47)
You just go down the list, there’s so many ramifications. Interesting thing is certain regions of the US transmitting power cost more than actually generating it because the grid is so slow to build. And the demand for power, and the ability to build power, and re-ramping on a natural gas plant or even a coal plant is easy enough to do, but transmitting the power’s really hard. So in some parts of the US like in Virginia, it costs more to transmit power than it costs to generate it, which is there’s all sorts of second-order effects that are insane here.
Lex Fridman
(03:46:16)
Can the power grid support this kind of growth?
Dylan Patel
(03:46:19)
Trump’s executive orders… There was a Biden executive order before the end of the year, but then Trump had some more executive orders, which hopefully reduced the regulations to where, yes, things can be built. But yeah, this is a big, big challenge. Is building enough power fast enough?
Lex Fridman
(03:46:33)
Are you going to basically have a nuclear power plant next to a data center for each one of these?
Dylan Patel
(03:46:39)
The fun thing here is this is too slow to build the power plant. To build a power plant or to reconfigure an existing power plant, it’s too slow. And so therefore, you must use…

(03:46:51)
Data center power consumption is flat, right? I mean, [inaudible 03:46:53].
Nathan Lambert
(03:46:53)
This is why nuclear is also good for it. Long term, nuclear is a very natural fit, but…
Dylan Patel
(03:46:58)
Yes.
Nathan Lambert
(03:46:58)
… data-

(03:46:59)
You can’t do solar or anything in the short term like that.
Dylan Patel
(03:47:03)
Because data center power’s like this, right? You’re telling me I’m going to buy tens of billions of dollars of GPUs and idle them because the power’s not being generated? Power’s cheap. If you look at the cost of a cluster, less than 20% of it is power. Most of it is the capital cost and depreciation of the GPUs. And so it’s like, “Well, screw it. I’ll just build natural gas plants.” This is what Meta is doing in Louisiana, this is what OpenAI is doing in Texas, and all these different places. They may not be doing it directly, but they are partnered with someone. And so, there is a couple of hopes.

(03:47:34)
One is… And Elon, what he’s doing in Memphis is to the extreme. They’re not just using dual combine cycle gas which is super efficient, he’s also just using single cycle and mobile generators and stuff which is less efficient. But there’s also the flip side, which is solar power generation is like this, and wind is another like this. Different correlate different. So if you stack both of those, plus you get a big chunk of batteries, plus you have a little bit of gas, it is possible to run it more green. It’s just the time scales for that is slow. So, people are trying. But Meta basically said, “Whatever. I don’t care about my sustainability pledge.” Or they’ll buy a power… It’s called a PPA, Power Purchasing Agreement, where there’ll be a massive wind farm or solar farm wherever. And then, they’ll just pretend like those electrons are being consumed by the data center. But in reality, they’re paying for the power here and selling it to the grid, and they’re buying power here.

(03:48:26)
And then another thing is Microsoft quit on some of their sustainability pledges. Elon, what he did with Memphis is objectively somewhat dirty, but he is also doing it in an area where there’s a bigger natural gas plant right next door and a sewer next… Or not a sewer, but a wastewater treatment and a garbage dump nearby. And he’s obviously made the world a lot more clean than that one data center is going to do, so I think it’s fine to some extent. And maybe, AGI solves global warming and stuff, whatever it is.

(03:48:55)
This is the attitude that people at the labs have, which is like, “Yeah, it’s great. We’ll just use gas,” because the race is that important. And if we lose, that’s way worse.
Lex Fridman
(03:49:05)
I should say that I got a chance to visit the Memphis data center.
Dylan Patel
(03:49:08)
Oh, wow.
Lex Fridman
(03:49:10)
And it’s incredible. I mean, I visited with Elon. Just the teams and the rate of innovation there is insane. My sense is that nobody’s ever done anything of this scale, and nobody has certainly ever done anything of this scale at the rate that xAI is doing. So, they’re figuring out…

(03:49:31)
I was sitting in on all of these meetings where they’re brainstorming. It’s insane. It’s exciting because they’re trying to figure out what the bottlenecks are, how to remove the bottlenecks, how to make sure that… There’s just so many really cool things about putting together a data center because everything has to work. The people that do the sys admin, the machine learning and all of that is the exciting thing, so on. But really, the people that run everything are the folks that know the low-level software and hardware that runs everything, the networking, all of that. So, you have to make sure you have procedures that test everything. I think they’re using ethernet. I don’t know how they’re doing the networking, but-
Dylan Patel
(03:50:15)
They’re using NVIDIA Spectrum-X Ethernet. I think the unsung heroes are the cooling in electrical systems which are just glossed over.
Lex Fridman
(03:50:23)
Yeah, exactly.
Dylan Patel
(03:50:25)
But I think one story that maybe exemplifies how insane this stuff is, is when you’re training, you’re always doing… You’re running through the model a bunch, in the most simplistic terms. Running through the model a bunch, and then you’re going to exchange everything and synchronize the weights. So, you’ll do a step. This is like a step-in model training. And every step, your loss goes down hopefully, and it doesn’t always.

(03:50:48)
But in the simplest terms, you’ll be computing a lot and then you’ll exchange. The interesting thing is GPU power is most of it, networking power is some but it’s a lot less. So while you’re computing, your power for your GPUs is here. But then when you’re exchanging weights, if you’re not able to overlap communications and compute perfectly, there may be a time period where your GPUs are just idle, and you’re exchanging weights and you’re like, “Hey, the model’s updating.” So, you’re exchanging the radiance, you do the model update, and then you start training again. So, the power goes… Right? And it’s super spiky.

(03:51:17)
And so funnily enough, when you talk about the scale of data center power, you can blow stuff up so easily. And so, Meta actually has accidentally upstreamed something to code in PyTorch where they added an operator. And I kid you not, whoever made this, I want to hug the guy because it says PyTorch… It’s like PyTorch.powerplant no blow up equals 0 or equal 1. And what it does is amazing, right?
Lex Fridman
(03:51:44)
Yeah.
Dylan Patel
(03:51:44)
Either when you’re exchanging the weights, the GPU will just compute fake numbers so the power doesn’t spike too much, and so then the power plants don’t blow up because the transient spikes screw stuff up.
Lex Fridman
(03:51:54)
Well, that makes sense. You have to do that kind thing. [inaudible 03:51:57] You have to make sure they’re not idle.
Dylan Patel
(03:51:59)
And Elon’s solution was like, “Let me throw a bunch of Tesla Megapacks and a few other things.”
Lex Fridman
(03:52:03)
Yeah, to symbolize that.
Dylan Patel
(03:52:03)
Everyone has different solutions, but Meta’s, at least, was publicly and openly known, which is just like, set this operator. And what this operator does is it just makes the GPUs compute nothing so that the power doesn’t spike.
Lex Fridman
(03:52:14)
But that just tells you how much power you’re working with. I mean, it’s insane. It’s insane.
Nathan Lambert
(03:52:18)
People should just go to Google, like scale or what does X watts do, and go through all the scales from 1 watt to a kilowatt to a megawatt. You look and stare at that, and you’re how high on the list a gigawatt is, it’s mind-blowing.
Lex Fridman
(03:52:34)
Can you say something about the cooling? I know Elon’s using liquid cooling, I believe, in all cases. That’s a new thing. Most of them don’t use liquid cooling. Is there something interesting to say about the cooling?
Dylan Patel
(03:52:46)
Yeah. So, air cooling has been the de facto standard. Throw a bunch of metal heat pipes, et cetera, and fans, and that’s cold. That’s been enough to cool it. People have been dabbling in water cooling. Google’s TPUs are water- cooled. So, they’ve been doing that for a few years. But with GPUs, no one’s ever done… And no one’s ever done the scale of water cooling that Elon just did. Now, next generation NVIDIA is for the highest-end GPU, it is mandatory water cooling. You have to water-cool it.

(03:53:16)
But Elon did it on this current generation, and that required a lot of stuff. If you look at some of the satellite photos and stuff of the Memphis facility, there’s all these external water chillers that are sitting. Basically, it looks like a semi truck pod thing. What’s it called? The container? But really, those are water chillers, and he has 90 of those water chillers just sitting outside. Ninety different containers that chill the water, bring it back to the data center, and then you distribute it to all the chips, pull all the heat out and then send it back. And this is both a way to cool the chips, but also, it’s an efficiency thing.

(03:53:49)
And going back to that three vector thing, there is Memory Bandwidth FLOPS and interconnect. The closer the chips are together, the easier it is to do high-speed interconnects. And this is also a reason why you want to go water cooling is because you can just put the chips right next to each other, and therefore get higher speed connectivity.
Lex Fridman
(03:54:13)
I got to ask you, in one of your recent posts, there’s a section called cluster measuring contest. So…
Dylan Patel
(03:54:22)
There’s another word there, but I won’t say it.
Lex Fridman
(03:54:28)
Who’s got the biggest now and who’s going to have the biggest?
Dylan Patel
(03:54:31)
Today, individual largest is Elon. Right?
Lex Fridman
(03:54:36)
Right. Elon’s cluster.
Dylan Patel
(03:54:36)
Elon’s cluster in Memphis, 200,000 GPUs. Meta has 128,000, OpenAI has 100,000 now. Now to be clear, other companies have more GPUs than Elon. They just don’t have them in one place. And for training, you want them tightly connected. There’s some techniques that people are researching and working on that let you train across multiple regions. But for the most part, you want them all in one area so you can connect them highly with high-speed networking.

(03:55:02)
And so, Elon today has 200,000 H100s, 100,000 H100s and 100,000 H200s. Meta, OpenAI, and Amazon all have on the scale of a hundred thousand, a little bit less. But next this year, people are building much more. Anthrophic and Amazon are building a cluster of 400,000 trainium 2, which is Amazon-specific chip trying to get away from NVIDIA. Meta and OpenAI have scales for hundreds of thousands. But by next year, you’ll have 500,000 to 700,000 GPU clusters. And note, those GPUs are much higher power consumption than existing ones. Hopper’s 700 watts, Blackwell goes to 1,200 watts.

(03:55:45)
So, the power per chip is growing and the number of chips is growing.
Lex Fridman
(03:55:50)
Nuts. Elon said he’ll get to a million. Do you think that’s actually feasible?
Dylan Patel
(03:55:56)
I mean, I don’t doubt Elon. The filings that he has for the power plant and the Tesla battery packs, it’s clear he has some crazy plans for Memphis. Permits and stuff is open record, but it’s not quite clear what the time scales are. I just never doubt Elon. He’s going to surprise us.
Lex Fridman
(03:56:16)
So, what’s the idea with these clusters? If you have a million GPUs, what percentage in a, let’s say 2 or 3 years, is used for training? What percent pre-training, and what percent is used for the actual computation?
Dylan Patel
(03:56:31)
These mega clusters make no sense for inference. You could route inference there and just not train. But most of the inference capacity is being, “Hey, I’ve got a 30-megawatt data center here, I’ve got 50 megawatts here, I’ve got 100 here.” Whatever. I’ll just throw inference in all of those because the mega clusters, multi-gigawatt data centers, I want to train there because that’s where all of my GPUs are co-located where I can put them at a super high networking speed connected together. Because that’s what you need for training.

(03:56:58)
Now with pre-training, this is the old scale. You can increase parameters, you did increase data, model gets better. That doesn’t apply anymore because there’s not much more data in the pre-training side. Yes, there’s video and audio and image that has not been fully taken advantage of, so there’s a lot more scaling. But a lot of people have transcript, taken transcripts out of YouTube videos, and that gets you a lot of the data. It doesn’t get you all of the learning value out of the video and image data, but…

(03:57:23)
There’s still scaling to be done on pre-training, but this post-training world is where all the FLOPS are going to be spent. The model’s going to play with itself, it’s going to self-play, it’s going to do verifiable tasks, it’s going to do computer use in sandboxes. It might even do simulated robotics things. All of these things are going to be environments where compute is spent in “post-training.” But I think it’s going to be good. We’re going to drop the post from post-training.
Nathan Lambert
(03:57:48)
Yeah. Wow.
Dylan Patel
(03:57:49)
It’s going to be pre-training and it’s going to be training, I think, at some point. [inaudible 03:57:53] At some point. Because for bulk of the last few years, pre-training has dwarfed post-training. But with these verifiable methods, especially ones that scale really potentially infinitely, like computer use in robotics, not just math and coding where you can verify what’s happening, those infinitely verifiable tasks, it seems you can spend as much compute as you want on this.
Nathan Lambert
(03:58:13)
Especially at the context length increase because the end of pre-training is when you increase the context length for these models. And we’ve talked earlier in the conversation about how the context length, when you have a long input, is much easier to manage than output. And a lot of these post-training and reasoning techniques rely on a ton of sampling, and it’s becoming increasingly long context. So just like effectively, your compute efficiency goes down.

(03:58:36)
I think FLOPS is the standard for how you measure it. But with RL, and you have to do all of these things where you move your weights around in a different way than at pre-training and just generation, it’s going to be become less efficient and FLOPS is going to be less of a useful term. And then as the infrastructure gets better, it’s probably going to go back to FLOPS.
Lex Fridman
(03:58:57)
So, all of the things we’ve been talking about is most likely going to be NVIDIA, right? Is there any competitors of GPU?
Dylan Patel
(03:59:03)
Google kind of ignored them. I was getting-
Nathan Lambert
(03:59:06)
I was like, “Ah?”
Lex Fridman
(03:59:08)
What’s the story with TPU? What’s the…
Dylan Patel
(03:59:10)
TPU is awesome. It’s great. Google is, they’re a bit more tepid on building data centers for some reason. They’re building big data centers, don’t get me wrong, and they actually have the biggest cluster. I was talking about NVIDIA clusters. They actually have the biggest cluster. Period.

(03:59:25)
But the way they do it is very interesting. They have two data center super regions in that the data center isn’t physically… All of the GPUs aren’t physically on one site but they’re like 30 miles from each other. And they’re not GPUs, TPUs. In Iowa and Nebraska, they have four data centers that are just right next to each other.
Lex Fridman
(03:59:44)
Why doesn’t Google flex its cluster size?
Dylan Patel
(03:59:48)
Go to multi-data center training, there’s good images in there. I’ll show you what I mean. It’s just semi-analysis multi-data center.

(03:59:56)
This is an image of what a standard Google data center looks like. By the way, their data centers look very different than anyone else’s data centers.
Lex Fridman
(04:00:01)
What are we looking at here?
Dylan Patel
(04:00:03)
So if you see this image, in the center, there are these big rectangular boxes. Those are where the actual chips are kept. And then if you scroll down a little bit further, you can see there’s these water pipes, there’s these chiller cooling towers in the top, and a bunch of diesel generators. The diesel generators are backup power. The data center itself look physically smaller than the water chillers. The chips are actually easier to keep together, but then cooling all the water for the water cooling is very difficult.

(04:00:33)
So, Google has a very advanced infrastructure that no one else has for the TPU. And what they do is they’ve stamped a bunch of these data centers out in a few regions. So if you go a little bit further down… This is a Microsoft. This is in Arizona. This is where GPT-5 “will be trained.”
Nathan Lambert
(04:00:52)
If it doesn’t exist already.
Dylan Patel
(04:00:54)
Yeah, if it doesn’t exist already. But each of these data centers, I’ve shown a couple images of them, they’re really closely co-located in the same region. Nebraska, Iowa. And then they also have a similar one in Ohio complex. And so, these data centers are really close to each other. And what they’ve done is they’ve connected them super high bandwidth with fiber. And so, these are just a bunch of data centers.

(04:01:15)
And the point here is that Google has a very advanced infrastructure, very tightly connected in a small region. So, Elon will always to have the biggest cluster fully connected because it’s all in one building, and he’s completely right on that. Google has the biggest cluster but you have to spread over three sites, and by a significant margin. We have to go across multiple sites.
Lex Fridman
(04:01:35)
Why doesn’t Google compete with NVIDIA? Why don’t they sell TPUs?
Dylan Patel
(04:01:41)
I think there’s a couple of problems with it. It’s like, one, TPU has been a form of allowing search to be really freaking cheap and build models for that. And so, a big chunk of the search, GPU purchases or TPU purchases or big chunk of Google’s purchases and usage, all of it is for internal workloads. Whether it be search, now Gemini, YouTube, all these different applications that they have ads. These are where all their TPUs are being spent and that’s what they’re hyper-focused on. And so, there’s certain aspects of the architecture that are optimized for their use case that are not optimized elsewhere.

(04:02:21)
One simple one is they’ve open sourced a Gemma model, and they called it Gemma-7B. But then, it’s actually 8 billion parameters because the vocabulary is so large. And the reason they made the vocabulary so large is because TPUs matrix multiply unit is massive because that’s what they’ve optimized for. And so they decided, “Oh, well, I’ll just make the vocabulary large, too.” Even though it makes no sense to do so in such a small model, because that fits on their hardware. Gemma doesn’t run it as efficiently on a GPU as a Llama does. But vice versa, Llama doesn’t run as efficiently on a TPU as a Gemma does.

(04:02:53)
There’s certain aspects of hardware, software co-design. All their search models are there, ranking and recommendation models, all these different models that are AI but not like gen AI have been hyper optimized with TPUs forever. The software stack is super optimized. But all of this software stack has not been released publicly at all. Very small portions of it. JAX and XLA have been. But the experience when you’re inside of Google and you’re training on TPUs as a researcher, you don’t need to know anything about the hardware in many cases, right? It’s pretty beautiful.
Nathan Lambert
(04:03:24)
They all loved it.
Dylan Patel
(04:03:24)
But as soon as you step outside-
Nathan Lambert
(04:03:25)
A lot of them go back. They leave Google and then they go back.
Lex Fridman
(04:03:28)
Yeah.
Dylan Patel
(04:03:29)
Yeah. They leave and they start a company because they have all of these amazing research ideas. And they’re like, “Wait. Infrastructure’s hard, software is hard.” And this is on GPUs. Or if they try to use TPUs, same thing, because they don’t have access to all this code. And so it’s like, how do you convince a company whose golden goose is search where they’re making hundreds of billions of dollars from, to start selling GPU or TPUs which they used to only buy a couple of billion of…

(04:03:51)
I think in 2023, they bought a couple of billion. And now, they’re buying like 10 billion to $15 billion worth. But how do you convince them that they should just buy twice as many and figure out how to sell them, and make $30 billion? Who cares about making $30 billion?
Lex Fridman
(04:04:05)
Won’t that 30 billion exceed actually the search profit eventually?
Dylan Patel
(04:04:11)
You’re always going to make more money on services than…
Lex Fridman
(04:04:14)
Always.
Dylan Patel
(04:04:15)
I mean, yeah. To be clear, today, people are spending a lot more on hardware than they are with the services because the hardware front runs the service spend. But-
Lex Fridman
(04:04:25)
You’re investing, yeah.
Dylan Patel
(04:04:27)
… if there’s no revenue for AI stuff or not enough revenue, then obviously, it’s going to blow up. People won’t continue to spend on GPUs forever. And NVIDIA is trying to move up the stack with software that they’re trying to sell and licensed and stuff. But Google has never had that DNA of like, “This is a product we should sell.” The Google Cloud, which is a separate organization from the TPU team, which is a separate organization from the DeepMind team, which is a separate organization from the Search team. There’s a lot of bureaucracy here.
Lex Fridman
(04:04:52)
Wait. Google Cloud is a separate team than the TPU team?
Dylan Patel
(04:04:55)
Technically, TPU sits under infrastructure, which sits under Google Cloud. But Google Cloud, for renting stuff-
Dylan Patel
(04:05:00)
… But Google cloud for renting stuff and TPU architecture are very different goals, and hardware and software, all of this, right? The Jax XLA teams do not serve Google’s customers externally. Whereas NVIDIA’s various CUDA teams for things like NCCL serve external customers. The internal teams like Jax and XLA and stuff, they more so serve DeepMind and Search, right? And so their customer is different. They’re not building a product for them.
Lex Fridman
(04:05:27)
Do you understand why AWS keeps winning versus Azure for cloud versus Google Cloud?
Dylan Patel
(04:05:34)
Yeah, there’s-
Lex Fridman
(04:05:35)
Google Cloud is tiny, isn’t it, relative to AWS?
Dylan Patel
(04:05:37)
Google Cloud is third. Yeah. Microsoft is the second biggest, but Amazon is the biggest, right?
Lex Fridman
(04:05:37)
Yeah.
Dylan Patel
(04:05:43)
And Microsoft deceptively sort of includes Microsoft Office 365 and things like that, some of these enterprise-wide licenses. So in reality, the gulf is even larger. Microsoft is still second though, right? Amazon is way bigger. Why? Because using AWS is better and easier. And in many cases, it’s cheaper-
Nathan Lambert
(04:06:00)
It was first.
Dylan Patel
(04:06:00)
And it’s first. It was first.
Lex Fridman
(04:06:00)
Yeah. But there’s a lot of things that are first that lose the-
Nathan Lambert
(04:06:03)
Well, it’s harder to switch than it is to-
Lex Fridman
(04:06:05)
Yeah, okay.
Dylan Patel
(04:06:05)
AWS is-
Lex Fridman
(04:06:05)
Because there’s large-
Nathan Lambert
(04:06:07)
There’s big fees for switching too.
Dylan Patel
(04:06:09)
AWS generates over 80% of Amazon’s profit. I think over 90%.
Lex Fridman
(04:06:13)
That’s insane.
Dylan Patel
(04:06:13)
The distribution centers are just like one day we’ll decide to make money from this, but they haven’t yet, right? They make tiny little profit from it.
Nathan Lambert
(04:06:20)
Yeah, one day Amazon Prime will triple in price.
Lex Fridman
(04:06:22)
You would think they would improve AWS interface because it’s horrible. It’s clunky, but everybody is.
Nathan Lambert
(04:06:31)
Yeah, one would think.
Dylan Patel
(04:06:33)
I think actually Google’s interface is sometimes nice, but it’s also they don’t care about anyone besides their top customers.
Lex Fridman
(04:06:38)
Exactly.
Dylan Patel
(04:06:39)
And their customer service sucks and they have a lot less-
Lex Fridman
(04:06:42)
I mean, all these companies, they optimize for the big customers. Yeah, it’s supposed to be for business.
Dylan Patel
(04:06:47)
Amazon has always optimized for the small customer too though. Obviously they optimize a lot for the big customer, but when they started, they just would go to random Bay Area things and give out credits or just put in your credit card and use us back in the early days. The business has grown with them and [inaudible 04:07:04]. Why is Snowflake all over Amazon? Because Snowflake in the beginning, when Amazon didn’t care about them, was still using Amazon. And then of course one day Snowflake and Amazon has a super huge partnership, but this is the case. Amazon’s user experience and quality is better.

(04:07:17)
Also, a lot of the silicon they’ve engineered makes them have a lower cost structure in traditional cloud, storage, CPU networking, that kind of stuff than in databases. I think four of Amazon’s top five revenue products, margin products like gross profit products are all database-related products like Redshift and all these things. So Amazon has a very good silicon to user experience like entire pipeline with AWS. I think Google, their silicon teams, they have awesome silicon internally, TPU, the YouTube chip, some of these other chips that they’ve made. And the problem is they’re not serving external customers, they’re serving internal customers, right?
Nathan Lambert
(04:07:58)
I mean, NVIDIA’s entire culture is designed from the bottom up to do this. There’s this recent book, The NVIDIA Way by Tae Kim, that details this and how they look for future opportunities and ready their CUDA software libraries to make it so that new applications of high-performance computing can very rapidly be evolved on CUDA and NVIDIA chips. And that is entirely different than Google as a services business.
Lex Fridman
(04:08:24)
I mean NVIDIA, it should be said, is a truly special company. I mean there’s the culture of everything. They’re really optimized for that kind of thing. Speaking of which, is there somebody that can even challenge NVIDIA hardware-wise? Intel? AMD?
Dylan Patel
(04:08:39)
I really don’t think so. We went through a very long process of working with AMD on training on their GPUs inference and stuff. And they’re decent, their hardware is better in many ways than in NVIDIA’s. The problem is their software is really bad and I think they’re getting better, right? They’re getting better, faster, but the gulf is so large and they don’t spend enough resources on it or haven’t historically, right? Maybe they’re changing their tune now, but for multiple months we were submitting the most bugs like us semi-analysis like what the fuck? Why are we submitting the most bugs? Because they only cared about their biggest customers and so they’d ship them a private image, blah, blah, blah. And it’s like, “Okay, but I am just using PyTorch and I want to use the publicly available libraries,” and you don’t care about that. So they’re getting better, but I think AMD is not possible. Intel is obviously in dire straits right now and needs to be saved somehow. Very important for national security, for American technology comments.
Lex Fridman
(04:09:39)
Can you explain the obviously, so why are they in dire straits?
Dylan Patel
(04:09:41)
Going back to earlier, only three companies can R&D, right? Taiwan Hsinchu, Samsung [inaudible 04:09:49], and then Intel Hillsboro. Samsung’s doing horribly. Intel’s doing horribly. We could be in a world where there’s only one company that can do R& and that one company already manufactures most of chips. They’ve been gaining market share anyways, but that’s a critical thing. So what happens to Taiwan means the rest of the world, semiconductor industry and therefore tech relies on Taiwan and that’s obviously precarious as far as Intel, they’ve been slowly, steadily declining. They were on top of servers and PCs, but now Apple’s done the M1 and Nvidia’s releasing a PC chip and Qualcomm’s releasing a PC chip.

(04:10:21)
And in servers, hyperscalers are all making their own ARM-based server chips and Intel has no AI silicon like wins. They have very small wins and they never got into mobile because they said no to the iPhone and all these things have compounded and they’ve lost their process technology leadership. They were ahead for 20 years and now they’re behind by at least a couple years and they’re trying to catch back up and we’ll see if their 18A, 14A strategy works out where they try and leapfrog TSMC like and Intel is just losing tons of money anyways, and they just fired their CEO, even though the CEO was the only person who understood the company well, right? We’ll see. He was not the best, but he was pretty good relatively technical guy.
Lex Fridman
(04:11:01)
Where does Intel make most of its money? The CPUs though.
Dylan Patel
(04:11:04)
PCs and data center CPUs, yeah, but data center CPUs are all going cloud and Amazon, Microsoft, Google are making ARM-based CPUs. And then PC side, AMD’s gained market share, Nvidia’s launching a chip, that’s not going to be a success, right? MediaTek, Qualcomm ever launched chips. Apple’s doing well. They could get squeezed a little bit in PC, although PC generally I imagine will just stick Intel mostly for Windows side.

Who wins the race to AGI?

Lex Fridman
(04:11:27)
Let’s talk about the broad AI race. Who do you think wins? We talked about Google, Meta.
Nathan Lambert
(04:11:33)
The default leader has been Google because of their infrastructure advantage.
Lex Fridman
(04:11:37)
Well, in the news, OpenAI is the leader.
Nathan Lambert
(04:11:40)
They’re the leading in the narrative.
Dylan Patel
(04:11:42)
They have the best model.
Nathan Lambert
(04:11:43)
They have the best model that people can use and they’re experts-Experts.
Dylan Patel
(04:11:47)
And they have the most AI revenue.
Nathan Lambert
(04:11:48)
Yeah. OpenAI is winning.
Lex Fridman
(04:11:51)
So who’s making money on AI right now? Is anyone making money?
Dylan Patel
(04:11:55)
So accounting profit-wise, Microsoft is making money, but they’re spending a lot of CapEx and that gets depreciated over years. Meta’s making tons of money with recommendation systems, which is AI, but not with Llama, right? Llama’s losing money for sure. I think Anthropic and OpenAI are obviously not making money otherwise they wouldn’t be raising money. They have to raise money to build more. Although theoretically they are making money. You spent a few hundred million dollars on GPT-4 and it’s doing billions in revenue. So obviously it’s making money. Although they had to continue to research to get the compute efficiency wins and moved down the curve to get that 1200x that has been achieved for GPT-3. Maybe we’re only at a couple hundred X now, but know with GPT-4 Turbo and 4.0 And there’ll be another one probably cheaper than GPT-4.0 even that comes out at some point.
Lex Fridman
(04:12:45)
And that research costs a lot of money.
Dylan Patel
(04:12:48)
Yep, exactly.
Lex Fridman
(04:12:49)
That’s the thing that I guess is not talked about with the cost, that when you’re referring to the cost of the model, it’s not just the training or the test runs, it’s the actual research, the manpower.
Dylan Patel
(04:13:02)
Yeah, to do things like reasoning right now that exists. They’re going to scale it. They’re going to do a lot of research still. I think people focus on the payback question, but it’s really easy to just be like, well, GDP is humans and industrial capital. And if you can make intelligence cheap, then you can grow a lot, right? That’s the sort of dumb way to explain it. But that’s sort of what basically the investment thesis is. I think only Nvidia is actually making tons of money and other hardware vendors, the hyperscalers are all on paper making money, but in reality they’re spending a lot more on purchasing the GPUs, which you don’t know if they’re still going to make this much money on each GPU in two years, right?

(04:13:40)
You don’t know if all of a sudden OpenAI goes kapoof and now Microsoft has hundreds of thousands of GPUs they were renting to OpenAI that they paid for themselves with their investment in them that no longer have a customer. This is always a possibility. I don’t believe that. I think OpenAI will keep raising money. I think others will keep raising money because the returns from it are going to be eventually huge once we have AGI.
Lex Fridman
(04:14:08)
So do you think multiple companies will get, let’s assume-
Dylan Patel
(04:14:11)
I don’t think it’s winner take all.
Lex Fridman
(04:14:12)
Okay, so let’s not call it AGI whatever. It’s like a single day. It’s a gradual thing-
Nathan Lambert
(04:14:18)
Powerful AI. Super powerful AI.
Lex Fridman
(04:14:20)
But it’s a gradually increasing set of features that are useful and make-
Nathan Lambert
(04:14:20)
Rapidly increasing set of features.
Lex Fridman
(04:14:25)
Rapidly increasing set of features. So you’re saying a lot of companies will be… It just seems absurd that all of these companies are building gigantic data centers.
Nathan Lambert
(04:14:41)
There are companies that will benefit from AI but not because they train the best model. Meta has so many avenues to benefit from AI and all of their services. People are there. People spend time on that as platforms, and it’s a way to make more money per user per hour.
Lex Fridman
(04:14:54)
It seems like Google X/X AI/ Tesla important to say. And then Meta will benefit not directly from the AI like the LLMs, but from the intelligence, like the additional boost of intelligence to the products they already sell. So whether that’s the recommendation system or for Elon who’s been talking about Optimus, the robot, potentially the intelligence of the robot, and then you have personalized robots in the home, that kind of thing. He thinks it’s a 10 plus trillion dollars business, which…
Nathan Lambert
(04:15:30)
At some point, maybe. Not soon, but who knows when robotics will use for-
Dylan Patel
(04:15:36)
Let’s do a TAM analysis, 8 billion humans and let’s get 8 billion robots and let’s pay them the average salary. And there we go. 10 trillion. More than 10 trillions.
Lex Fridman
(04:15:46)
Yeah, I mean if there’s robots everywhere, why does it have to be just 8 billion robots?
Dylan Patel
(04:15:52)
Yeah, yeah, of course. Of course. I’m going to have one robot. You’re going to have like 20.
Lex Fridman
(04:15:57)
Yeah, I mean I see a use case for that. So yeah, so I guess the benefit would be in the products they sell, which is why OpenAI’s in a trickier position because they-
Nathan Lambert
(04:16:06)
All of the value of OpenAI right now as a brand is in ChatGPT and for most users, there’s not that much of a reason that they need OpenAI to be spending billions and billions of dollars on the next best model when they could just license Llama 5 and for be way cheaper. So that’s kind of like ChatGPT is an extremely valuable entity to them, but they could make more money just off that.
Dylan Patel
(04:16:31)
The chat application clearly does not have tons of room to continue. The standard chat where you’re just using it for a random question and stuff. The cost continues to collapse. V3 is the latest one.
Nathan Lambert
(04:16:41)
It’ll go down with the ads.
Dylan Patel
(04:16:43)
But it’s going to get supported by ads. Meta already serves 405B and probably loses the money, but at some point the models are going to get so cheap that they can just serve them for free with ad supported and that’s what Google is going to be able to do. And obviously they’ve got a bigger reach. Chat is not going to be the only use case. It’s like these reasoning, code, agents, computer use, all this stuff is where OpenAI has to actually go to make money in the future otherwise they’re kaputs.
Lex Fridman
(04:17:10)
But X, Google, and Meta have these other products. So isn’t it likely that OpenAI and Anthropic disappear eventually?
Dylan Patel
(04:17:22)
Unless they’re so good at models, which they are.
Lex Fridman
(04:17:24)
But it’s such a cutting edge. I mean-
Nathan Lambert
(04:17:25)
It depends on where you think AI capabilities are going.
Lex Fridman
(04:17:28)
You have to keep winning.
Dylan Patel
(04:17:30)
Yes.
Lex Fridman
(04:17:30)
You have to keep winning as you climb, even if the AI capabilities are going super rapidly awesome into the direction of AGI, there’s still a boost for X in terms of data, Google in terms of data, Meta in terms of data, in terms of other products and the money and there’s just huge amounts of money.
Dylan Patel
(04:17:50)
The whole idea is human data is kind of tapped out. We don’t care. We all care about self-play, verifiable task.
Nathan Lambert
(04:17:57)
Think about AWS.
Lex Fridman
(04:17:58)
Yes, self-play, which is an RNG problem.
Nathan Lambert
(04:17:58)
AWS does not make a lot of money on each individual machine. And the same can be said for the most powerful AI platform, which is even though the calls to the API are so cheap, there’s still a lot of money to be made by owning that platform. And there’s a lot of discussions as it’s the next compute layer.
Dylan Patel
(04:18:15)
You have to believe that. And there’s a lot of discussions that tokens and tokenomics and LLM, APIs are the next compute layer, are the next paradigm for the economy like energy and oil was. But you have to sort of believe that APIs and chat are not where AI is stuck. It is actually just tasks and agents and robotics and computer use, and those are the areas where all the value will be delivered, not API, not chat application.
Lex Fridman
(04:18:42)
So is it possible you have it all just becomes a commodity and you have the very thin wrapper like Perplexity, just joking.
Nathan Lambert
(04:18:54)
There are a lot of wrappers making a lot of money.
Lex Fridman
(04:18:57)
But do you think it’s possible that people would just even forget what OpenAI and Anthropic is just there’ll be wrappers around the API and it just dynamically-
Dylan Patel
(04:19:06)
If model progress is not rapid, yeah. It’s becoming a commodity, right? DeepSeek V3 shows this, but also the GPT-3 chart earlier, Kurt [inaudible 04:19:14] showed this, right? Llama 3B is 1200X cheaper than GPT-3. Anyone whose business model was GPT-3 level capabilities is dead. Anyone whose business models GPT-4 level capabilities is dead.
Nathan Lambert
(04:19:26)
It is a common saying that the best businesses being made now are ones that are predicated on models getting better.
Lex Fridman
(04:19:32)
Right. Which would be like wrappers, thing that is riding the wave of the models.
Nathan Lambert
(04:19:37)
The short-term that company that could make the most money is the one that figures out what advertising targeting method works for language model generations. We have the Meta ads which are hyper-targeted in feed, not within specific pieces of content. And we have search ads that are used by Google and Amazon has been rising a lot on search. But within a return from ChatGPT, it is not clear how you get a high-quality placed ad within the output. And if you can do that with model costs coming down, you can just get super high revenue. That revenue is totally untapped and it’s not clear technically how it’s done.
Lex Fridman
(04:20:12)
Yeah, that is, I mean sort of the AdSense innovation that Google did, the one day you’ll have in GPT output an ad and that’s going to make billions, if not-
Nathan Lambert
(04:20:25)
And it could be very subtle, it could be in conversation, we have voice mode now. It could be some way of making it so the voice introduces certain things. It’s much harder to measure and it takes imagination, but yeah.
Lex Fridman
(04:20:35)
And it wouldn’t come off shady so that you would receive public blowback, that kind of thing. So you have to do it loud enough to where it’s clear it’s an ad and balance all of that. So that’s the open question they’re trying to solve. Anthropic and OpenAI, they need to-
Nathan Lambert
(04:20:51)
They might not say that they’re trying-
Dylan Patel
(04:20:53)
I don’t think they care about that at all.
Nathan Lambert
(04:20:53)
They don’t care about it right now. I think it’s places like Perplexity are experimenting on that more.
Lex Fridman
(04:20:59)
Oh, interesting. Yeah, for sure.
Dylan Patel
(04:21:01)
Perplexity, Google, Meta care about this. I think OpenAI and Anthropic are purely laser focused on-
Lex Fridman
(04:21:07)
AGI.
Dylan Patel
(04:21:08)
Yeah. Like agents and AGI, and if I build AGI, I can make tons of money or I can pay for everything. And it’s just predicated back on the export control thing. If you think AGI is five, 10 years away or less, these labs think it’s two, three years away. Obviously your actions are, if you assume they’re rational actors, which they are mostly what you do in a two-year AGI versus five year versus 10 years, very, very, very different. Right?

AI agents

Lex Fridman
(04:21:39)
Do you think agents are promising? We have to talk about this. This is the excitement of the year that agents are going to rev.. This is the generic hype term that a lot of business folks are using. AI agents are going to revolutionize everything.
Nathan Lambert
(04:21:57)
Okay. So mostly the term agent is obviously overblown. We’ve talked a lot about reinforcement learning as a way to train for verifiable outcomes. Agents should mean something that is open-ended and is solving a task independently on its own and able to adapt to uncertainty. There’s a lot of the term agent applied to things like Apple Intelligence, which we still don’t have after the last WWDC, which is orchestrating between apps and that type of tool use thing is something that language models can do really well. Apple Intelligence I suspect will come eventually. It’s a closed domain. It’s your messages app integrating with your photos with AI in the background. That will work. That has been described as an agent by a lot of software companies to get into the narrative.

(04:22:40)
The question is what ways can we get language models to generalize to new domains and solve their own problems in real time. Maybe some tiny amount of training when they’re doing this with fine-tuning themselves or in context learning, which is the idea of storing information in a prompt. And you can use learning algorithms to update that and whether or not you believe that that is going to actually generalize to things like me saying, “Book my trip to go to Austin in two days. I have XYZ constraints,” and actually trusting it. I think there’s an HCI problem coming back for information.
Lex Fridman
(04:23:19)
Well, what’s your prediction there? Because my gut says we’re very far away from that.
Dylan Patel
(04:23:24)
I think OpenAI’s statement, I don’t know if you’ve seen the five levels where it’s chat is level one, reasoning is level two, and then agents is level three. And I think there’s a couple more levels, but it’s important to note, we were in chat for a couple years. We just theoretically got to reasoning, we’ll be here for a year or two, and then agents, but at the same time, people can try and approximate capabilities of the next level, but the agents are doing things autonomously, doing things for minutes at a time, hours at a time, et cetera, right? Reasoning is doing things for tens of seconds at a time and then coming back with an output that I still need to verify and use and try check out. And the biggest problem is of course, it’s the same thing with manufacturing. There’s the whole six sigma thing, how many nines do you get?

(04:24:14)
And then you compound the nines onto each other and it’s like if you multiply by the number of steps that are six sigma, you get to a yield or something. So in semiconductor manufacturing, tens of thousands of steps, 9999999 is not enough. You multiply by that many times you actually end up with 60% yield, right? Really low yield or zero. And this is the same thing with agents, right? Chaining tasks together each time, even the best LLMs in particularly pretty good benchmarks don’t get 100%, right? They get a little bit below that because there is a lot of noise. And so how do you get to enough nines, right? This is the same thing with self-driving. We can’t have self-driving because without it being super geofenced like Google’s and even then they have a bunch of teleoperators to make sure it doesn’t get stuck. But you can’t do that because it doesn’t have enough nines.
Lex Fridman
(04:25:07)
Self-driving has quite a lot of structure because roads have rules, it’s well-defined, there’s regulation. When you’re talking about computer use for the open web, for example, or the open operating system, it’s a mess. So the possibility… I’m always skeptical of any system that is tasked with interacting with the human world, with the open messaging world.
Nathan Lambert
(04:25:36)
That’s the thing. If we can’t get intelligence that’s enough to solve the human world on its own, we can create infrastructure like the human operators for Waymo over many years that enable certain workflows.
Dylan Patel
(04:25:47)
There is a company, I don’t remember it, but it is, but that’s literally their pitch is, “Yeah, we’re just going to be the human operator when agents fail and you just call us and we fix it.” Same thing an API call, and it’s hilarious.
Nathan Lambert
(04:25:57)
There’s going to be teleoperation markets when we get human robots, which is there’s going to be somebody around the world that’s happy to fix the fact that it can’t finish loading my dishwasher when I’m unhappy with it. But that’s just going to be part of the Tesla service package.
Lex Fridman
(04:26:10)
I’m just imagining an AI agent talking to another AI agent. One company has an AI agent that specializes in helping other AI agents.
Nathan Lambert
(04:26:20)
But if you can make things that are good at one step, you can stack them together. So that’s why if it takes a long time, we’re going to build infrastructure that enables it. You see the operator launch, they have partnerships with certain websites, with DoorDash, with OpenTable, with things like this. Those partnerships are going to let them climb really fast. Their model’s going to get really good at those things. It’s going to proof of concept that might be a network effect where more companies want to make it easier for AI. Some companies will be like, “No, let’s put blockers in place.” And this is the story of the internet we’ve seen, we see it now with training data for language models where companies are like, “No, you have to pay.” Business working it out.
Lex Fridman
(04:27:00)
That said, I think airlines and hotels have high incentive to make their site work really well, and they usually don’t. If you look at how many clicks it takes to order airplane ticket, it’s insane.
Nathan Lambert
(04:27:14)
You actually can’t call an American Airlines agent anymore. They don’t have a phone number.
Lex Fridman
(04:27:20)
I mean, it’s horrible on the interface front. And to imagine that agents will be able to deal with that website when I, as a human, struggle, like I have an existential crisis every time I try to book an airplane ticket. I think it’s going to be extremely difficult to build an AI agent that’s robust in that way.
Nathan Lambert
(04:27:40)
But think about it, United has accepted the Starlink term, which is they have to provide Starlink for free and the users are going to love it. What if one airline is like, “We’re going to take a year and we’re going to make our website have white text that works perfectly for the AIs.” Every time anyone asks about an AI flight, they buy whatever airline it is.
Dylan Patel
(04:28:00)
They’re just like, “Here’s an API and it’s only exposed to AI agents and if anyone queries it, the price is 10% higher for any flight, but we’ll let you see any of our flights and you can just book any of them. Here you go.”
Nathan Lambert
(04:28:11)
And then that’s it.
Dylan Patel
(04:28:12)
It’s like, “Oh, and I made 10% higher price. Awesome.” And am I willing to say that for like, “Hey, book me a flight to [inaudible 04:28:18].” Right? And it’s like, yeah, whatever. I think computers and real world and the open world are really, really messy, but if you start defining the problem in narrow regions, people are going to be able to create very, very productive things and ratchet down cost massively, right? Now, crazy things like robotics in the home, those are going to be a lot harder to do just like self-driving because there’s just a billion different failure modes, but agents that can navigate a certain set of websites and do certain sets of tasks or take a photo of your fridge or upload your recipes and then it figures out what to order from Amazon/Whole Foods food delivery, and that’s going to be pretty quick and easy to do, I think. So it’s going to be a whole range of business outcomes and it’s going to be tons of optimism around people can just figure out ways to make money.
Nathan Lambert
(04:29:14)
To be clear, these sandboxes already exist in research. There are people who have built clones of all the most popular websites of Google, Amazon, blah, blah, blah, to make it so that there’s… And I mean open AI probably has them internally to train these things. It’s the same as DeepMind’s robotics team for years has had clusters for robotics where you interact with robots fully, remotely. They just have a lab in London and you send tasks to it, arrange the blocks, and you do this research. Obviously there’s techs there that fix stuff, but we’ve turned these cranks of automation before.

(04:29:46)
You go from sandbox to progress and then you add one more domain at a time and generalize, I think. And the history of NLP and language processing instruction, tuning and tasks per language model used to be like one language model did one task, and then in the instruction tuning literature, there’s this point where you start adding more and more tasks together where it just starts to generalize to every task. And we don’t know where on this curve we are. I think for reasoning with this RL and verifiable domains, we’re early, but we don’t know where the point is where you just start training on enough domains and poof, more domains just start working. And you’ve crossed the generalization barrier.

Programming and AI

Lex Fridman
(04:30:22)
Well, what do you think about the programming context? So software engineering, that’s where I personally, and I know a lot of people interact with AI the most.
Dylan Patel
(04:30:34)
There’s a lot of fear and angst too from current CS students, but that is the area where probably the most AI revenue and productivity gains have come, right? Whether it be Copilots or Cursor or what have you, or just standard ChatGPT. I know very few programmers who don’t have ChatGPT and actually many of them have the $200 tier because that’s what it’s so good for. I think that in that world, we already see it like SWE-bench. And if you’ve looked at the benchmark made by some Stanford students, I wouldn’t say it’s really hard, but I wouldn’t say it’s easy either. I think it takes someone who’s been through at least a few years of CS or a couple years of programming to do SWE-bench, well, and the models went from 4% to 60% in a year, and where are they going to go to next year? It’s going to be higher. It probably won’t be a hundred percent because again, that nines is really hard to do, but we’re going to get to some point where that’s, and then we’re going to need harder software engineering benchmarks and so on and so forth.

(04:31:34)
But the way that people think of it now is it can do code completion. Easy. It can do some function generation. I have to review it. Great. But really the software engineering agents I think can be done faster sooner than any other agent because it is a verifiable domain. You can always unit test or compile, and there’s many different regions of it can inspect the whole code base at once, which no engineer really can. Only the architects can really think about this stuff, the really senior guys, and they can define stuff and then the agent can execute on it. So I think software engineering costs are going to plummet like crazy. And one interesting aspect of that is when software engineering costs are really low, you get very different markets. So in the US, you have all these platform SaaS companies, Salesforce and so on and so forth. In China, no one uses platform SaaS. Everyone just builds their own stack because software engineering is much cheaper in China and partially because people, number of STEM graduates, et cetera. So it’s generally just cheaper to do.

(04:32:38)
And so at the same time, code LLMs have been adopted much less in China because the cost of an engineer there is much lower. But what happens when every company can just invent their own business logic really cheaply and quickly? You stop using platform SaaS, you start building custom tailored solutions, you change them really quickly. Now all of a sudden your business is a little bit more efficient too, potentially because you’re not dealing with the hell that is. Some random platform SaaS company stuff not working perfectly and having to adjust workflows or random business automation cases that aren’t necessarily AI required.

(04:33:08)
It’s just logic that needs to be built that no one has built. All of these things can go happen faster. And so I think software and then the other domain is industrial, chemical, mechanical engineers suck at coding just generally. And their tools like semiconductor engineers, their tools are 20 years old. All the tools run on XP including ASML lithography tools run on Windows XP. And a lot of the analysis happens in Excel, right? It’s just like, “Guys, you guys can move 20 years forward with all the data you have and gathered and do a lot better.” You need the engineering skills for software engineering to be delivered to the actual domain expert engineer. So I think that’s the area where I’m super-duper bullish of generally AI creating value.
Nathan Lambert
(04:33:47)
The big picture is that I don’t think it’s going to be a cliff. I think a really good example of how growth changes is when Meta added stories. So Snapchat was on an exponential, they added stories, it flatlined. Software engineers then up until the right, AI is going to come in, it’s probably just going to be flat. It’s not like everyone’s going to lose their job. It’s hard because the supply corrects more slowly. So the amount of students is still growing, and that’ll correct on a multi-year, like a year delay, but the amount of jobs will just turn and then maybe in 20, 40 years, it’ll be well down. But in the few years, there’ll never going to be the snap moment where it’s like software engineers aren’t useful.
Lex Fridman
(04:34:30)
I think also the nature of what it means to be a programmer and what kind of jobs programmers do changes, because I think there needs to be a human in the loop of everything you’ve talked about. There’s a really important human in that picture of correcting the code, fixing-
Dylan Patel
(04:34:49)
Thinking larger than the context length.
Lex Fridman
(04:34:51)
And debugging also, like debugging by reading the code, understanding the steering the system. No, no, no. You missed the point. Adding more to the prompt like, yes, adding the human-
Nathan Lambert
(04:35:05)
Designing the perfect Google button. Google’s famous for having people design buttons that are so perfect, and it’s like how is AI going to do that? They could give you all the ideas. Perfect, fine.
Lex Fridman
(04:35:17)
I mean, that’s the thing. You can call it taste. One thing humans can do is figure out what other humans enjoy better than AI systems. That’s where the preference you loading that in. But ultimately, humans are the greatest preference generator. That’s where the preference comes from.
Nathan Lambert
(04:35:32)
And humans are actually very good at reading or judging between two things versus… This goes back to the core of what RLHF and preference tuning is that it’s hard to generate a good answer for a lot of problems, but it’s easy to see which one is better. And that’s how we’re using humans for AI now is judging which one is better, and that’s what software engineering could look like. The PR review, here’s a few options, here are some potential pros and cons, and they’re going to be judges.
Lex Fridman
(04:35:59)
I think the thing I would very much recommend is programmers start using AI and embracing that role of the supervisor of the AI system and partner the AI system versus writing from scratch or not learning coding at all and just generating stuff because I think there actually has to be a pretty high level of expertise as a programmer to be able to manage increasingly intelligent systems.
Dylan Patel
(04:36:24)
I think it’s that and then becoming a domain expert in something.
Lex Fridman
(04:36:27)
Sure. Yeah.
Dylan Patel
(04:36:28)
Because seriously, if you go look at aerospace or semiconductors or chemical engineering, everyone is using really crappy platforms, really old software. The job of a data scientist is a joke in many cases. In many cases, it’s very real, but it’s like bring what the forefront of human capabilities are to your domain. And even if the forefront is from the AI, your domain, you’re at the forefront. So it’s like you have to be at the forefront of something and then leverage the rising tide that is AI for everything else.
Lex Fridman
(04:36:57)
Oh, yeah. There’s so many low hanging fruit everywhere in terms of where software can help automate a thing or digitize a thing in the legal system. That’s why DOGE is exciting. I got to hang out with a bunch of the DOGE folks, and I mean, government is so old school. It’s like begging for the modernization of software, of organizing the data, all this kind of stuff. I mean, in that case it’s by design because bureaucracy protects centers of power and so on. But software breaks down those barriers, so it hurts those that are holding onto power, but ultimately benefits humanity. So there’s a bunch of domains of that kind. One thing we didn’t fully finish talking about is open source. So first of all, congrats. You released a new model.

Open source

Nathan Lambert
(04:37:58)
Yeah, this is-
Lex Fridman
(04:37:58)
Tülu.
Nathan Lambert
(04:38:00)
I’ll explain what a tülu is. A tülu is a hybrid camel when you breed a dromedary with a Bactrian camel. Back in the early days after ChatGPT, there was a big wave of models coming out like Alpaca, Vicuna, et cetera, that were all named after various mammalian species. Tülu, the brand, is multiple years old, which comes from that.

(04:38:19)
And we’ve been playing at the frontiers of post-training with open source code. And this first part of this release was in the fall where we’ve built on Llama’s, open models, open weight models, and then we add in our fully open code or fully open data. There’s a popular benchmark that is Chatbot Arena. And that’s generally the metric by which how these chat models are evaluated. And it’s humans compare random models from different organizations. And if you looked at the leaderboard in November or December, among the top 60 models from tens to twenties of organizations, none of them had open code or data for just post-training.

(04:38:58)
Among that, even fewer or none have pre-training data and code available. Post-training is much more accessible at this time. It’s still pretty cheap, and you can do it. And the thing is, how high can we push this number where people have access to all the code and data? So that’s kind of the motivation of the project. We draw in lessons from Llama. Nvidia had a Nemotron model where the recipe for their post-training was fairly open with some data and a paper, and it’s putting all these together to try to create a recipe that people can fine tune models like GPT-4 to their domain.
Lex Fridman
(04:39:28)
To be clear, in the case of Tülu, maybe you can talk about Llama too, but in the case of Tülu, you’re taking Llama 3, 405B.
Nathan Lambert
(04:39:37)
Tülu has been a series of recipes for post-training. So we’ve done multiple models over years.
Lex Fridman
(04:39:44)
And so you’re open sourcing everything.
Nathan Lambert
(04:39:46)
Yeah. If you start with an open weight based model, their whole model technically isn’t open source because you don’t know what Llama put into it, which is why we have the separate thing that we’ll get to, but it’s just getting parts of the pipeline where people can zoom in and customize. I know I hear from startups and businesses, they’re like, “Okay, I can take this post-training-“
Nathan Lambert
(04:40:00)
… I know I hear from startups and businesses, they’re like, “Okay, I can take this post-training and try to apply it to my domain.” We talk about verifiers a lot. We use this idea which is reinforcement learning with verifiable rewards RLVR, kind of similar to RLHF. And we applied it to MAP and the model today, which is we applied it to the Llama 405B base model from last year. And we have our other stuff. We have our instruction tuning and our preference tuning. But the math thing is interesting, which is it’s easier to improve this math benchmark. There’s a benchmark, M-A-T-H, MATH, all capitals, tough name on the benchmark name is the area that you’re evaluating. We’re researchers, we’re not brand strategists.

(04:40:44)
And this is something that the DeepSeek paper talked about as well is at this bigger model, it’s easier to elicit powerful capabilities with this RL training. And then they distill it down from that big model to the small model. And this model we released today, we saw the same thing. We’re at AI2, we don’t have a ton of compute. We can’t train 405B models all the time. So we just did a few runs and they tend to work. And it just shows that there’s a lot of room for people to play in these things and that’s –
Dylan Patel
(04:41:12)
And they crushed Llama’s actual release, they’re way better than it.
Nathan Lambert
(04:41:16)
… Yeah. So our eval numbers, I mean we have extra months in this, but our eval numbers are much better than the Llama instruct model that they released.
Lex Fridman
(04:41:24)
And then you also said better than DeepSeek V3?
Nathan Lambert
(04:41:26)
Yeah, on our eval benchmark. DeepSeek V3 is really similar. We have a safety benchmark to understand if it will say harmful things and things like that. And that’s what draws down most of the way. It’s still-
Dylan Patel
(04:41:37)
It’s like an amalgamation of multiple benchmarks or what do you mean?
Nathan Lambert
(04:41:40)
… Yeah, so we have a 10 evaluator. This is standard practice in post-training is you choose your evaluations you care about. In academics, in smaller labs you’ll have fewer evaluations. In companies, you’ll have a really one domain that you really care about. In Frontier Labs, you’ll have tens to 20s to maybe even 100 evaluations of specific things. So we choose a representative suite of things that look like chat, precise instruction following, which is like respond only in emojis, just model follow weird things like that, math, code. And you create a suite like this. So safety would be one of 10 in that type of suite where you have what does the broader community of AI care about? And for example, in comparison to DeepSeek it would be something like our average eval for our model would be 80, including safety and similar without. And DeepSeek would be like 79% average score without safety and their safety score would bring it down to like 70 or there abouts.
Dylan Patel
(04:42:31)
Oh, so you’d beat them even ignoring safety.
Nathan Lambert
(04:42:33)
Yeah. So this is something that internally, it’s like I don’t want to win only by how you shape the eval benchmark. So if there’s something that’s like people may or may not care about safety in their model, safety can come downstream, safety can be when you host the model for an API, like safety is addressed in a spectrum of locations in AI applications. So it’s like if you want to say that you have the best recipe, you can’t just gate it on these things that some people might not want.

(04:42:56)
And this is, it’s like the time of progress and we benefit if we can release a model later, we have more time to learn new techniques like this RL technique, we had started this in the fall, it’s now really popular reasoning models. The next thing to do for open source post-training is to scale up verifiers, to scale up data to replicate some of DeepSeek’s results. And it’s awesome that we have a paper to draw on and it makes it a lot easier. And that’s the type of things that is going on among academic and closed frontier research in AI.
Lex Fridman
(04:43:28)
Since you’re pushing open source, what do you think is the future of it? Do you think DeepSeek actually changes things since it’s open source or open weight or is pushing the open source movement into the open direction?
Nathan Lambert
(04:43:39)
This goes very back to license discussion. So DeepSeek R1 with a friendly license is a major reset. So it’s like the first time that we’ve had a really clear frontier model that is open weights and with a commercially friendly license with no restrictions on downstream use cases since that data distillation, whatever.This has never been the case at all in the history of AI in the last few years since ChatGPT. There have been models that are off the frontier or models with weird licenses that you can’t really use them.
Dylan Patel
(04:44:04)
So is it Meta’s license pretty much permissible except for five companies?
Nathan Lambert
(04:44:10)
So this goes to what open source AI is, which is there’s also use case restrictions in the Llama license, which says you can’t use it for specific things. So if you come from an open source software background, you would say that that is not an open source license.
Dylan Patel
(04:44:22)
What kind of things are those, though? Are they like-
Nathan Lambert
(04:44:25)
At this point, I can’t pull them off the top of my head, but it’d be like-
Lex Fridman
(04:44:28)
Stuff like competitors?
Nathan Lambert
(04:44:29)
It used to be military use was one and they removed that for scale, it’ll be like CSAM like child abuse material. That’s the type of thing that is forbidden there. But that’s enough from an open source background to say it’s not an open source license.And also the Llama license has this horrible thing where you have to name your model Llama if you touch it to the Llama model. So it’s like the branding thing. So if a company uses Llama, technically the license says that they should say built with Llama at the bottom of their application. And from a marketing perspective, that just hurts. I could suck it up as a researcher, I’m like, oh, it’s fine. It says Llama dash on all of our materials for this release. But this is why we need truly open models, which is we don’t know DeepSeek R1’s data, but-
Dylan Patel
(04:45:12)
Wait, so you’re saying I can’t make a cheap copy of Llama and pretend it’s mine, but I can do this with the Chinese model?
Nathan Lambert
(04:45:18)
… Hell, yeah. That’s what I’m saying. And that’s why it’s like we want this whole open language model thing, he Olmo thing is to try to keep the model where everything is open with the data as close to the frontier as possible. So we’re compute constrained, we’re personnel constrained. We rely on getting insights from people like John Schulman tells us to do URL and outputs. We can make these big jumps, but it just takes a long time to push the frontier of open source. And fundamentally, I would say that that’s because open source AI does not have the same feedback loops as open source software. We talked about open source software for security. Also it’s just because you build something once and can reuse it. If you go into a new company, there’s so many benefits, but if you open source a language model, you have this data sitting around, you have this training code, it’s not like that easy for someone to come and build on and improve because you need to spend a lot on compute, you need to have expertise.

(04:46:12)
So until there are feedback loops of open source AI, it seems like mostly an ideological mission. People like Mark Zuckerberg, which is like America needs this and I agree with him, but in the time where the motivation ideologically is high, we need to capitalize and build this ecosystem around, what benefits do you get from seeing the language model data? And there’s not a lot about that. We’re going to try to launch a demo soon where you can look at an OMO model and a query and see what pre-training data is similar to it, which is legally risky and complicated, but it’s like what does it mean to see the data that the AI was trained on? It’s hard to parse. It’s terabytes of files. It’s like I don’t know what I’m going to find in there, but that’s what we need to do as an ecosystem if people want open source AI to be financially useful.

Stargate

Lex Fridman
(04:47:01)
We didn’t really talk about Stargate. I would love to get your opinion on what the new administration, the Trump administration, everything that’s being done from the America side and supporting AI infrastructure and the efforts of the different AI companies. What do you think about Stargate? What are we supposed to think about Stargate and does Sam have the money?
Dylan Patel
(04:47:23)
Yeah, so I think Stargate is a opaque thing. It definitely doesn’t have $500 billion, doesn’t even have $100 billion dollars. So what they announced is this $500 billion number, Larry Ellison, Sam Altman and Trump said it. They thanked Trump and Trump did do some executive actions that do significantly improve the ability for this to be built faster. One of the executive actions he did is on federal land, you can just basically build data centers in power pretty much like that. And then permitting process is basically gone or you file after the fact. So again, I had of schizo take earlier, another schizo take, if you’ve ever been to the Presidio in San Francisco, beautiful area, you could build a power plant in a data center there if you wanted to because it is federal land. It used to be a military base, but obviously this would people off. It’s a good fit. Anyways, Trump has made it much easier to do this, right? Generally, Texas has the only unregulated grid in the nation as well.
Lex Fridman
(04:48:24)
Let’s go Texas.
Dylan Patel
(04:48:25)
And so therefore ERCOT enables people to build faster as well in addition, the federal regulations are coming down and so Stargate is predicated, and this is why that whole show happened. Now how they came up with a $500 billion number is beyond me. How they came up with $100 billion dollars number makes sense to some extent. And there’s actually a good table in here that I would like to show in that Stargate piece that I had. It’s the most recent one. So anyways, Stargate, it’s basically, it’s a table about cost. There, you passed it already. It’s that one. So this table is kind of explaining what happens. So Stargate is in Abilene, Texas, the first $100 billion of it. That site is 2.2 gigawatts of power in, about 1.8 gigawatts of power consumed. Per GPU, Oracle is already building the first part of this before Stargate came about. To be clear, they’ve been building it for a year.

(04:49:32)
They tried to rent it to Elon in fact, but Elon was like, “It’s too slow. I need it faster.” So then he went and did his Memphis thing, and so OpenAI was able to get it with this weird joint venture called Stargate. They initially signed a deal with just Oracle for the first section of this cluster. This first section of this cluster is roughly $5 billion to $6 billion of server spend, and then there’s another billion or so of data center spend. And then likewise, if you fill out that entire 1.8 gigawatts with the next two generations of NVIDIA’s chips, GB 200, GB 300, VR 200, and you fill it out completely, that ends up being roughly $50 billion of server cost. Plus there’s data center costs plus maintenance costs, plus operation costs plus all these things. And that’s where OpenAI gets to their $100 billion announcement that they had. Because they talked about $100 billion dollars is phase one. That’s this Abilene, Texas data center, right? $ 100 billion of “total cost of ownership.” So it’s not CapEx, it’s not investment, it’s a $100 billion of total cost of ownership.

(04:50:39)
And then there will be future phases. They’re looking at other sites that are even bigger than this 2.2 gigawatts by the way, in Texas and elsewhere. And so they’re not completely ignoring that, but the number of $100 billion that they save for phase one, which I do think will happen. They don’t even have the money for that. Furthermore, it’s not $100 billion dollars, it’s $50 billion of spend and then $50 billion of operational cost power, et cetera, rental pricing, et cetera, because they’re renting it. OpenAI is renting the GPUs from the Stargate joint venture. What money do they actually have, right? SoftBank is going to invest, Oracle is going to invest. OpenAI is going to invest. OpenAI is on the line for $19 billion. Everyone knows that they’ve only got 46 billion in their last round and $4 billion of debt. But there, there’s news of Softbank maybe investing $25 billion into OpenAI. So that’s part of it. So $19 billion can come from there.

(04:51:32)
So OpenAI does not have the money at all to be clear. Ink is not dried on anything. OpenAI has $0 for this, 50 billion in which they’re legally obligated to put 19 billion of CapEx into the joint venture, and then the rest they’re going to pay via renting the GPUs from the joint venture. And then there’s Oracle. Oracle has a lot of money. They’re building the first section completely. They were spending for it themselves, this $6 billion of CapEx, $10 billion of TCO, and they were going to do that first section. They’re paying for that, right? As far as the rest of the section, I don’t know how much Larry wants to spend. At any point he could pull out. This is again, it is completely voluntary. So at any point, there’s no signed ink on this, but he potentially could contribute tens of billions of dollars to be clear. He’s got the money, Oracle’s got the money.

(04:52:17)
And then there’s like MGX is the UAE fund, which technically has $1.5 trillion for investing in AI. But again, I don’t know how real that money is and there’s no ink signed for this, SoftBank does not have $25 billion of cash. They have to sell down their stake in arm, which is the leader in CPUs and they IPO’d it. This is obviously what they’ve always wanted to do, they just didn’t know where they’d redeploy the capital. Selling down the stake in ARM makes a ton of sense. So they can sell that down and invest in this if they want to and invest in OpenAI if they want to. As far as money secured, the first 100,000 GB 200 cluster can be funded. Everything else after that-
Lex Fridman
(04:52:57)
Up in the air.
Dylan Patel
(04:52:58)
… is up in the air. Money’s coming. I believe the money will come. I personally do.
Lex Fridman
(04:53:03)
It’s a belief.
Dylan Patel
(04:53:04)
It’s a belief that they’re going to release better models and be able to raise more money. But the actual reality is that Elon’s right, the money does not exist.
Lex Fridman
(04:53:12)
What does the US government have to do with anything? What does Trump have to do with everything? He’s just a hype man?
Dylan Patel
(04:53:17)
Trump, he’s reducing the regulation so they can build it faster and he’s allowing them to do it because any investment of this side is going to involve antitrust stuff. So obviously he’s going to allow them to do it. He’s going to enable the regulations to actually allow it to be built. I don’t believe there’s any US government dollars being spent on this though.
Lex Fridman
(04:53:37)
So I think he’s also just creating a general vibe that regulation will go down and this is the era of building. So if you’re a builder, you want to create stuff, you want to launch stuff, this is the time to do it.
Dylan Patel
(04:53:50)
And so we’ve had this 1.8 gigawatt data center in our data for over a year now, and we’ve been sending it to all of our clients, including many of these companies that are building the multi gigawatts. But that is at a level that’s not quite, maybe executives seeing $500 billion, $100 billion dollars, and then everyone’s asking them. So it could spur an even faster arms race. Because there’s already an arms race, but this 100 billion, $500 billion number, Trump talking about it on TV, it could spur the arm race to be even faster and more investors to flood in and et cetera, et cetera. So I think you’re right in that sense that open AI or Trump is sort of championing, people are going to build more and his actions are going to let people build more.

Future of AI

Lex Fridman
(04:54:31)
What are you excited about these several years that are upcoming in terms of cluster build outs, in terms of breakthroughs in AI, the best possible future you can imagine in the next couple of years, two, three, four years? What does that look like? It could be very specific technical things like breakthroughs on post-training or it could be just size, big impressive clusters.
Dylan Patel
(04:55:01)
I really enjoy tracking supply chain and who’s involved and what, I really do. It’s really fun to see the numbers, the cost, who’s building what capacity, helping them figure out how much capacity they should build winning deals, strategic stuff. That’s really cool. I think technologically, there’s a lot around the networking side that really excites me with optics and electronics kind of getting closer and closer, whether it be co-packaged optics or some sort of forms of new forms of switching.
Lex Fridman
(04:55:28)
This is internal to a cluster?
Dylan Patel
(04:55:30)
A cluster, yeah. Also multi-data center training. People are putting so much fiber between these data centers and lighting it up with so much bandwidth that there’s a lot of interesting stuff happening on that end. Telecom has been really boring since 5G, and now it’s really exciting again on the hardware side.
Lex Fridman
(04:55:48)
Can you educate me a little bit about the speed of things? So the speed of memory versus the speed of interconnect versus the speed of fiber between data centers. Are these orders of magnitude different? Can we at some point converge towards a place where it all just feels like one computer?
Dylan Patel
(04:56:04)
No, I don’t think that’s possible. It’s only going to get harder to program, not easier. It’s only going to get more difficult and complicated and more layers. The general image that people like to have is this hierarchy of memory, so on-chip is really close, localized within the chip, you have registers and those are shared between some compute elements and then you’ll have caches which are shared between more compute elements. Then you have memory like HBM or DRAM like DDRR memory or whatever it is, and that’s shared between the whole chip. And then you can have pools of memory that are shared between many chips and then storage and you keep zoning out. The access latency across data centers, within the data center within a chip is different. So you’re always going to have different programming paradigms for this. It’s not going to be easy. Programming this stuff is going to be hard, maybe AI can help with programming this.

(04:56:55)
But the way to think about it is that there is sort of the more elements you add to a task, you don’t get strong scaling. If I double the number of chips, I don’t get two exit performance. This is just a reality of computing because there’s inefficiencies.And there’s a lot of interesting work being done to make it not to make it more linear, whether it’s making the chips more networked together more tightly or cool programming models or cool algorithmic things that you can do on the model side. DeepSeek did some of these really cool innovations because they were limited on interconnect, but they still needed to parallelize. Everyone’s always doing stuff. Google’s got a bunch of work and everyone’s got a bunch of work about this. That stuff is super exciting on the model and workload and innovation side. Hardware, solid-state transformers are interesting. For the power side, all sorts of stuff on batteries and there’s all sorts of stuff on.

(04:57:53)
I think if you look at every layer of the compute stack, whether it goes from lithography and etch all the way to fabrication, to optics, to networking, to power, to transformers, to cooling, to a networking, and you just go on up and up and up and up the stack, even air conditioners for data centers are innovating. Copper cables are innovating. You wouldn’t think it, but copper cables, there’s some innovations happening there with the density of how you can pack them and it’s like all of these layers of the stack, all the way up to the models, human progress is at a pace that’s never been seen before.
Lex Fridman
(04:58:24)
I’m just imagining you sitting back in a layer somewhere with screens everywhere, just monitoring the supply chain where all these clusters, all the information you’re gathering, you’re incredible.
Dylan Patel
(04:58:34)
There’s a big team, there’s a big team.
Lex Fridman
(04:58:38)
You do quite incredible work with semi analysis. I mean just keeping your finger on the pulse of human civilization in the digital world. It’s pretty cool just to watch, feel that.
Dylan Patel
(04:58:51)
Yeah, thank you. I guess.
Lex Fridman
(04:58:53)
Feel all of us doing shit. Epic shit.
Dylan Patel
(04:58:57)
The AGI, yeah.
Lex Fridman
(04:58:57)
I feel from meme to reality. Nathan, is there breakthroughs that you’re looking forward to potentially?
Nathan Lambert
(04:59:07)
I had a while to think about this while listening to Dylan’s beautiful response.
Dylan Patel
(04:59:10)
He did listen to me. He was so into it.
Nathan Lambert
(04:59:12)
No, I knew this was coming and it’s like realistically training models is very fun because there’s so much low-hanging fruit. And the thing that makes my job entertaining, I train models, I write analysis about what’s happening with models and it’s fun because there is obviously so much more progress to be had. And the real motivation, why I do this somewhere where I can share things is that there’s just, I don’t trust people that are like, “Trust me bro, we’re going to make AI good.”

(04:59:39)
It’s like we’re the ones that it’s like, we’re going to do it and you can trust us and we’re just going to have all the AI, and it’s just like, I would like a future where more people have a say in what AI is and can understand it, and it’s a little bit less fun that it’s not a positive thing of this is just all really fun. Training models is fun and bring people in as fun, but it’s really AI if it is going to be the most powerful technology of my lifetime, it’s like we need to have a lot of people involved in making that and-
Lex Fridman
(05:00:09)
Making it open helps with that. As accessible as possible, as open as possible, yeah.
Nathan Lambert
(05:00:14)
… In my read of the last few years is that more openness would help the AI ecosystem in terms of having more people understand what’s going on. Rather that’s researchers from non-AI fields to governments to everything. It doesn’t mean that openness will always be the answer. I think then it’ll reassess of what is the biggest problem facing AI and tack on a different angle to the wild ride that we’re on.
Lex Fridman
(05:00:36)
And for me, just from even the user experience, anytime you have like Aparthi said, the aha moments, the magic, seeing the reasoning, the chain of thought, it’s like there’s something really just fundamentally beautiful about that. It’s putting a mirror to ourselves and seeing like, oh, shit. It is solving intelligence as the cliche goal of these companies is, and you get to understand why we humans are special. The intelligence within us is special. And for now also why we’re special in terms of we seem to be conscious and the AI systems for now, and we get to explore that mystery, so it’s just really cool to get to explore these questions that I don’t think I would’ve never imagined would be even possible back when just watching with excitement, deep blue beat Kasparov, I wouldn’t have ever thought this kind of AI would be possible in my lifetime. This is really feels like AI.
Nathan Lambert
(05:01:43)
Yeah.
Lex Fridman
(05:01:43)
It’s incredible.
Nathan Lambert
(05:01:44)
I started with AI learning to fly a silly, a quadrotor, it’s like learning to fly and it learned to fly up. It would hit the ceiling and stop and catch it. It’s like, okay, that is really stupid compared to what’s going on now.
Lex Fridman
(05:01:57)
And now you could probably with natural language tell it to learn to fly and it’s going to generate the control algorithm required to do that probably.
Nathan Lambert
(05:02:05)
There’s low level blockers. We have to do some weird stuff for that, but you can, you definitely can.
Lex Fridman
(05:02:09)
Back to our robotics conversation, yeah, when you have to interact in the actual physical world, that’s hard. What gives you hope about the future of human civilization looking into the next 10 years, 100 years, 1000 years, how long do you think we’ll make it? You think we’ve got 1000 years?
Nathan Lambert
(05:02:28)
I think humans will definitely be around in a 1000 years, I think. There’s ways that very bad things could happen. There’ll be way fewer humans, but humans are very good at surviving. There’s been a lot of things that that is true. I don’t think necessarily we’re good at long-term credit assignment of risk, but when the risk becomes immediate, we tend to figure things out.
Lex Fridman
(05:02:28)
Oh, yeah.
Nathan Lambert
(05:02:49)
And for that reason, there’s physical constraints to things like AGI, like recursive improvement to kill us all type stuff. For the physical reasons and for how humans have figured things out before, I’m not too worried about AI takeover. There are other international things that are worrying, but there’s just fundamental human goodness and trying to amplify that. I think we’re on a tenuous time. And I mean if you look at humanity as a whole, there’s been times where things go backwards, there’s times when things don’t happen at all, and we’re on what should be very positive trajectory right now.
Lex Fridman
(05:03:28)
Yeah, there seems to be progress, but just like with power, there’s like spikes of human suffering and we want to try to minimize the amount of spikes.
Dylan Patel
(05:03:38)
Generally, humanity is going to suffer a lot less, I’m very optimistic about that. I do worry of like techno-fascism type stuff arising. As AI becomes more and more prevalent and powerful and those who control it can do more and more, maybe it doesn’t kill us all, but at some point, every very powerful human is going to want to brain- computer interface so that they can interact with the AGI and all of its advantages in many more way and merge its mind and its capabilities or that person’s capabilities can leverage those much better than anyone else and therefore be, it won’t be one person rule them all, but it will be, the thing I worry about is it’ll be few people, hundreds, thousands, tens of thousands, maybe millions of people rule whoever’s left and the economy around it.

(05:04:27)
And I think that’s the thing that’s probably more worrisome is human-machine amalgamations. This enables an individual human to have more impact on the world and that impact can be both positive and negative. Generally, humans have positive impacts on the world, at least societally, but it’s possible for individual humans to have such negative impacts. And AGI, at least as I think the labs define it, which is not a runaway sentient thing, but rather just something that can do a lot of tasks really efficiently amplifies the capabilities of someone causing extreme damage. But for the most part, I think it’ll be used for profit-seeking motives, which will increase the abundance and supply of things and therefore reduce suffering, right? That’s the goal.
Lex Fridman
(05:05:12)
Scrolling on a timeline, just drowning in dopamine-
Dylan Patel
(05:05:16)
Scrolling open stasis.
Nathan Lambert
(05:05:18)
Scrolling holds the status quo of the world.
Dylan Patel
(05:05:20)
That is a positive outcome, right? If I have food tubes and lung down scrolling and I’m happy, that’s a positive outcome.
Lex Fridman
(05:05:28)
While expanding out into the cosmos. Well, this is a fun time to be alive. And thank you for pushing the forefront of what is possible in humans, and thank you for talking today. This was fun.
Dylan Patel
(05:05:29)
Thanks for having us.
Nathan Lambert
(05:05:41)
Thanks for having us.
Lex Fridman
(05:05:44)
Thanks for listening to this conversation with Dylan Patel and Nathan Lambert. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Richard Feynman. “For a successful technology, reality must take precedence over public relations, for nature cannot be fooled.” Thank you for listening and I hope to see you next time.

Transcript for Marc Andreessen: Trump, Power, Tech, AI, Immigration & Future of America | Lex Fridman Podcast #458

This is a transcript of Lex Fridman Podcast #458 with Marc Andreessen.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Introduction

Marc Andreessen
(00:00:00)
I mean, look, we’re adding a trillion dollars to the national debt every 100 days right now, and it’s now passing the size of the Defense Department budget and it’s compounding, and pretty soon it’s going to be adding a trillion dollars every 90 days, and then it’s going to be adding a trillion dollars every 80 days, and then it’s going to be a trillion dollars every 70 days. And then if this doesn’t get fixed, at some point, we enter a hyper-inflationary spiral and we become Argentina or Brazil. And …
Lex Fridman
(00:00:22)
The following is a conversation with Marc Andreessen, his second time on the podcast. Marc is a visionary tech leader and investor who fundamentally shaped the development of the internet and the tech industry in general over the past 30 years. He’s the co-creator of Mosaic, the first widely used web browser, co-founder of Netscape, co-founder of the legendary Silicon Valley venture capital firm, Andreessen Horowitz, and is one of the most influential voices in the tech world, including at the intersection of technology and politics. This is the Lex Fridman Podcast. To support it, please check out our sponsors in the description. And now, dear friends, here’s Marc Andreessen.

Best possible future

Lex Fridman
(00:01:09)
All right, let’s start with optimism. If you were to imagine the best possible 1 to 2 years, 2025, ’26 for tech, for big tech and small tech, what would it be? What would it look like? Lay out your vision for the best possible scenario trajectory for America?
Marc Andreessen
(00:01:28)
The roaring 20s.
Lex Fridman
(00:01:28)
The roaring 20s.
Marc Andreessen
(00:01:29)
The roaring 20s. I mean, look, couple of things. It is remarkable over the last several years with all of the issues including not just everything in politics, but also COVID and every other thing that’s happened. It’s really amazing, the United States just kept growing. If you just look at economic growth charts, the US just kept growing and very significantly, many other countries stopped growing. So Canada stopped growing, the UK has stopped growing, Germany has stopped growing, and some of those countries may be actually growing backwards at this point. And there’s a very long discussion to be had about what’s wrong with those countries. And there’s of course, plenty of things that are wrong with our country, but the US is just flat out primed for growth. And I think that’s a consequence of many factors, some of which are lucky and some of which through hard work.

(00:02:11)
And so the lucky part is just number one, we just have incredible physical security by being our own continent. We have incredible natural resources. There’s this running joke now that whenever it looks like the US is going to run out of some rare earth material, some farmer in North Dakota kicks over a hay bale and finds a $2 trillion deposit. I mean, we’re just blessed with geography and with natural resources. Energy. We can be energy independent anytime we want. This last administration decided they didn’t want to be, they wanted to turn off American energy. This new administration has declared that they have a goal of turning it on in a dramatic way. There’s no question we can be energy independent, we can be a giant net energy exporter. It’s purely a question of choice, and I think the new administration’s going to do that. And then I would say two other things.

(00:02:56)
One is we are the beneficiaries, and you’re an example of this. We’re a beneficiary. We’re the beneficiary of 50, 100, 200 years of the basically most aggressive driven, smartest people in the world, most capable people moving to the US and raising their kids here. And so we’re by far the most dynamic population, most aggressive, we’re the most aggressive set of characters, certainly in any Western country and have been for a long time, and certainly are today.

(00:03:23)
And then finally, I would just say, look, we are overwhelmingly the advanced technology leader. We have our issues and we have, I would say particular issue with manufacturing, which we could talk about. But for anything in software, anything in AI, anything in all these … Advanced biotech, all these advanced areas of technology, we’re by far the leader. Again, in part because many of the best scientists and engineers in those fields come to the US. And so we have all of the preconditions for just a monster boom. I could see economic growth going way up, I could see productivity growth going way up, rate of technology adoption going way up. And then we can do a global tour, if you like. But basically, all of our competitors have profound issues, and we could go through them one by one, but the competitive landscape just is … It’s like it’s remarkable how much better positioned we are for growth.
Lex Fridman
(00:04:13)
What about the humans themselves? Almost a philosophical question. I travel across the world and there’s something about the American spirit, the entrepreneurial spirit that’s uniquely intense in America. I don’t know what that is. I’ve talked to Saagar who claims it might be the Scots-Irish blood that runs through the history of America. What is it? You, at the heart of Silicon Valley, is there something in the water? Why is there this entrepreneurial spirit?
Marc Andreessen
(00:04:42)
Yeah. So is this a family show or am I allowed to swear?
Lex Fridman
(00:04:44)
You can say whatever the fuck you want.
Marc Andreessen
(00:04:46)
Okay. So the great TV show, succession, the show of course, which you were intended to root for exactly zero of the characters.
Lex Fridman
(00:04:47)
Yes.
Marc Andreessen
(00:04:53)
The best line from succession was in the final episode of the first season when the whole family’s over in Logan Roy’s ancestral homeland of Scotland. And they’re at this castle for some wedding. And Logan is just completely miserable because he’s been in New York for 50 years, he’s totally miserable being back in Scotland. And he gets in some argument with somebody and he says, finally just says, “My God, I cannot wait to get out of here and go back to America where we can fuck without condoms.”
Lex Fridman
(00:05:21)
Was that a metaphor or … Okay
Marc Andreessen
(00:05:23)
Exactly. No, but it’s exactly the thing, and everybody instantly knows what … Everybody watching that instantly starts laughing because you know what it means, which it’s exactly this. I think there’s an ethnographic way of it. There’s a bunch of books on, like you said, the Scots-Irish, like all the different derivations of all the different ethnic groups that have come to the US over the course of the last 400 years. But what we have is this sort of amalgamation of the Northeast Yankees who are super tough and hardcore. Yeah, the Scots-Irish are super aggressive. We’ve got the Southerners and the Texans and the whole blended kind of Anglo-Hispanic thing, super, incredibly tough, strong, driven, capable characters. The Texas Rangers, we’ve got the California, we’ve got the wild, we’ve got the incredibly inventive hippies, but we also have the hardcore engineers. We’ve got the best rocket scientists in the world. We’ve got the best artists in the world, creative professionals, the best movies.

(00:06:17)
So yeah, I would say all of our problems, I think are basically, in my view, to some extent, attempts to basically sand all that off and make everything basically boring and mediocre. But there is something in the national spirit that basically keeps bouncing back. And basically what we discover over time is we basically just need people to stand up at a certain point and say, “It’s time to build, it’s time to grow, it’s time to do things.” And there’s something in the American spirit that just roars right back to life. And I’ve seen it before. I saw it as a kid here in the early 80s because the 70s were horribly depressing in the US. They were a nightmare on many fronts. And in a lot of ways, the last decade to me has felt a lot like the 70s just being mired in misery and just this self-defeating negative attitude and everybody’s upset about everything. And then by the way, energy crisis and hostage crisis and foreign wars and just demoralization.

(00:07:17)
The low point in the 70s was Jimmy Carter who just passed away, he went on TV and he gave this speech known as the Malaise Speech, and it was like the weakest possible, trying to rouse people back to a sense of passion, completely failed. And we had the hostages in Iran for I think 440 days. And every night on the nightly news, it was lines around the block, energy crisis, depression, inflation. And then Reagan came in and Reagan was a very controversial character at the time. And he came in and he’s like, “Yep, nope, it’s morning in America and we’re the shining city on the hill, and we’re going to do it.” And he did it, and we did it. And the national spirit came roaring back and roared really hard for a full decade. I think that’s exactly what … I think we’ll see, but I think that’s what could happen here.
Lex Fridman
(00:07:57)
And I just did a super long podcast on Milton Friedman with Jennifer Burns, who’s this incredible professor at Stanford, and he was part of the Reagan … So there’s a bunch of components to that, one of which is economic, and one of which maybe you can put a word on it of not to be romantic or anything, but freedom, individual freedom, economic freedom, political freedom, and just in general, individualism.
Marc Andreessen
(00:08:22)
Yeah, that’s right. Yeah. As you know this, America has this incredible streak of individualism. Individualism in America probably peaked, I think between roughly call it the end of the Civil War, 1865 through to probably call it 1931 or something. And there was this incredible rush. I mean that period, we now know that period as the second Industrial Revolution, and it’s when the United States basically assumed global leadership and basically took over technological and economic leadership from England. And then that led to ultimately then therefore being able to not only industrialize the world, but also win World War II and then win the Cold War. And yeah, there’s a massive individualistic streak. By the way, Milton Friedman’s old videos are all on YouTube. They are every bit as compelling and inspiring as they were then. He’s a singular figure. And many of us, I never knew him, but he was actually at Stanford for many years at the Hoover Institution, but I never met him, but I know a lot of people who worked with him and he was a singular figure. But all of his lessons live on or are fully available.

(00:09:25)
But I would also say it’s not just individualism, and this is one of the big things. It’s playing out in a lot of our culture and kind of political fights right now, which is basically this feeling, certainly that I have and I share with a lot of people, which is it’s not enough for America to just be an economic zone, and it’s not enough for us to just be individuals, and it’s not enough to just have line go up, and it’s not enough to just have economic success. There are deeper questions at play, and also there’s more to a country than just that. And quite frankly, a lot of it is intangible. A lot of it involves spirit and passion. And like I said, we have more of it than anybody else, but we have to choose to want it.

(00:10:05)
The way I look at, it’s like all of our problems are self-inflicted. Decline is a choice. All of our problems are basically demoralization campaigns, basically people telling us, people in positions of authority telling us that, “We shouldn’t stand out, we shouldn’t be adventurous, we shouldn’t be exciting, we shouldn’t be exploratory, we shouldn’t this, that, and the other thing. And we should feel bad about everything that we do.” And I think we’ve lived through a decade where that’s been the prevailing theme. And I think quite honestly, as of November, I think people are done with it.

History of Western Civilization

Lex Fridman
(00:10:33)
If we could go on a tangent of a tangent, since we’re talking about individualism, and that’s not all that it takes. You’ve mentioned in the past the book The Ancient City by, if I could only pronounce the name, French historian Numa Denis Fustel de Coulanges. I don’t know.
Marc Andreessen
(00:10:48)
That was amazing.
Lex Fridman
(00:10:49)
Okay. All right. From the 19th century. Anyway, you said this is an important book to understand who we are and where we come from.
Marc Andreessen
(00:10:54)
So what that book does, it’s actually quite a striking book. So that book is written by this guy as a [inaudible 00:11:02] Let Lex do the pronunciations, the foreign language pronunciations for the day. He was a professor of classics at the Sorbonne in Paris, the top university, actually in the 1860s, so actually right around after the US Civil War. And he was a savant of a particular kind, which is he, and you can see this in the book is he had apparently read, and sort of absorbed and memorized every possible scrap of Greek and Roman literature. And so is like a walking index on basically everything we know about Greek and Roman culture, and that’s significant. The reason this matters is because basically none of that has changed. And so he had access to the exact same written materials that we have access to, and so we’ve learned nothing.

(00:11:41)
And then specifically what he did is he talked about the Greeks and the Romans, but specifically what he did is he went back further. He reconstructed the people who came before the Greeks and the Romans and what their life and society was like. And these were the people who were now known as the Indo-Europeans. And you may have heard of these, these are the people who came down from the steppes. And so they came out of what’s now Eastern Europe around sort of the outskirts of what’s now Russia. And then they sort of swept through Europe. They ultimately took over all of Europe, by the way, almost many of the ethnicities in the Americas, the hundreds of years that follow are Indo-European. And so they were basically this warrior, basically class that came down and swept through and essentially populated much of the world. And there’s a whole interesting saga there. And then from there came basically what we know as the Greeks and the Romans were kind of evolutions off of that.

(00:12:27)
And so what he reconstructs, what life was like, at least in the West for people in their kind of original social state. And the significance of that is the original social state is living in the state of the absolute imperative for survival with absolutely no technology. No modern systems, no nothing. You’ve got the clothes on your back, you’ve got whatever you can build with your bare hands. This predates basically all concepts of technology as we understand them today. And so these are people under maximum levels of physical survival pressure. And so what social patterns did they evolve to be able to do that? And the social pattern basically was as follows, is a three-part social structure, family, tribe and city, and zero concept of individual rights and essentially no concept of individualism. And so you were not an individual. You were a member of your family, and then a set of families would aggregate into a tribe and then a set of tribes would aggregate into a city.

(00:13:24)
And then the morality was completely … It was actually what Nietzsche talks about. The morality was entirely master morality, not slave morality. And so in their morality, anything that was strong was good, and anything that was weak was bad. And it’s very clear why that is. It is because strong equals good equals survive. Weak equals bad equals die. And that led to what became known later as the master-slave dialectic, which is, is it more important for you to live on your feet as a master even at the risk of dying? Or are you willing to live as a slave on your knees in order to not die? And this is sort of the derivation of that moral framework. Christianity later inverted that moral framework. But the original framework lasted for many, many thousands of years.

(00:14:01)
No concept of individualism. The head of the family had total life and death control over the family, the head of the tribe, same thing, head of the city, same thing. And then you were morally obligated to kill members of the other cities on contact. You were morally required to. If you didn’t do it, you were a bad person.

(00:14:16)
And then the form of the society was basically maximum fascism combined with maximum communism. And so it was maximum fascism in the form of this absolute top-down control where the head of the family tribe or city could kill other members of the community at any time with no repercussions at all. So maximum hierarchy, but combined with maximum communism, which is no market economy and so everything gets shared. And sort of the point of being in one of these collectives is that it’s a collective and people are sharing, and of course that limited how big they could get because the problem with communism is it doesn’t scale. It works at the level of a family. It’s much harder to make it work at the level of a country. Impossible. Maximum fascism, maximum communism.

(00:14:55)
And then it was all intricately tied into their religion. And their religion was in two parts. It was veneration of ancestors and it was veneration of nature. And the veneration of ancestors is extremely important because it was basically the ancestors were the people who got you to where you were. The ancestors were the people who had everything to teach you. And then it was veneration of nature because of course, nature is the thing that’s trying to kill you. And then you had your ancestor, every family, tribe or city had their ancestor gods and then they had their nature gods.

(00:15:25)
So fast-forward to today, we live in a world that is radically different, and the book takes you through what happened from that through the Greeks and Romans through to Christianity. But it’s very helpful to kind of think in these terms because the conventional view of the progress through time is that we are … The cliche is the arc of the moral universe bends towards justice or so-called wig history, which is that the arc of progress is positive. And so what you hear all the time, what you’re taught in school and everything is every year that goes by, we get better and better and more and more moral and more and more pure and a better version of ourselves. Our Indo-European ancestors would say, ” Oh no, you people have fallen to shit. You people took all of the principles of basically your civilization and you have diluted them down to the point where they barely even matter and you’re having children out of a wedlock and you regularly encounter people of other cities and you don’t try to kill them.” And how crazy is that?

(00:16:16)
And they would basically consider us to be living like an incredibly diluted version of this sort of highly religious, highly cult-like, highly organized, highly fascist, communist society. I can’t resist noting that as a consequence of basically going through all the transitions we’ve been through going all the way through Christianity coming out the other end of Christianity, Nietzsche declares God is dead. We’re in a secular society that still has tinges of Christianity, but largely prides itself on no longer being religious in that way. We being the sort of most fully evolved, modern secular experts, scientists and so forth have basically re-evolved or fallen back on the exact same religious structure that the Indo-Europeans had, specifically ancestor worship, which is identity politics and nature worship, which is environmentalism. And so we have actually worked our way all the way back to their cult religions without realizing it. And it just goes to show that in some ways we have fallen far from the family tree, but in some cases we’re exactly the same.
Lex Fridman
(00:17:16)
You kind of described this progressive idea of wokeism and so on as worshiping ancestors.
Marc Andreessen
(00:17:23)
Identity politics is worshiping ancestors. It’s tagging newborn infants with either benefits or responsibilities or levels of condemnation based on who their ancestors were. The Indo-Europeans would’ve recognized it on sight. We somehow think it’s super socially progressive.
Lex Fridman
(00:17:39)
And it is not.
Marc Andreessen
(00:17:41)
I mean, I would say obviously not. Get nuanced which is where I think you’re headed, which is, is the idea that you can completely reinvent society every and have no regard whatsoever for what came before you? That seems like a really bad idea. That’s like the Cambodians with Year Zero under Pol Pot and death follows. Obviously the Soviets tried that. The utopian fantasists who think that they can just rip up everything that came before and create something new in the human condition and human society have a very bad history of causing enormous destruction. So on the one hand, it’s like, okay, there is a deeply important role for tradition.

(00:18:14)
And the way I think about that is the process of evolutionary learning, which is what tradition ought to be, is the distilled wisdom of all. And this is what Indo-Europeans thought about. It should be the distilled wisdom of everybody who came before you. All those important and powerful lessons learned. And that’s why I think it’s fascinating to go back and study how these people lived is because part of the history and part of the learning that got us to where we’re today.

(00:18:36)
Having said that, there are many cultures around the world that are mired in tradition to the point of not being able to progress. And in fact, you might even say globally, that’s the default human condition, which is a lot of people are in societies in which there’s absolute seniority by age, kids are completely … In the US, for some reason we decided kids are in charge of everything and they’re the trendsetters and they’re allowed to set all the agendas and set all the politics and set all the culture and maybe that’s a little bit crazy. But in a lot of other cultures, kids have no voice at all, no role at all. The old people who are in charge of everything, they’re gerontocracies, and it’s all a bunch of 80 year olds running everything, which by the way, we have a little bit of that too.

(00:19:15)
And so what I would say is there’s a real downside. Full traditionalism is communitarianism, it’s ethnic particularism, it’s ethnic chauvinism, it’s this incredible level of resistance to change. It just doesn’t get you anywhere. It may be good and fine at the level of an individual tribe, but as a society living in the modern world, you can’t evolve, you can’t advance, you can’t participate in all the good things that have happened. And so I think probably this is one of those things where extremists on either side is probably a bad idea, but this needs to be approached in a sophisticated and nuanced way.

Trump in 2025

Lex Fridman
(00:19:52)
So the beautiful picture you painted of the roaring 20s, how can the Trump administration play a part in making that future happen?
Marc Andreessen
(00:20:00)
So look, a big part of this is getting the government boot off the neck of the American economy, the American technology industry, the American people. And again, this is a replay of what happened in the 60s and 70s, which is for what started out looking like, I’m sure good and virtuous purposes, we ended up both then and now with this, what I describe as sort of a form of soft authoritarianism. The good news is it’s not like a military dictatorship. It’s not like you get thrown into Lubyanka. For the most part, [inaudible 00:20:28] not coming at four in the morning. You’re not getting dragged off to a cell. So it’s not hard authoritarianism, but it is soft authoritarianism. And so it’s this incredible suppressive blanket of regulation rules, this concept of a vetocracy. What’s required to get anything done? You need to get 40 people to sign off on anything, any one of them can veto it. There’s a lot of [inaudible 00:20:47] political system works.

(00:20:49)
And then just this general idea of progress is bad, and technology is bad, and capitalism is bad, and building businesses is bad and success is bad. Tall poppy syndrome, basically, anybody who sticks their head up deserves to get it chopped off. Anybody who’s wrong about anything deserves to get condemned forever. Just this very kind of grinding repression. And then coupled with specific government actions such as censorship regimes and debanking and Draconian, deliberately kneecapping critical American industries, and then congratulating yourselves on the back for doing it or having these horrible social policies, like let’s let all the criminals out of jail and see what happens. And so we’ve just been through this period, I call it a demoralization campaign. We’ve just been through this period, whether it started that way or not, it ended up basically being this comprehensive message that says, “You’re terrible and if you try to do anything, you’re terrible and fuck you.” And the Biden administration reached the full pinnacle of that in our time. They got really bad on many fronts at the same time. And so just relieving that and getting back to a reasonably optimistic, constructive, pro-growth frame of mind, there’s so much pent-up energy and potentially in the American system, that alone is going to, I think cause growth and spirit to take off. And then there’s a lot of things proactively that could be done.
Lex Fridman
(00:22:13)
So how do you relieve that? To what degree has the thing you describe ideologically permeated government and permeated big companies?
Marc Andreessen
(00:22:23)
Disclaimer at first, which is I don’t want to predict anything on any of this stuff because I’ve learned the hard way that I can’t predict politics or Washington at all. But I would just say that the plans and intentions are clear and the staffing supports it, and all the conversations are consistent with the due administration and that they plan to take very rapid action on a lot of these fronts very quickly. They’re going to do as much as they can through executive orders, and then they’re going to do legislation and regulatory changes for the rest. And so they’re going to move, I think, quickly on a whole bunch of stuff. You can already feel, I think a shift in the national spirit, or at least, let’s put it this way, I feel it for sure in Silicon Valley. I mean, we just saw a great example of this with what Mark Zuckerberg is doing, and obviously I’m involved with his company, but we just saw it kind of in public, the scope and speed of the changes are reflective of a lot of these shifts.

(00:23:08)
But I would say that same conversation, those same kinds of things are happening throughout the industry. And so the tech industry itself, whether people were pro-Trump or anti-Trump, there’s just a giant vibe shift, mood shift that’s kicked in already. And then I was with a group of Hollywood people about two weeks ago, and they were still people who at least vocally were still very anti-Trump, but I said, “Has anything changed since November 6th?” And they immediately said, “Oh, it’s completely different. It feels like the ice has thawed. Woke is over.” They said that all kinds of projects are going to be able to get made now they couldn’t before, that Hollywood’s going to start making comedies again. It is just like an incredible immediate environmental change. And as I talk to people, certainly throughout the economy, people who run businesses, I hear that all the time, which is just this last 10 years of misery is just over.

(00:23:57)
I mean, the one that I’m watching that’s really funny. I mean, Facebook’s getting a lot, Meta’s getting a lot of attention, but the other funny one is BlackRock, which I don’t know him, but I’ve watched for a long time. And so Larry Fink, who’s the CEO of BlackRock, was first in as a major investment CEO on every dumb social trend and rule set every … I’m going for it. Every retarded thing you can imagine, every ESG and every possible … Saddling companies with every aspect of just these crazed ideological positions. And he was coming in, he literally had aggregated together trillions of dollars of shareholdings that were his customer’s rights, and he seized their voting control of their shares and was using it to force all these companies to do all of this crazy ideological stuff. And he was like the Typhoid Mary of all this stuff in corporate America. And he in the last year has been backpedaling from that stuff as fast as he possibly can.

(00:24:55)
And just an example, last week, he pulled out of the whatever the Corporate Net-Zero Alliance, he pulled out of the crazy energy stuff. And so he’s backing away as fast as he can. Remember, the Richard Pryor backwards walk? Richard Pryor had this way where he could back out of a room while looking like he was walking forward.

(00:25:11)
And so even there doing that and just the whole thing. I mean, if you saw the court recently ruled that NASDAQ had these crazy board of directors composition rules. One of the funniest moments of my life is when my friend Peter Thiel and I were on the Meta board and these NASDAQ rules came down, mandated diversity on corporate boards. And so we sat around the table and had to figure out which of us counted as diverse. And the very professional attorneys at Meta explained with 100% complete straight face that Peter Thiel counts as diverse by virtue of being LGBT. And this is a guy who literally wrote a book called The Diversity Myth. He literally looked like he’d swallowed a live goldfish, and this was imposed. I mean, this was so incredibly offensive to him that it was just absolutely appalling and I felt terrible for him. But the look in his face was very funny.

(00:26:03)
And it was imposed by NASDAQ, your stock exchange imposing this stuff on you, and then the court, whatever, the Court of Appeals just nuked that. So these things basically are being ripped down one by one. And what’s on the other side of it is basically finally being able to get back to everything that everybody always wanted to do, which is run their companies, have great products, have happy customers, succeed, achieve, outperform, and work with the best and the brightest and not be made to feel bad about it. And I think that’s happening in many areas of American society.
Lex Fridman
(00:26:34)
It’s great to hear that Peter Thiel is fundamentally a diversity hire.
Marc Andreessen
(00:26:38)
Well, there was a moment. So Peter, of course, is publicly gay, has been for a long time, but there are other men on the board, and we’re sitting there and we’re all looking at it, and we’re like, all right, okay, LGBT, and we keep coming back to the B, and it’s like all I’m willing to do a lot for this company, but …
Lex Fridman
(00:27:05)
It’s all about sacrifice for diversity.
Marc Andreessen
(00:27:08)
Well, yeah. And then it’s like, okay, is there a test?
Lex Fridman
(00:27:13)
Oh yeah, exactly. How do you prove it?
Marc Andreessen
(00:27:15)
The questions that got asked.
Lex Fridman
(00:27:18)
What are you willing to do for the greater good?
Marc Andreessen
(00:27:20)
I’ve become very good at asking lawyers completely absurd questions with a totally straight face.
Lex Fridman
(00:27:26)
And do they answer with a straight face [inaudible 00:27:29]?
Marc Andreessen
(00:27:28)
Sometimes. I think in fairness, they have trouble telling when I’m joking.

TDS in tech

Lex Fridman
(00:27:32)
So you mentioned the Hollywood folks, maybe people in Silicon Valley and the vibe shift. Maybe you can speak to preference falsification. What do they actually believe? How many of them actually hate Trump? What percent of them are feeling this vibe shift and are interested in creating the roaring 20s in the way they’ve described?
Marc Andreessen
(00:27:57)
So first we should maybe talk population. So there’s all of Silicon Valley, and the way to just measure that is just look at voting records and what that shows consistently is Silicon Valley is just, at least historically, my entire time there has been overwhelmingly majority just straight up Democrat. The other way to look at that is political donation records. And again, the political donations in the Valley range from 90 to 99% to 1 side. And so I just bring it up because we’ll see what happens with the voting and with donations going forward.

(00:28:25)
We can maybe talk about the fire later, but I can tell you there is a very big question of what’s happening in Los Angeles right now. I don’t want to get into the fire, but it’s catastrophic. And there was already a rightward shift in the big cities in California, and I think a lot of people in LA are really thinking about things right now as they’re trying to literally save their houses and save their families. But even in San Francisco, there was a big shift to the right in the voting in ’24. So we’ll see where that goes, but you observe that by just looking at the numbers over time.

(00:28:55)
The part that I’m more focused on is, and I don’t know how to exactly describe this, but it’s like the top 1,000 or the top 10,000 people. I don’t have a list, but it’s all the top founders, top CEOs, top executives, top engineers, top VCs, and then into the ranks, the people who kind of built and run the companies. And I don’t have numbers, but I have a much more tactile feel for what’s happening. So the big thing I have now come to believe is that the idea that people have beliefs is mostly wrong. I think that most people just go along, and I think even most high status people just go along. And I think maybe the most high status people are the most prone to just go along because they’re the most focused on status. And the way I would describe that is one of the great forbidden philosophers of our time is the Unabomber, Ted Kaczynski. And amidst his madness, he had this extremely interesting articulation. He was an insane lunatic murderer, but he was also a Harvard super genius. Not that those are in conflict.
Lex Fridman
(00:30:03)
Shots fired man.
Marc Andreessen
(00:30:04)
But he was a very bright guy, and he did this whole thing where he talked about, basically he was very right-wing and talked about leftism a lot. And he had this great concept that’s just stuck in my mind ever since I read it, which is he had this concept just called over-socialization. And so most people are socialized. We live in a society, most people learn how to be part of a society. They give some deference to the society. There’s something about modern Western elites where they’re over-socialized and they’re just overly oriented towards what other people like themselves think and believe. And you can get a real sense of that if you have a little bit of an outside perspective, which I just do, I think as a consequence of where I grew up.

(00:30:47)
Even before I had the views that I have today, there was always just this weird thing where it’s like, why does every dinner party have the exact same conversation? Why does everybody agree on every single issue? Why is that agreement precisely what was in the New York Times today? Why are these positions not the same as they were five years ago? But why does everybody snap into agreement every step of the way? And that was true when I came to Silicon Valley, and it’s just as true today 30 years later. And so I think most people are just literally, I think they’re taking their cues from, it’s some combination of the press, the universities, the big foundations. So it’s basically, it’s like The New York Times, Harvard, the Ford Foundation, and I don’t know, a few CEOs and a few public figures and maybe the President if your party is in power. And whatever that is, everybody who’s sort of good and proper and elite and good standing and in charge of things, and a sort of correct member of, let’s call it coastal American society, everybody just believes those things.

(00:31:45)
And then the two interesting things about that is, number one, there’s no divergence among the organs of power. So Harvard and Yale believed the exact same thing. The New York Times and The Washington Post believe the exact same thing. The Ford Foundation and the Rockefeller Foundation believe the exact same thing. Google and whatever, Microsoft believe the exact same thing. But those things change over time, but there’s never conflict in the moment. And so The New York Times and The Washington Post agreed on exactly everything in 1970, 1980, 1990, 2000, 2010, and 2020, despite the fact that the specifics changed radically. The lockstep was what mattered. And so I think basically we in the Valley we’re on the tail end of that, in the same way Hollywood’s on the tail end of that, in the same way New York’s on the tail end of that, the same way the media’s on the tail end of that. It’s like some sort of collective hive mind thing.

(00:32:33)
And I just go through that to say, I don’t think most people in my orbit, or let’s say the top 10,000 people in the Valley or the top 10,000 people in LA, I don’t think they’re sitting there thinking basically, I have rock … I mean, they probably think they have rocks solid beliefs, but they don’t actually have some inner core of rock solid beliefs. And then they kind of watch reality change around them and try to figure out how to keep their beliefs correct. I don’t think that’s what happens. I think what happens is they conform to the belief system around them, and I think most of the time they’re not even aware that they’re basically part of a herd.
Lex Fridman
(00:33:01)
Is it possible that the surface chatter …
Marc Andreessen
(00:33:00)
That they’re basically part of a herd.
Lex Fridman
(00:33:01)
Is it possible that the surface chatter of dinner parties, underneath that there is a turmoil of ideas and thoughts and beliefs that’s going on, but you’re just talking to people really close to you or in your own mind, and the socialization happens at the dinner parties? When you go outside the inner circle of one, two, three, four people who you really trust, then you start to conform. But inside there, inside the mind, there is an actual belief or a struggle, attention within New York Times or with the listener. For the listener, there’s a slow smile that overtook Mark Andreessen’s face.
Marc Andreessen
(00:33:41)
So look, I’ll just tell you what I think, which is at the dinner parties and at the conferences, no, there’s none of that. What there is that all of the heretical conversations, anything that challenges the status quo, any heretical ideas and any new idea is a heretical idea, any deviation is either discussed one-on-one, face-to-face, it’s like a whisper network or it’s like a real life social network. There’s a secret handshake. Which is like, okay, you meet somebody and each other a little bit, but not well, and you’re both trying to figure out if you can talk to the other person openly or whether you have to be fully conformist. It’s a joke.
Lex Fridman
(00:34:15)
Well, yeah, humor 100%.
Marc Andreessen
(00:34:16)
Somebody cracks a joke, right? Somebody cracks a joke. If the other person laughs, the conversation is on.
Lex Fridman
(00:34:21)
Yeah.
Marc Andreessen
(00:34:22)
If the other person doesn’t laugh, back slowly away from the scene, I didn’t mean anything by it. And then by the way, it doesn’t have to be a super offensive joke. It just has to be a joke that’s just up against the edge of one of the, use the Sam Bankman-Fried term, one of the chivalrous. It has to be up against one of the things, one of the things that you’re absolutely required to believe to be the dinner parties. And then at that point, what happens is have a peer-to-peer network. You have a one-to-one connection with somebody, then you have your little conspiracy of thought criminality, and then you’ve probably been through this, you have your network of thought criminals, and then they have their network of thought criminals, and then you have this delicate mating dance as to whether you should bring the thought criminals together.
Lex Fridman
(00:35:05)
And the fundamental mechanism of the dance is humor.
Marc Andreessen
(00:35:09)
Yeah, it’s humor. Well, of course.
Lex Fridman
(00:35:09)
Memes. Yeah.
Marc Andreessen
(00:35:10)
Well, for two reasons. Number one, humor is a way to have deniability, right? Humor is a way to discuss serious things without having deniability. Oh, I’m sorry. It was just a joke. So that’s part of it, which is one of the reasons why comedians can get away with saying things the rest of us can’t, they can always fall back on, oh yeah, I was just going for the laugh. But the other key thing about humor is that laughter is involuntary. You either laugh or you don’t. And it’s not a conscious decision whether you’re going to laugh. And everybody can tell when somebody’s fake laughing and every professional comedian knows this. The laughter is the clue that you’re onto something truthful.

(00:35:41)
People don’t laugh at made-up bullshit stories. They laugh because you’re revealing something that they either have not been allowed to think about or have not been allowed to talk about or is off limits. And all of a sudden it’s like the ice breaks and it’s like, oh yeah, that’s the thing. And it’s funny and I laugh, and then of course, this is why of course live comedy is so powerful is because you’re all doing that at the same time, so you start to have the safety of numbers. It’s no surprise to me, for example, Joe has been as successful as he has because they have this hack that the rest of us who are not professional comedians don’t have, but you have your in-person version of it, and then you’ve got the question of whether you can join the networks together.

(00:36:17)
And then you’ve probably been to this, is then at some point there’s like the Alt dinner party, the [inaudible 00:36:23] dinner party, and you get six or eight people together and you join the networks. And those are the happiest, at least in the last decade, those are the happiest moments of everybody’s lives. Everybody’s just ecstatic because they’re like, I don’t have to worry about getting yelled at and shamed for every third sentence that comes out of my mouth, and we can actually talk about real things. So that’s the live version of it. And then of course, the other side of it’s the group chat phenomenon. And then basically the same thing played out until Elon bought X and until Substack took off, which were really the two big breakthroughs in free speech online, the same dynamic played out online, which is you had absolute conformity on the social networks, literally enforced by the social networks themselves through censorship, and then also through cancellation campaigns and mobbing and shaming.

(00:37:05)
But then group chats grew up to be the equivalent of Samizdat. Anybody who grew up in the Soviet Union under communism, note, they had the hard version of this. It’s like, how do you know who you could talk to? And then how do you distribute information? And again, that was the hard authoritarian version of this. And then we’ve been living through this weird mutant soft authoritarian version, but with some of the same patterns.
Lex Fridman
(00:37:26)
And WhatsApp allows you to scale and make it more efficient to build on these groups of heretical ideas bonded by humor.
Marc Andreessen
(00:37:36)
Yeah, exactly. Well, and this is the thing, and well, this is the running kind of thing about group chats. It’s not even a joke. It’s true. If you’ve noticed this principle of group chats, every group chat ends up being about memes and humor. And the game of group chat is to get as close to the line of being actually objectionable as you can get without actually tripping it. And literally every group chat that I have been in for the last decade, even if it starts some other direction, what ends up happening is it becomes the absolute comedy fest where, butt they walk right up the line and they’re constantly testing. And every once in a while somebody will trip the line and people will freak out. And it’s like, oh, too soon. Okay, we got to wait until next year to talk about that. They walk it back.

(00:38:17)
And so it’s that same thing. And then group chats is a technological phenomenon. It was amazing to see. Number one, it was obviously the rise of smartphones, then it was the rise of the new messaging services, then it was the rise specifically of I would say combination of WhatsApp and Signal. And the reason for that is those were the two big systems that did the full encryption, so you actually felt safe. And then the real breakthrough I think was disappearing messages, which hit Signal probably four or five years ago and hit WhatsApp three or four years ago. And then the combination of encryption and disappearing messages I think really unleashed it. Well, then there’s the fight over the length of disappearing messages. And so it’s like I often get behind on my thing, so I set to seven day disappearing messages and my friends who are like, no, that’s way too much risk. It’s got to be a day. And then every once in a while somebody will set to five minutes before they send something particularly inflammatory.
Lex Fridman
(00:39:12)
Yeah, 100%. One of the things that bothers me about WhatsApp, the choice is between 24 hours and seven days, one day or seven days. And I have to have an existential crisis deciding whether I can last for seven days with what I’m about to say.
Marc Andreessen
(00:39:29)
Exactly. Now, of course, what’s happening right now is the big thaw. The vibe shift. So what’s happening on the other side of the election is Elon on Twitter two years ago and now Mark with Facebook and Instagram. And by the way, with the continued growth of Substack and with other new platforms that are emerging, I think it may be, I don’t know that everything just shifts back into public, but a tremendous amount of the verboten conversations can now shift back into public view. And this is one of those things, quite frankly, even if I was opposed to what people are saying, and I’m sure I am in some cases, I would argue still net better for society that those things happen in public instead of private. Does you want to know? And then look, it’s just I think, clearly much healthier to live in a society in which people are not literally scared of what they’re saying.

Preference falsification

Lex Fridman
(00:40:19)
I mean, to push back, to come back to this idea that we’re talking about, I do believe that people have beliefs and thoughts that are heretical, like a lot of people. I wonder what fraction of people have that? To me, the preference falsification is really interesting. What is the landscape of ideas that human civilization has in private as compared to what’s out in public? Because the dynamical system that is the difference between those two is fascinating. Throughout history the fall of communism and multiple regimes throughout Europe is really interesting. Everybody was following the line until not. But for sure, privately, there was a huge number of boiling conversations happening, where this is the bureaucracy of communism, the corruption of communism, all of that was really bothering people more and more and more and more. And all of a sudden there’s a trigger that allows the vibe shift to happen.

(00:41:22)
To me, the interesting question here is, what is the landscape of private thoughts and ideas and conversations that are happening under the surface of Americans? Especially, my question is how much dormant energy is there for this roaring twenties? What people are like, no more bullshit, let’s get done.
Marc Andreessen
(00:41:43)
So we’ll go through the theory of preference falsification just to-
Lex Fridman
(00:41:47)
Yes. By the way, amazing. The books on this is fascinating.
Marc Andreessen
(00:41:49)
Yeah, yeah. So this is one of the all time great books. Incredible. About 20, 30-year-old book, but it’s completely modern and current in what it talks about as well as very deeply historically informed. So it’s called Private Truths, Public Lies, and it’s written by a social science professor named Timur Kuran, at I think Duke, and his definitive work on this. And so he has this concept, he calls Preference Falsification. And so preference falsification is two things, and you get it from the title of the book, Private Truths, Public Lies. So preference falsification is when you believe something and you can’t say it, or, and this is very important, you don’t believe something and you must say it. And the commonality there is in both cases, you’re lying. You believe something internally, and then you’re lying about it in public.

(00:42:36)
And there’s the two classic forms of it. For example, there’s the, I believe communism is rotten, but I can’t say it version of it. But then there’s also the famous parable of the real life example, but the thing that Vaclav Havel talks about in the other good book on this topic, which is The Power of the Powerless, who is an anti-communist resistance fighter who ultimately became the president of Czechoslovakia after the fall of the wall. But he wrote this book and he describes the other side of this, which is workers of the world unite. And so he describes what he calls the Parable of the Greengrocer, which is you’re a greengrocer in Prague in 1985, and for the last 50 years, it’s been absolutely mandatory to have a sign in the window of your store that says Workers of the World Unite.

(00:43:22)
And it’s 1985, it is crystal clear that the workers of the world are not going to unite. Of all the things that could happen in the world, that is not going to happen. The Commies have been at that for 70 years, it is not happening. But that slogan had better be in your window every morning because if it’s not in your window every morning, you are not a good communist. The secret police are going to come by and they’re going to get you. And so the first thing you do when you get to the store is you put that slogan in the window and you make sure that it stays in the window all day long. But he says, the thing is, the greengrocer knows the slogan is fake. He knows it’s a lie. Every single person walking past the slogan knows that it’s a lie. Every single person walking past the store knows that the greengrocer is only putting it up there because he has to lie in public. And the greengrocer has to go through the humiliation of knowing that everybody knows that he’s caving into the system and lying in public.

(00:44:07)
And so it turns into the moralization campaign. In fact, it’s not ideological enforcement anymore because everybody knows it’s fake. The authorities know it’s fake, everybody knows it’s fake. It’s not that they’re enforcing the actual ideology of the workers of the world uniting. It’s that they’re enforcing compliance and compliance with the regime. And you fuck you, you will comply. And so anyway, that’s the other side of that. And of course, we have lived in the last decade through a lot of both of those. I think anybody listening to this could name a series of slogans that we’ve all been forced to chant for the last decade that everybody knows at this point are just simply not true. I’ll let the audience speculate on their own group chats.
Lex Fridman
(00:44:50)
Send Marc your memes online as well, please.
Marc Andreessen
(00:44:52)
Yes, yes, exactly. Okay. So anyway, so it’s the two sides of that, right? So it’s Private Truth, Public Lies. So then what preference falsification does is it talks about extending that from the idea of the individual experience in that to the idea of the entire society experiencing that, right? And this gets to your percentages question. Which is like, okay, what happens in a society in which people are forced to lie in public about what they truly believe? What happens number one is that individually they’re lying in public and that’s bad. But the other thing that happens is they no longer have an accurate gauge at all or any way to estimate how many people agree with them. And again, this literally is how you get something like the communist system, which is like, okay, you end up in a situation in which 80 or 90 or 99% of a society can actually all be thinking individually, I really don’t buy this anymore.

(00:45:34)
And if anybody would just stand up and say it, I would be willing to go along with it, but I’m not going to be the first one to put my head on the chopping block. But because of the suppression censorship, you have no way of knowing how many of the people agree with you. And if the people agree with you are 10% of the population and you become part of a movement, you’re going to get killed. If 90% of the people agree with you, you’re going to win the revolution. And so the question of what the percentage actually is is a really critical question. And then basically in any sort of authoritarian system, you can’t run a survey to get an accurate result. And so you actually can’t know until you put it to the test. And then what he describes in the book is it’s always put to the test in the same way.

(00:46:10)
This is exactly what’s happened for the last two years, like 100% of exactly what’s happened. It’s like straight out of this book. Which is somebody, Elon, sticks his hand up and says, the workers of the world are not going to unite. Or the emperor is actually wearing no clothes, that famous parable. So one person stands up and does it, and literally that person is standing there by themselves, and everybody else in the audience is like, Ooh, I wonder what’s going to happen to that guy? But again, nobody knows. Elon doesn’t know, the first guy doesn’t know, other people don’t know which way is this going to go. And it may be that that’s a minority position and that’s the way to get yourself killed. Or it may be that that’s the majority position and you are now the leader of a revolution.

(00:46:49)
And then basically, of course, what happens is, okay, the first guy does that doesn’t get killed, the second guy does… Well, a lot of the time that guy does get killed, but when the guy doesn’t get killed, then a second guy pops his head up, says the same thing. All right, now you’ve got two. Two leads to four, four leads to eight, eight leads to 16. And then as we saw with the fall of the Berlin Wall, this is what happened in Russia and Eastern Europe in ’89, when it goes, it can go, and then it rips. And then if it turns out that you had a large percentage of the population that actually believed the different thing, it turns out all of a sudden everybody has this giant epiphany that says, oh, I’m actually part of the majority. And at that point, you were on the freight train to revolution, right? It is rolling. Now, the other part of this is the distinction between the role of the elites and the masses.

(00:47:34)
And here the best book is called the True Believer, which is the Eric Hoffer book. And so the nuance you have to put on this is the elites play a giant role in this because the elites do idea formation and communication, but the elites, by definition are a small minority. And so there’s also this giant role played by the masses, and the masses are not necessarily thinking these things through in the same intellectualized formal way that the elites are, but they are for sure experiencing these things in their daily lives, and they for sure have at least very strong emotional views on them.

(00:48:01)
And so when you really get the revolution, it’s when you get the elites lined up with, or either the current elites change or the new set of elites, a new set of counter elites, basically come along and say, “No, there’s actually a different and better way to live.” And then the people basically decide to follow the counter elite. So that’s the other dimension to it. And of course, that part is also happening right now. And again, case study one of that would be Elon, and who turns out in truly massive following.
Lex Fridman
(00:48:26)
And he has done that over and over in different industries, not just saying crazy shit online, but saying crazy shit in the realm of space, in the realm of autonomous driving, in the realm of AI, just over and over and over again. Turns out saying crazy shit is one of the ways to do a revolution and to actually make progress.
Marc Andreessen
(00:48:43)
Yeah. And it’s like, well, but then there’s the test. Is it crazy or is it the truth?
Lex Fridman
(00:48:46)
Yeah.
Marc Andreessen
(00:48:49)
And this is where there are many specific things about Elon’s genius, but one of the really core ones is an absolute dedication to the truth. And so when Elon says something, it sounds like crazy shit, but in his mind it’s true. Now, is he always right? No. Sometimes the rockets crash, sometimes he’s wrong. He’s human, he’s like anybody else. He’s not right all the time. But at least my through line with him both in what he says in public and what he says in private, which by the way are the exact same things. He does not do this. He doesn’t lie in public about what he believes in private, or at least he doesn’t do that anymore. He’s 100% consistent in my experience. By the way, there’s two guys who are 100% consistent like that that I know. Elon and Trump.

(00:49:26)
Whatever you think of them, what they say in private is 100% identical to what they say in public. They’re completely transparent, they’re completely honest in that way. Again, it’s not like they’re perfect people, but they’re honest in that way. And it makes them potentially both as they have been, very powerful leaders of these movements because they’re both willing to stand up and say the thing that if it’s true, it turns out to be the thing in many cases that many or most or almost everyone else actually believes, but nobody was actually willing to say out loud. And so they can actually catalyze these shifts. I think this framework is exactly why Trump took over the Republican Party. I think Trump stood up there on stage with all these other kind of conventional Republicans, and he started saying things out loud that it turned out the base really was they were either already believing or they were prone to believe, and he was the only one who was saying them.

(00:50:06)
And so again, elite masses, he was the elite, the voters of the masses, and the voters decided, no. No more bushes, we’re going this other direction. That’s the mechanism of social change. What we just described is the actual mechanism of social change. It is fascinating to me that we have been living through exactly this. We’ve been living through everything exactly what Timur Kuran describes, everything that Vaclav Havel described. Black Squares and Instagram, like the whole thing, right? All of it. And we’ve been living through the true believer elites masses thing, with a set of basically incredibly corrupt elites wondering why they don’t have the masses anymore, and a set of new elites that are running away with things.

(00:50:43)
And so we’re living through this incredible applied case study of these ideas. And if there’s a moral of the story, it is I think fairly obvious, which is it’s a really bad idea for a society to wedge itself into a position in which most people don’t believe the fundamental precepts of what they’re told they have to do to be good people like that. That is just not a good state to be in.
Lex Fridman
(00:51:03)
So one of the ways to avoid that in the future maybe, is to keep the delta between what’s said in private and what’s said in public small.
Marc Andreessen
(00:51:10)
Yeah. Well, this is sort of the siren song of censorship is we can keep people from saying things, which means we can keep people from thinking things.

(00:51:17)
And by the way, that may work for a while. I mean, again, the hard form, Soviet Union, pre photocopiers, there were mimeograph machines that were used to make Samizdat and underground newspapers, which is the mechanism of written communication of radical ideas, radical ideas. Ownership of a mimeograph machine was punishable by death. So that’s the hard version. The soft version is somebody clicks a button in Washington and you were erased from the internet, which good news, you’re still alive. Bad news is, shame about not being able to get a job. Too bad your family now hates you and won’t talk to you, whatever the version of cancellation it’s been. And so does that work? Maybe it works for a while. It worked for the Soviet Union for a while in its way, especially when it was coupled with official state power. But when it unwinds, it can unwind with incredible speed and ferocity. Because to your point, there’s all this bottled up energy.

(00:52:12)
Now, your question was what are the percentages? What’s the breakdown? And so my rough guess, just based on what I’ve seen in my world, is it’s something like 20, 60, 20. It’s like you’ve got 20% true believers in whatever is the current thing. You got 20% of people who are just true believers of whatever’s in the New York Times, Harvard professors and the Ford Foundation, they’re just… Maybe it’s 10, maybe it’s five, but let’s say generously it’s 20. So 20% kind of full-on revolutionaries. And then you’ve got, let’s call it 20% on the other side that are like, no, I’m not on board with this. This is crazy. I’m not signing up for this. But their view of themselves is they’re in a small minority, and in fact, they start out in a small minority, because what happens is the 60% go with the first 20%, not the second 20%. So you’ve got this large middle of people.

(00:53:03)
And it’s not that the people in the middle are not smart or anything like that, that they just have normal lives and they’re just trying to get by and they’re just trying to go to work each day and do a good job and be a good person and raise their kids and have a little bit of time to watch the game, and they’re just not engaged in the cut and thrust of political activism or any of this stuff. It’s just not their thing. But that’s where the over socialization comes in. It’s just like, okay, by default, the 60% will go along with the 20% of the radical revolutionaries at least for a while, and then the counter elite is in this other 20%. And over time, they build up a theory and network and ability to resist in a new set of representatives, in a new set of ideas. And then at some point there’s a contest and then, and then the question is, what happens in the middle? What happens in the 60%?

(00:53:51)
And it is kind of my point. It’s not even really does the 60% change their beliefs as much as it’s like, okay, what is the thing that that 60% now decides to basically fall into step with? And in the valley, that 60% for the last decade decided to be woke and extremely, I would say, on edge on a lot of things. And that 60% is pivoting in real time. They’re just done. They’ve just had it.
Lex Fridman
(00:54:17)
And I would love to see where that pivot goes because there’s internal battles happening right now.
Marc Andreessen
(00:54:24)
So this is the other thing. So there’s two forms of things, and Timur has actually talked about this, Professor Kuran has talked about this. So one is he said, this is the kind of unwind where what you’re going to have is you’re now going to have people in the other direction. You’re going to have people who claim that they supported Trump all along, who actually didn’t, right? So it’s going to swing the other way.

(00:54:42)
And by the way, Trump’s not the only part of this, but he’s just a convenient shorthand for a lot of this. But whatever it is, you’ll have people who will say, well, I never supported the EI, or I never supported ESG, or I never thought we should have canceled that person, where of course, they were full on a part of the mob at that moment. So anyway, so you’ll have preference falsification happening in the other direction. His prediction, I think, basically is you’ll end up with the same quote, “problem” on the other side. Now, will that happen here? I don’t know. How far is American society willing to go on any of these things? I don’t know. But there is some question there.

(00:55:15)
And then the other part of it is, okay, now you have this elite that is used to being in power for the last decade. And by the way, many of those people are still in power and they’re in very important positions. And the New York Times is still the New York Times, and Harvard is still Harvard, and those people haven’t changed, like at all. Bureaucrats in the government and senior democratic politicians and so forth. And they’re sitting there right now feeling like reality has just smacked them hard in the face because they lost the election so badly. But they’re now going into, and specifically the Democratic Party, is going into a Civil War. And that form of the Civil War is completely predictable and it’s exactly what’s happening, which is half of them are saying, we need to go back to the center. We need to de-radicalize because we’ve lost the people. We’ve lost the people in the middle, and so we need to go back to the middle in order to be able to get 50% plus one in an election.

(00:56:03)
And then the other half of them are saying, no, we weren’t true to our principles. We were too weak. We were too soft. We must become more revolutionary. We must double down and we must celebrate murders in the street of health insurance executives. And that right now is a real fight.

Self-censorship

Lex Fridman
(00:56:15)
If I could tell you a little personal story that breaks my heart a little bit, there’s a professor, a historian, I won’t say who, who I admire deeply, love his work. He’s kind of a heretical thinker. And we were talking about having a podcast, on doing a podcast, and he eventually said that, “You know what, at this time, given your guest list, I just don’t want the headache of being in the faculty meetings in my particular institution.” And I asked, “Who are the particular figures in this guest list?” He said, “Trump.” And the second one, he said, “That you announced your interest to talk to Vladimir Putin.” So I just don’t want the headache. Now, I fully believe it would surprise a lot of people if I said who it is. This is a person who’s not bothered by the guest list. And I should also say that 80 plus percent of the guest list is left wing.

(00:57:20)
Nevertheless, he just doesn’t want the headache. And that speaks to the thing that you’ve kind of mentioned, that you just don’t want the headache. You just want to just have a pleasant morning with some coffee and talk to your fellow professors. And I think a lot of people are feeling that in universities and in other contexts, in tech companies. And I wonder if that shifts, how quickly that shifts? And there, the percentages you mentioned, 20, 60, 20 matters, and the contents of the private groups matters, and the dynamics of how that shifts matters. Because it’s very possible, nothing really changes in universities and in major tech companies. Where just, there’s a kind of excitement right now for potential revolution and these new ideas, these new vibes, to reverberate through these companies and universities, but it’s possible the wall will hold.
Marc Andreessen
(00:58:14)
Yeah. So he’s a friend of yours, I respect that you don’t want to name him. I also respect you don’t want to beat on him, so I would like to beat on him on your behalf. Does he have tenure?
Lex Fridman
(00:58:23)
Yes. He should use it.
Marc Andreessen
(00:58:27)
So this is the thing. This is the ultimate indictment of the corruption and the rot at the heart of our education system, at the heart of these universities. And it’s, by the way, it’s across the board. It’s all the top universities. Because the siren song for what it’s been for 70 years, whatever, of the tenure system, peer review system, tenure system, which is like, yeah, you work your butt off as an academic to get a professorship and then to get tenure, because then you can say what you actually think. Then you can do your work and your research and your speaking and your teaching without fear of being fired. Without fear of being canceled. Academic freedom. I mean, think of the term academic freedom, and then think of what these people have done to it. It’s gone. That entire thing was fake and is completely rotten. And these people are completely giving up the entire moral foundation of the system that’s been built for them, which by the way, is paid for virtually 100% by taxpayer money.
Lex Fridman
(00:59:34)
What’s the inkling of hope in this? This particular person and others who hear this, what can give them strength, inspiration, and courage?
Marc Andreessen
(00:59:44)
That the population at large is going to realize the corruption in their industry and it’s going to withdraw the funding.
Lex Fridman
(00:59:49)
Okay, so desperation.
Marc Andreessen
(00:59:51)
No, no, no, no, no. Think about what happens next. Okay, so let’s go through it. So the universities are funded by four primary sources of federal funding. The big one is a federal student loan program, which is in the many trillions of dollars at this point, and then only spiraling way faster than inflation. That’s number one. Number two is federal research funding, which is also very large. And you probably know that when a scientist at the university gets a research grant, the university rakes as much as 70% of the money for central uses.
Lex Fridman
(01:00:20)
Yeah.
Marc Andreessen
(01:00:20)
Number three is tax exemption at the operating level, which is based on the idea that these are nonprofit institutions as opposed to, let’s say, political institutions. And then number four is tax exemptions at the endowment level, which is the financial buffer that these places have. Anybody who’s been close to a university budget will basically see that what would happen if you withdrew those sources of federal taxpayer money, and then for the state schools, the state money, they all instantly go bankrupt. And then you could rebuild. Then you could rebuild. Because the problem right now, the folks at University of Austin are mounting a very valiant effort, and I hope that they succeed and I’m cheering for them, but the problem is you’re now inserting. Suppose you and I want to start a new university, and we want to hire all the free thinking professors, and we want to have the place that fixes all this, practically speaking, we can’t do it because we can’t get access to that money.

(01:01:09)
I’ll give you the most direct reason we can’t get access to that money, we can’t get access to federal student funding. Do you know how universities are accredited for the purpose of getting access to federal student funding? Federal student loans? They’re accredited by the government, but not directly, indirectly. They’re not accredited by the Department of Education. Instead, what happens is the Department of Education accredits accreditation bureaus that are nonprofits that do the accreditation. Guess what the composition of the accreditation bureaus is? The existing universities. They are in complete control. The incumbents are in complete control as to who gets access to federal student loan money. Guess how enthusiastic they are about accrediting a new university? Right.

(01:01:49)
And so we have a government funded and supported cartel that has gone… It’s just obvious now. It’s just gone sideways in basically any possible way it could go sideways, including, I mean, literally, as you know, students getting beaten up on campus for being the wrong religion. They’re just wrong in every possible way at this point. And it’s all on the federal taxpayer back. And there is no way, I mean, my opinion, there is no way to fix these things without replacing them. And there’s no way to replace them without letting them fail. And by the way, it’s like everything else in life. I mean, in a sense, this is the most obvious conclusion of all time, which is what happens in the business world when a company does a bad job is they go bankrupt and another company takes its place, and that’s how you get progress. And of course, below that is what happens is this is the process of evolution. Why does anything ever get better? Things are tested and tried, and then the things that are good survive. And so these places, they’ve been allowed to cut themselves off, both from evolution of the institutional level and evolution of the individual level as shown by the just widespread abuse of tenure. And so we’ve just stalled out. We built an ossified system, an ossified, centralized, corrupt system, where we’re surprised by the results. They are not fixable in their current form.
Lex Fridman
(01:03:01)
I disagree with you on that. Maybe it’s grounded in hope that I believe you can revolutionize a system from within, because I do believe Stanford and MIT are important.
Marc Andreessen
(01:03:11)
But that logic doesn’t follow at all. That’s underpants gnome logic.
Lex Fridman
(01:03:15)
Underpants gnome, can you explain what that means?
Marc Andreessen
(01:03:17)
Underpants gnomes logic. So I just started watching a key touchstone of American culture with my nine-year-old, which of course is South Park.
Lex Fridman
(01:03:23)
Yes. Wow.
Marc Andreessen
(01:03:26)
Which by the way is a little aggressive for a nine-year-old.
Lex Fridman
(01:03:28)
Very aggressive.
Marc Andreessen
(01:03:28)
But he likes it. He’s learning all kinds of new words.
Lex Fridman
(01:03:32)
And all kinds of new ideas. But yeah, go on.
Marc Andreessen
(01:03:34)
I told him, I said, “You’re going to hear words on here that you are not allowed to use.”
Lex Fridman
(01:03:37)
Right. Education.
Marc Andreessen
(01:03:39)
And I said, “You know how we have an agreement that we never lie to mommy?” I said, “Not using a word that you learn in here does not count as lying. And keep that in mind.”
Lex Fridman
(01:03:51)
Wow. This is Orwellian redefinition of lying. But yes, go ahead.
Marc Andreessen
(01:03:55)
And of course, in the very opening episode, in the first 30 seconds, one of the kids calls the other kid a dildo. We’re off to the races.
Lex Fridman
(01:04:01)
Yep. Let’s go.
Marc Andreessen
(01:04:02)
“Daddy, what’s a dildo?”
Lex Fridman
(01:04:04)
Yep.
Marc Andreessen
(01:04:08)
“Sorry son. I don’t know.”
Lex Fridman
(01:04:09)
Yeah.
Marc Andreessen
(01:04:11)
So the-
Lex Fridman
(01:04:13)
Underpants gnome.
Marc Andreessen
(01:04:14)
So famous episode of South Park, the underpants gnomes. All the kids basically realize that their underpants are going missing from their dresser drawers, somebody’s stealing the underpants. And it’s just like, well, who on earth would steal the underpants? And it turns out it’s the underpants gnomes. And it turns out the underpants gnomes have come to town and they’ve got this little underground warren of tunnels in storage places for all the underpants. And so they go out at night, they steal the underpants, and the kids discover that the underpants gnomes, and they’re, “What are you doing? What’s the point of this?” And so the underpants gnomes present their master plan, which is a three-part plan, which is step one, collect underpants, step three, profit, step two, question mark. So you just proposed the underpants gnome. Which is very common in politics.

(01:04:57)
So the form of this in politics is we must do something. This is something, therefore we must do this. But there’s no causal logic chain in there at all to expect that that’s actually going to succeed because there’s no reason to believe that it is.
Lex Fridman
(01:05:10)
Yeah, but-
Marc Andreessen
(01:05:12)
But this is what I hear all the time, and I will let you talk as the host of the show in a moment, but I hear this all the time. I have friends who are on these boards, very involved in these places, and I hear this all the time, which is like, “Oh, these are very important. We must fix them.” And so therefore, they are fixable. There’s no logic chain there at all.
Lex Fridman
(01:05:32)
If there’s that pressure that you described in terms of cutting funding, then you have the leverage to fire a lot of the administration and have new leadership that steps up that aligns with this vision that things really need to change at the heads of the universities. And they put students and faculty at primary, fire a lot of the administration, and realign and reinvigorate this idea of freedom of thought and intellectual freedom, because there is already-
Lex Fridman
(01:06:00)
And intellectual freedom. Because there is already a framework of great institutions that’s there, and the way they talk about what it means to be a great institution is aligned with this very idea that you’re talking about, meaning like intellectual freedom, the idea of tenure. On the surface it’s aligned, underneath it’s become corrupted.
Marc Andreessen
(01:06:23)
If we say free speech and academic freedom often enough, sooner or later these tenured professors will get brave.
Lex Fridman
(01:06:27)
Wait, do you think that universities are fundamentally broken? Okay, so how do you fix it? How do you have institutions for educating 20-year-olds and institutions that host researchers that have the freedom to do epic shit, like research-type shit that’s outside the scopes of R&D departments and inside companies? So how do you create an institution like that?
Marc Andreessen
(01:06:51)
How do you create a good restaurant when the one down the street sucks?
Lex Fridman
(01:06:55)
Right. You invent something new?
Marc Andreessen
(01:06:57)
You open a new restaurant.
Lex Fridman
(01:06:58)
Yeah.
Marc Andreessen
(01:07:00)
How often in your life have you experienced a restaurant that’s just absolutely horrible, and it’s poisoning all of its customers and the food tastes terrible, and then three years later you go back and it’s fantastic? Charlie Munger actually had the best comment, this great investor, Charlie Munger has great comment. He was once asked, it’s like General Electric was going through all these challenges, and he was asked at a Q&A. It said, “How would you fix the culture at General Electric?” And he said, “Fix the culture at General Electric?” He said, “I couldn’t even fix the culture at a restaurant.”

(01:07:25)
It’s insane, like obviously you can’t do it. Nobody in business thinks you can do that, it’s impossible. Now look, having said all that, I should also express this because I have a lot of friends who work at these places and are involved in various attempts to fix these. I hope that I’m wrong, I would love to be wrong, I would love for the underpants gnome step two to be something clear and straightforward that they can figure out how to do. I would love to fix it, I’d love to see them come back to their spoken principles, I think that’d be great, I’d love to see the professors with tenure get bravery, it would be fantastic. My partner and I have done a lot of public speaking on this topic, it’s been intended to not just be harsh, but also be like, okay, these challenges have to be confronted directly.

(01:08:07)
By the way, let me also say something positive, especially post-October seventh, there are a bunch of very smart people who are major donors and board members of these institutions like Mark Rowan who are really coming in trying to, I think legitimately trying to fix these places. I have a friend on the executive committee at one of the top technical universities. He’s working overtime to try to do this.

(01:08:26)
Man, I hope they can figure it out. But the counter question would just be like, do you see it actually happening at a single one of these places?
Lex Fridman
(01:08:34)
I’m a person that believes in leadership. If you have the right leadership, the whole system can be changed.
Marc Andreessen
(01:08:41)
So here’s a question for your friend who have tenure at one of these places, which is who runs his university?
Lex Fridman
(01:08:46)
You know how I think runs it? Whoever the fuck says they run it, that’s what great leadership is. A president has that power.
Marc Andreessen
(01:08:54)
But how does-
Lex Fridman
(01:08:54)
President of university has the leverage because they can mouth off like Elon can.
Marc Andreessen
(01:08:58)
Can they fire the professors?
Lex Fridman
(01:08:59)
They can fire them through being vocal publicly, yes.
Marc Andreessen
(01:09:02)
Can they fire the professors?
Lex Fridman
(01:09:04)
What are you talking about legally? Can we fire? No, they cannot fire the professors.
Marc Andreessen
(01:09:07)
Then we know who runs the university.
Lex Fridman
(01:09:07)
The professors?
Marc Andreessen
(01:09:08)
Yeah, the professors. The professors and the students, the professors and the feral students. Then they’re of course in a radicalization feedback cycle driving each other crazy.
Lex Fridman
(01:09:16)
You said feral students?
Marc Andreessen
(01:09:16)
The feral students. Yeah, the feral students. What happens when you’re put in charge of your bureaucracy where the thing that the bureaucracy knows is that they can outlast you? The thing that the tenured professors at all these places know is it doesn’t matter who the president is because they can outlast them because they cannot get fired. By the way, it’s the same thing that bureaucrats in the government know. It’s the same thing that the bureaucrats in the Department of Education know. They know the exact same thing. They can outlast you. I mean it’s the whole thing that, it’s the resistance. They can be the resistance. They can just sit there and resist, which is what they do. They’re not fireable.
Lex Fridman
(01:09:47)
That’s definitely a crisis that needs to be solved. That’s a huge problem. And I also don’t like that I’m defending academia here. I agree with you that the situation is dire, but I just think that institutions are important. And I should also add context since you’ve been grilling me a little bit, you were using restaurants as an analogy and earlier offline in this conversation you said the Dairy Queen is a great restaurant. So let’s [inaudible 01:10:12].
Marc Andreessen
(01:10:11)
I didn’t say Dairy Queen is a great restaurant.
Lex Fridman
(01:10:12)
Let the listener take-
Marc Andreessen
(01:10:13)
I said Dairy Queen is the best restaurant.
Lex Fridman
(01:10:15)
The best restaurant. There you go. So everything that Marc Andreessen is saying today, put that into, cont-
Marc Andreessen
(01:10:20)
You should go order a Blizzard. One day, you should walk down there and order a Blizzard.
Lex Fridman
(01:10:23)
Yeah.
Marc Andreessen
(01:10:24)
They can get like 4,000 calories in a cup.
Lex Fridman
(01:10:26)
They can and they’re delicious.
Marc Andreessen
(01:10:26)
It’s amazing.
Lex Fridman
(01:10:28)
They are truly delicious. And they-
Marc Andreessen
(01:10:28)
They’re really fantastic. And they’ll put anything in there you want. Okay. But anyway, let me just close by saying, look, my friends in the university system, I would just say, “Look, this is the challenge.” I would just pose this as the challenge. To me having had a lot of these conversations, this is the bar in my view, this is the conversation that actually has to happen. This is the bar that actually has to be hit. These problems need to be confronted directly because I there’s think there’s been way too much, I mean, I’m actually worried on the other side. There’s too much happy talk in these conversations.

(01:10:55)
I think the taxpayers do not understand this level of crisis, and I think if the taxpayers come to understand it, I think the funding evaporates. And so I think the fuse is going through no fault of any of ours, but the fuse is going and there’s some window of time here to fix this and address it and justify the money because just normal taxpayers sitting in normal towns in normal jobs are not going to tolerate this for that much longer.

Censorship

Lex Fridman
(01:11:18)
You’ve mentioned censorship a few times. Let us, if we can go deep into the darkness of the past and how censorship mechanism was used. So you are a good person to speak about the history of this because you were there on the ground floor in 2013-ish Facebook. I heard that you were there when they invented or maybe developed the term hate speech in the context of censorship on social media. So take me through that history if you can, the use of censorship.
Marc Andreessen
(01:11:55)
So I was there on the ground in 1993.
Lex Fridman
(01:12:00)
There’s multiple floors to this building, apparently.
Marc Andreessen
(01:12:02)
There are.
Lex Fridman
(01:12:03)
Yeah.
Marc Andreessen
(01:12:03)
So I got the first ask to implement censorship on the internet, which was in the web browser.
Lex Fridman
(01:12:08)
That is fascinating.
Marc Andreessen
(01:12:09)
Yeah, yeah. Actually 1992. I was asked to implement a nudity filter.
Lex Fridman
(01:12:14)
Did you have the courage to speak up back then?
Marc Andreessen
(01:12:16)
I did not have any problems speaking up back then. I was making $6.25 cents an hour. I did not have a lot to lose. No, I was asked at the time, and then look, in some sense, a legitimate request, which is working on a research project actually funded by the federal government at a public university. So I don’t think my boss was in any way out of line, but it was like, yeah, this web browser thing is great, but could it just make sure to not have any photos of naked people that show up? But if you think about this for a second, as a technologist, I had an issue, which is this was pre-image net. And so I had a brief period where I tried to imagine an algorithm that I referred to as the breast detection algorithm that I was going to have to design and then apparently a variety of other apparently body parts people are also sensitive about. And then I politely declined to do this.
Lex Fridman
(01:13:01)
For just the technical difficulties of it.
Marc Andreessen
(01:13:04)
Well, number one, I actually didn’t know how to do it, but number two is just like, no, I’m just not building a censorship engine. I’m just not doing it. And in those days, the internet generally was a free fire zone for everything. It was actually interesting as sort of pre-’93, the internet was such a specific niche community. It was the million kind of highest IQ nerds in the world. And so it actually didn’t really have a lot of issues that people were super interested in talking about like astrophysics and not very interested in even politics at that time so there really was not an issue there. But yeah, I didn’t want to start the process.

(01:13:39)
So I think the way to think about this, so first of all, yeah. So I was involved in this at Facebook, by the way, I’ve been involved in this at Facebook every step of the way. I joined the board there in 2007 so I’ve seen everything in the last almost 20 years every step of the way. But also I’ve been involved in most of the other companies over time so I was angel investor on Twitter. I knew them really well. We were the founding investor in Substack. Part of the Elon takeover of Twitter with X. I was an angel at LinkedIn. So I’ve been in, we were the funder of Pinterest. We were one of the main investors there, Reddit as well. And I was having these conversations with all these guys all the way through. So as much talk specifically about Facebook, but I can just tell you the general pattern. And for quite a while it was kind of all the same across these companies.

(01:14:20)
So basically the way to think about this, the true kind of nuanced view of this is that there is practically speaking, no internet service that can have zero censorship. And by the way, that also mirrors, there is no country that actually has unlimited free speech either. The U.S. First Amendment actually has 12 or 13 formal carve outs from the Supreme Court over time. So incitement to violence and terrorist recruitment and child abuse and child pornography and so forth, they’re not covered by the First Amendment. And just practically speaking, if you and I are going to start an internet company and have a service, we can’t have that stuff either because illegal or it will just clearly destroy the whole thing.

(01:14:56)
So you’re always going to have a censorship engine. I mean hopefully it’s not actually in the browser, but you’re going to have it for sure at the level of an internet service. But then what happens is now you have a machine. Now you have a system where you can put in rules saying, we allow this. We don’t allow that. You have enforcement, you have consequences. And once that system is in place, it becomes the ring of power, which is like, okay, now anybody in that company or anybody associated with that company or anybody who wants to pressure that company will just start to say, “Okay, you should use that machine for more than just terrorist recruitment and child pornography. You should use it for X, Y, Z.”

(01:15:35)
And basically that transition happened, call it 2012, 2013 is when there was this very, very kind of rapid pivot. I think the kickoff to it for some reason it was the beginning of the second Obama term. I think it also coincided with the sort of arrival of the first kind of super woke kids into these schools. It’s the kids that were in school between for the Iraq war and then the global financial crisis and they came out super radicalized. They came into these companies, they immediately started mounting these social crusades to ban and censor lots of things.

(01:16:08)
And then quite frankly, the Democratic Party figured this out. And they figured out that these companies were very subject to being controlled and the executive teams and boards of directors were almost all Democrats. And there’s tremendous circulation. A lot of Obama people from the first term actually came and worked in these companies. And a lot of FBI people and other law enforcement intelligence people came in and worked and they were all Democrats for that set. And so the ring of power was lying on the table. It had been built and they picked it up and put it on, and then they just ran.

(01:16:37)
And the original discussions were basically always on two topics. It was hate speech and misinformation. Hate speech was the original one. And the hate speech conversation started exactly like you’d expect, which is we can’t have the N word. And which the answer is fair enough, let’s not have the N word. Now, we’ve set a precedent, and Jordan Peterson has talked a lot about this. The definition of hate speech ended up being things that make people uncomfortable. So we can’t have things that make people uncomfortable. I, course people like me that are disagreeable raise their hands and say, “Well, that idea right there makes me uncomfortable.” But of course that doesn’t count as hate speech. So the ring of power is on one hand and not on the other hand.

(01:17:19)
And then basically that began this slide where it ended up being that completely anodyne is the point Mark has been making recently completely anodyne comments that are completely legitimate on television or on the Senate floor all of a sudden are hate speech can’t be said online so that the ring of power was wielded in grossly irresponsible ways. We could talk about all the stuff that happened there.

(01:17:39)
And then the other one was misinformation. And there was a little bit of that early on, but of course that really kicked in with Trump. So hate speech stuff pre-dated Trump by three or four years. The misinformation stuff was, it was a little bit later and it was a consequence of the Russiagate hoax. And then that was a ring of power that was even more powerful because hate speech, it’s like, okay, at some point if something offensive or not, at least you can have a question as to whether that’s the case. But the problem with misinformation is like, is it the truth or not? What do we know for 800 years or whatever western civilization it’s that there’s only a few entities that can determine the truth on every topic. There’s God, there’s the king. We don’t have those anymore and the rest of us are all imperfect and flawed.

(01:18:25)
And so the idea that any group of experts is going to sit around the table and decide on the truth is deeply anti-Western and deeply authoritarian. And somehow the misinformation kind of crusade went from the Russiagate hoax into just full-blown, we’re going to use that weapon for whatever we want.

(01:18:40)
And then of course, then the culminating moment on that, that really was the straw that broke the camel’s back was we’re going to censor all theories that the COVID virus might’ve been manufactured in a lab as misinformation. And inside these companies, that was the point where people for the first time, this is what, three years ago for the first time, they were like, that was when it sunk in where it’s just like, okay, this has spun completely out of control. But anyway, that’s how we got to where we are.

(01:19:04)
And then basically that spell lasted, that complex existed and got expanded basically from, call it 2013 to 2023. I think basically two things broke it. One is Substack, and I’m super proud of those guys because they started from scratch and declared right up front that they were going to be a free speech platform. And they came under intense pressure, including from the press, and they tried to beat them to the ground and kill them. And intense pressure, by the way, from let’s say certain of the platform companies basically threatening them. And they stood up to it. And sitting here today, they have the widest spectrum of speech and conversation anywhere on planet Earth. And they’ve done a great job. And it is worked by the way. It’s great. And then obviously Elon with X was the hammer blow. And then the third one now is what Mark is doing at Facebook.

Jon Stewart

Lex Fridman
(01:19:57)
And there’s also singular moments, I think you’ve spoken about this, which like Jon Stewart going on Stephen Colbert and talking about the lab leak theory.
Marc Andreessen
(01:20:08)
Yes.
Lex Fridman
(01:20:09)
There’s certain moments that just kind of shake everybody up, the right person the right time. It’s a wake-up call.
Marc Andreessen
(01:20:17)
So that there, and I will tell you, and I should say Jon Stewart attacked me recently, so I’m not that thrilled about him, but I would say I was a long run fan of Jon Stewart. I watched probably every episode of The Daily Show when he was on it for probably 20 years. But he did a very important public service and it was that appearance on the Colbert Show. And I don’t know how broadly this is, at the time, it was in the news briefly, but I don’t know how if people remember this, but I will tell you in the rooms where people discuss what is misinformation and these policies, that was a very big moment. That was probably actually the key catalyzing moment. And I think he exhibited, I would say, conspicuous bravery and had a big impact with that.

(01:20:51)
And for people who don’t recall what he did, and this was in the full-blown, you absolutely must lock down for two years. You absolutely must keep all the schools closed. You absolutely must have everybody work from home. You absolutely must wear a mask like the whole thing. And then one of those was you absolutely must believe that COVID was completely natural. You must believe that. And not believing that means you’re a fascist Nazi Trump supporter, MAGA, evil QAnon person. And uniformly, that was enforced by the social media companies. And like I said, that was the peak. And Jon Stewart went on the Colbert Show, and I don’t know if they planned it or not because Colbert looked shocked. I don’t know how much, it was a bit, but he went on there and he just had one of these, the Emperor’s wearing no clothes things where he said, “It’s just not plausible that you had the COVID super virus appear 300 yards down the street from the Wuhan Institute of lethal coronaviruses.” It’s just not plausible that certainly that you could just rule that out.

(01:21:46)
And then there was another key moment, actually, the more serious version was I think the author, Nicholson Baker wrote a big piece for New York Magazine. And Nicholson Baker is one of our great novelist, writers of our time. And he wrote the piece and he did the complete undressing of it. And that was the first, I think that was the first legit, there had been alt renegade, there had been people running around saying this, but getting censored all over the place. That was the first one that was in the mainstream press and he talked to all the heretics and he just laid the whole thing out. And that was a moment.

(01:22:13)
And I remember let’s say a board meeting at one of these companies after that where basically everybody looked around the table and was like, “All right, I guess we’re not, we don’t need to censor that anymore.” And then of course, what immediately follows from that is, “Well, wait a minute, why were we censoring that in the first place?” And then the downstream, not that day, but the downstream conversations were like, “Okay, if we made such a giant, in retrospect, if we all made such a giant collective mistake censoring that, then what does that say about the rest of our regime?” And I think that was the thread in the sweater that started to unravel it.

Mark Zuckerberg on Joe Rogan

Lex Fridman
(01:22:44)
I should say it again, I do think that the Jon Stewart appearance and the statement he made was a courageous act.
Marc Andreessen
(01:22:49)
Yeah, I agree.
Lex Fridman
(01:22:50)
I think we need to have more of that in the world. And like you said, Elon, everything he did with X is a series of courageous acts. And I think what Mark Zuckerberg did on Rogan a few days ago is a courageous act. Can you just speak to that?
Marc Andreessen
(01:23:12)
He has become, I think, an outstanding communicator, and he’s somebody who came in for a lot of criticism earlier in his career on that front. And I think he’s one of these guys who can sit down and talk for three hours and make complete sense. And as you do with all of your episodes, when somebody sit and talks for three hours, you really get a sense of somebody because it’s really hard to be artificial for that long and he’s now done that repeatedly. He’s really good at it. And then look again, I would maybe put him in the third category now certainly after that appearance, I would say I would put him up there now with kind of Elon and Trump in the sense of the public and the private are now synchronized. I guess I’d say that. He said on that show what he really believes. He said all the same things that he says in private. I don’t think there’s really any discrepancy anymore.

(01:23:55)
I would say he has always taken upon himself a level of obligation, responsibility to running a company the size of Meta and to running services that are that large. And I think his conception of what he’s doing, which I think is correct, is he’s running services that are bigger than any country. Over 3 billion people use those services. And then the company has many tens of thousands of employees and many investors, and it’s a public company and he thinks very deeply and seriously about his responsibilities. And so he has not felt like he has had, let’s just say, the complete flexibility that Elon has had. And people could argue that one way or the other, but he talked about a lot. He’s evolved a lot. A lot of it was he learned a lot.

(01:24:38)
And by the way, I’m going to put myself right back up there. I’m not claiming any huge foresight or heroism on any of this. I’ve also learned a lot, my views on things are very different than they were 10 years ago on lots of topics. And so I’ve been on a learning journey. He’s been on a learning journey. He’s a really, really good learner. He assimilates information as good as, or better than anybody else I know.

(01:25:02)
The other thing I guess I would just say is he talked on that show about something very important, which is when you’re in a role where you’re running a company like that, there are a set of decisions that you get to make and you deserve to be criticized for those decisions and so forth and it’s valid, but you are under tremendous external pressure as well. And by the way, you’re under tremendous internal pressure. You’ve got your employees coming at you, you’ve got your executives in some cases coming at you. You’ve got your board in some cases coming at you. You’ve got your shareholders coming at you, so you’ve got your internal pressures, but you also have the press coming at you. You’ve got academia coming at you, you’ve got the entire nonprofit complex activist complex coming at you.

(01:25:40)
And then really critically, he talked about in Rogan and these companies all went through this, in this last especially five years, you had the government coming at you. And that’s the really stinky end of the pool where the government was, in my view, illegally exerting just in flagrant violation of the First Amendment and federal laws on speech and coercion and conspiracy, forcing these companies to engage in activities. Again, in some cases they may have wanted to do, but in other cases they clearly didn’t want to do and felt like they had to do.

(01:26:11)
And the level of pressure, like I say, I’ve known every CEO Twitter, they’ve all had the exact same experience, which when they were in the job, it was just daily beatings. It’s just getting punched in the face every single day constantly. And Mark is very good at getting physically punched in the face and then-
Lex Fridman
(01:26:29)
Getting better and better.
Marc Andreessen
(01:26:31)
And he is. And he’s very good at taking a punch, and he has taken many, many punches. So I would encourage people to have a level of sympathy for these are not kings, these are people who operate with I would say extraordinary levels of external pressure. I think if I had been in his job for the last decade, I would be a little puddle on the floor. And so it says, I think a lot about him that he has risen to this occasion the way that he has.

(01:26:53)
And by the way, I should also say the cynicism of course is immediately out and it’s a legitimate thing for people to say, but it’s like, “Oh, you’re only doing this because of Trump or whatever.” And it’s just like, no, he has been thinking about and working on these things and trying to figure them out for a very long time. And so I think what you saw are legitimate, deeply held beliefs, not some sort of just-in-the-moment thing that could change at any time.
Lex Fridman
(01:27:15)
So what do you think it’s like to be him and other leaders of companies, to be you and withstand internal pressure and external pressure? What’s that life? Is it deeply lonely?
Marc Andreessen
(01:27:27)
That’s a great question. So leaders are lonely to start with. And this is one of those things where almost nobody has sympathy. Nobody feels sorry for a CEO. It’s not a thing. And again, legitimately so CEOs get paid a lot, the whole thing, there’s a lot of great things about it. So it’s not like they should be out there asking for a lot of sympathy, but it is the case that they are human beings and it is the case that it is a lonely job. And the reason it’s a lonely job is because your words carry tremendous weight and you are dealing with extremely complicated issues, and you’re under a tremendous amount of emotional, personal, emotional stress. And you often end up not being able to sleep well, and you end up not being able to keep up an exercise routine and all those things. And you come under family stress because you’re working all the time.

(01:28:08)
Or my partner Ben, he was CEO of our last company before we started the venture firm. He said the problem he had with his family life was even when he was home at night, he wasn’t home because he was in his head trying to solve all the business problems. And so he was supposed to be having dinner with his kids and he was physically there, but he wasn’t mentally there so you get that a lot. But the key thing is you can’t talk to people. You can. I mean, you can talk to your spouse and your kids, but they don’t understand that they’re not working in your company. They don’t understand, have the context to really help you. If you talk to your executives, they all have agendas and they can’t resist. It’s just human nature. And so you can’t necessarily rely on what they say. It’s very hard in most companies to talk to your board because they can fire you.

(01:28:52)
Now, Mark has the situation because he has control, it actually turns out he can talk to his board. And Mark talks to us about many things that most CEOs won’t talk to their boards about literally because we can’t fire him. But a general, including all the CEOs of Twitter, none of them had control and so they could all get fired. You can’t talk to the board members. They’re going to fire, you can’t talk to the shareholders because they’ll just dump your stock. Okay.

(01:29:16)
So every once in a while, what you find is basically the best case scenario they have is they can talk to other CEOs, and there’s these little organizations where they kind of pair up and do that and so they maybe get a little bit out of that. But even that’s fraught with peril because can you really talk about confidential information with another CEO, insider trading risk. And so it’s just a very lonely isolating thing to start with.

(01:29:35)
And then on top of that, you apply pressure, and that’s where it gets painful. And then maybe I’ll just spend a moment on this internal external pressure thing. My general experience with companies is that they can withstand most forms of external pressure as long as they retain internal coherence. So as long as the internal team is really bonded together and supporting each other, most forms of external pressure you can withstand. And by that I mean investors dump your stock, you lose your biggest customers, whatever negative article, negative headline, you can withstand all that. And basically, in fact, many of those forms of pressure can be bonding experiences for the team where they come out stronger.

(01:30:20)
What you 100% cannot withstand is the internal crack. And what I always look for in high pressure corporate situations now is the moment when the internal team cracks because I know the minute that happens, we’re in a different regime. It’s like the solid has turned into a liquid, we’re in a different regime, and the whole thing can unravel in the next week because then people turn, I mean, this is what’s happening in Los Angeles right now. The mayor and the fire chief turned on each other, and that’s it. That government is dysfunctional. It is never going to get put back together again. It is over. It is not going to work ever again. And that’s what happens to inside companies.

(01:30:56)
And so somebody like Mark is under profound internal pressure and external pressure at the same time. Now he’s been very good at maintaining the coherence of his executive team, but he has had over the years, a lot of activist employees as a lot of these companies have had and so that’s been continuous pressure.

(01:31:15)
And then the final thing I’d say is I said that companies can withstand most forms of external pressure, but not all [inaudible 01:31:21] though not all one is government pressure. Is it when your government comes for you? Yeah. Any CEO who thinks that they’re bigger than their government, has that notion beaten out of them in short order.

Government pressure

Lex Fridman
(01:31:32)
Can you just linger on that because it is maybe educating and deeply disturbing? You’ve spoken about it before, but we’re speaking about again this government pressure. So you think they’ve crossed the line into essentially criminal levels of pressure?
Marc Andreessen
(01:31:50)
Flagrant criminality, felonies, like obvious felonies. And I can actually cite the laws, but yes, absolute criminality.
Lex Fridman
(01:31:59)
Can you explain how those possible to happen and maybe on a hopeful note, how we can avoid that happening again?
Marc Andreessen
(01:32:07)
So just start with is a lot of this now is in the public record, which is good because it needs to be in the public record. And so there’s three forms of things that are in the public record that people can look at. So one is the Twitter files, which Elon put out with the set of journalists when he took over. And I will just tell you, the Twitter files are a hundred percent representative of what I’ve seen at every other one of these companies. And so you can just see what happened in Twitter and you can just assume that that happened in these other companies for the most part, certainly in terms of the kind of pressure that they got. So that’s number one. That stuff, you can just read it and you should if you haven’t.

(01:32:38)
The second is Mark referenced this in the Rogan podcast. There’s a congressman Jim Jordan who has a congressional committee called the Weaponization Committee. And they, in the last, whatever three years, have done a full-scale investigation of this. And Facebook produced a lot of documents into that investigation and many of those have now been made public and you can download those reports. And there’s 2000 pages worth of material on that. And that’s essentially the Facebook version of the Twitter files just arrived at with a different mechanism.

(01:33:06)
And then third is Mark himself talking about this on Rogan, so I’ll just defer to his comments there. But yeah, basically what those three forms of information show you is basically the government over time and then culminating in 2020, 2021 in the last four years, just decided that the First Amendment didn’t apply to them. And they just decided that federal laws around free speech and around conspiracies to take away the rights of citizens just don’t apply. And they just decided that they can just arbitrarily pressure, just like literally arbitrarily call up companies and threaten and bully and yell and scream and threaten repercussions and force them to censor.

(01:33:45)
And there’s this whole thing of like, well, the First Amendment only applies to, the government, it doesn’t apply to companies. It’s like, well, there’s actually a little bit of nuance to that. First of all, it definitely applies to the government. 100%, the First Amendment applies to the government. By the way, so does the Fourth Amendment and the Fifth Amendment, including the right to due process, also applies to the government. There was no due process at all to any of the censorship regime that was put in place. There was no due process put in place, by the way, for de-banking either. Those are just as serious violations as the free speech violations. And so this is just flagrant, flagrant, unconstitutional behavior.

(01:34:18)
And then there are specific federal statutes, 18 241 and 18 242, and one of them applies to federal employees, government employees, and the other one applies to private actors around what’s called deprivation of rights and conspiracy to deprive rights. And it is not legal according to the United States Criminal Code for government employees or in a conspiracy private entities to take away constitutional rights. And interestingly, some of those constitutional rights are enumerated, for example, in the First Amendment, freedom of speech. And then some of those rights actually do not need to be enumerated. If the government takes away rights that you have, they don’t need to be specifically enumerated rights in the Constitution in order to still be a felony. The Constitution very specifically does not say you only have the rights that it gives you. It says you have all the rights that have not been previously defined as being taken away from you. And so de-banking qualifies as a right, right to access to the financial system, is every bit something that’s subject to these laws as free speech. And so yeah, this has happened.

(01:35:18)
And then I’ll just add one final thing, which is we’ve talked about two parties so far. We talked about the government employees and then we’ve talked about the companies. The government employees for sure have misbehaved. The companies, there’s a very interesting question there as to whether they are victims or perpetrators or both. They will defend, they will argue, and I believe they have a good case, that they are victims, not perpetrators, right? They’re the downstream subjects of pressure, not the cause of pressure, but there’s a big swath of people who are in the middle and specifically the ones that are funded by the government that I think are in possibly pretty big trouble. And that’s all of these third-party censorship bureaus.

(01:35:53)
I mean, the one that is most obvious is the so-called Stanford Internet Observatory that got booted up there over the last several years. And they basically were funded by the federal government to be third-party censorship operations. And they’re private sector actors, but acting with federal funding. And so it puts them in this very interesting spot where there could be very obvious theory under which they’re basically acting as agents of the government. And so I think they’re also very exposed on this and have behaved in just flagrantly illegal ways.
Lex Fridman
(01:36:22)
So fundamentally, government should not do any kind of pressure, even soft pressure on companies to censor?
Marc Andreessen
(01:36:30)
Can’t. Not allowed.
Lex Fridman
(01:36:32)
It really is disturbing. It probably started soft, lightly slowly, and then it escalates as the old [inaudible 01:36:44] to power will instruct them to do. I mean, yeah, that’s why there’s protection because you can’t put a check on power for government, right?
Marc Andreessen
(01:36:54)
There are so many ways that they can get you. There are so many ways they can come at you and get you. And the thing here to think about is a lot of times when people think about government action, they think about legislation. So when I was a kid, we got trained, how does government work? There was this famous animated short, the thing we got shown was just a cartoon of how a bill becomes a law. It’s like this fancy little bill sneaked along and guess this-
Lex Fridman
(01:37:15)
I’m just the bill. Yeah.
Marc Andreessen
(01:37:16)
Exactly. It’s like, all right, number one, that’s not how it works at all. That doesn’t actually happen. We could talk about that. But even beyond that, mostly what we’re dealing with is not legislation. When we talk about government power these days, mostly it’s not legislation. Mostly it’s either regulation, which is basically the equivalent of legislation, but having not gone through the legislative process, which is a very big open legal issue. And one of the things that the DOGE is very focused on. Most government rules are not legislated. They’re regulated and there’s tons and tons of regulations that these companies are, this is another cliche you’ll hear a lot, which is, “Oh, private companies can do whatever they want.” It’s like, “Oh no can’t.”

(01:37:50)
There’s subject to tens of thousands of regulations that they have to comply with. And the hammer that comes down when you don’t comply with regulations is profound. They can completely wreck your company with no ability for you to do anything about it. So regulation is a big part of the way the power gets exercised.

(01:38:04)
And then there’s called just flat out administrative power, the term that you’ll hear and administrative power is just literally the government telling you, calling you and telling you what to do. Here’s an example of how this works. So Facebook had this whole program a few years back to do a global cryptocurrency for payments called Libra. And they built the entire system and it was this high-scale sort of new cryptocurrency, and they were going to build into every product, and they were going to be 3 billion people who could transact with Libra. And they went to the government and they went to all these different, trying to figure out how to make it so it’s fully compliant with anti-money laundering and all these controls and everything. And they had the whole thing ready to go.

(01:38:34)
Two senators wrote letters to the big banks saying, “We’re not telling you that you can’t work with Facebook on this, but if you do, you should know that every aspect of your business is going to come under greatly increased level of regulatory scrutiny,” which is of course the exact equivalent of it sure is a nice corner restaurant you have here. It would be a shame if somebody tossed a Molotov cocktail through the window and burned it down tonight, right?

(01:38:57)
And so what is that letter? It’s not a law. It’s not even a regulation, it’s just.
Marc Andreessen
(01:39:00)
It’s not a law, it’s not even a regulation, it’s just straight direct state power. And then it culminates in literally calls from the White House where they’re just flat out telling you what to do, which is of course what a king gets to do, but not what a president gets to do. Anyway. So what these companies experienced was they experienced the full panoply of this, but the level of intensity was in that order. It was actually, legislation was the least important part. Regulation was more important, administrative power was more important, and then just flat out demands and flat out threats were ultimately the most important. How do you fix it? Well, first of all, you have to elect people who don’t do it. As with all these things, ultimately the fault lies with the voters. And so you have to decide you don’t want to live in that regime.

(01:39:44)
I have no idea what part of this recent election mapped to the censorship regime. I do know a lot of people on the right got very angry about the censorship, but I think it probably at least helped with enthusiasm on that side. Maybe some people on the left will now not want their Democratic nominees to be so pro censorship. So the voters definitely get a vote, number one. Number two, I think you need transparency. You need to know what happened. We know some of what happened. Peter Thiel has written in the FT just now saying, after what we’ve been through in the last decade we need the broad-based truth and reconciliation efforts to really get to the root of things. So maybe that’s part of it. We need investigations for sure. Ultimately, we need prosecutions. Ultimately, we need people to go to jail. Because we need to set object lessons that say that you don’t get to do this. And on those last two, I would say those are both up to the new administration, and I don’t want to speak for them and I don’t want to predict what they’re going to do, but they for sure have the ability to do both of those things and we’ll see where they take it.
Lex Fridman
(01:40:43)
Yeah. It’s truly disturbing. I don’t think anybody wants this kind of overreach of power for government, including perhaps people that are participating in it. It’s like this dark momentum of power that you just get caught up in it. And that’s the reason there’s that kind of protection. Nobody wants that.
Marc Andreessen
(01:41:01)
I use the metaphor, the ring of power. And for people who don’t catch the reference, that’s Lord of the Rings. And the thing with the ring of power and Lord of the Rings, it’s the ring the Gollem has in the beginning and it turns you invisible. And it turns out it unlocks all this fearsome power. It’s the most powerful thing in the world, is to key to everything. And basically the moral lesson of Lord of the Rings, which was written by a guy who thought very deeply about these things is, yeah, the ring of power is inherently corrupting. The characters at one point, they’re like, “Gandalf, just put on the ring and fix this.” He will not put the ring on even to end the war because he knows that it will corrupt him. As it starts, the character of Gollem is the result of a normal character who ultimately becomes this incredibly corrupt and deranged version of himself.

(01:41:44)
I think you said something actually quite profound there, which is the ring of power is infinitely tempting. The censorship machine is infinitely tempting. If you have it, you are going to use it. It’s overwhelmingly tempting because it’s so powerful, and that it will corrupt you. Yeah. I don’t know whether any of these people feel any of this today. They should. I don’t know if they do. But yeah. You go out five or 10 years later, you would hope that you would realize that your soul has been corroded and you probably started out thinking that you were a patriot and you were trying to defend democracy, and you ended up being extremely authoritarian and anti-democratic and anti-western.

Nature of power

Lex Fridman
(01:42:20)
Can I ask you a tough question here? Staying on the ring of power is quickly becoming the most powerful human on earth.
Marc Andreessen
(01:42:34)
I’m not sure about that.
Lex Fridman
(01:42:35)
You don’t think he is.
Marc Andreessen
(01:42:37)
Well, he doesn’t have the nukes so.
Lex Fridman
(01:42:39)
Nukes. Yeah. There’s different definitions and perspectives on power, right?
Marc Andreessen
(01:42:45)
Yeah.
Lex Fridman
(01:42:45)
How can he and or Donald Trump avoid the corrupting aspects of this power?
Marc Andreessen
(01:42:53)
I think the danger is there with power. It’s flat out there. I would say with Elon, we’ll see. I would say with Elon, and I would say by the way, overwhelmingly, I would say so far so good. I’m extremely, extremely thrilled by what he’s done on almost every front for the last 30 years. But including all this stuff recently. I think he’s been a real hero on a lot of topics where we needed to see heroism. But look, I would say, I guess the case that he has this level of power is some combination of the money and the proximity to the president. And obviously both of those are instruments of power. The counter argument to that is I do think a lot of how Elon is causing change in the world right now … There’s the companies he’s running directly where I think he’s doing very well, and we’re investors in multiple of them and doing very well.

(01:43:36)
But I think a lot of the stuff that gets people mad at him is like, it’s the social and political stuff, and it’s his statements, and then it’s the downstream effects of his statements. So for example, for the last couple of weeks, it’s been him weighing in on this rape gang scandal, this organized child rape thing in the UK. It’s a preface cascade. It’s one of these things where people knew there was a problem, they weren’t willing to talk about it, it got suppressed. And then Elon brought it up, and then all of a sudden there’s now in the UK this massive explosion of basically open conversation about it for the first time. It’s like this catalyzing, all of a sudden everybody’s woken up and being like, “Oh my God, this is really bad.” And there will be now pretty clearly big changes as a result.

(01:44:19)
And Elon, he played the role of the boy who said, the emperor has no clothes. But here’s the thing, here’s my point. He said it about something that was true. And so had he said it about something that was false, he would get no credit for it. He wouldn’t deserve any credit for it. But he said something that was true. And by the way, everybody over there instantly, they were like, “Oh, yeah, he’s right.” They’re just arguing the details now. So number one, it’s like, okay, he says true things. And so it’s like, okay, how far … Put it this way. How worried are we about somebody becoming corrupt by virtue of their power being that they get to speak the truth? And I guess I would say, especially in the last decade of what we’ve been through where everybody’s been lying all the time about everything, I’d say, I think we should run this experiment as hard as we can to get people to tell the truth. And so I don’t feel that bad about that.

(01:45:05)
And then the money side, this rapidly gets into the money in politics question. And the money in politics question is this very interesting question because it seems like there’s a clear cut case that the more money in politics, the worse things are and the more corrupted the system is. That was a very popular topic of public conversation up until 2016 when Hillary outspent Trump three to one and lost. You’ll notice that money in politics has almost vanished as a topic in the last eight years. And once again, Kamala raised and spent 1.5 billion on top of what Biden had spent. So they were at, I don’t know, something like three billion total and Trump, I think spent again, a third or a fourth of that. So the money in politics topic has vanished from the popular conversation in the last eight years. It has come back a little bit now that Elon is spending. But again, it’s like, okay, he’s spending, but the data would seem to indicate, at least in the last eight years, that money doesn’t win the political battles. The voters actually have a voice and they actually exercise it, and they don’t just listen to ads. And so again, there, I would say, yeah, clearly there’s some power there, but I don’t know if it’s some weapon that he can just turn on and use in a definitive way.
Lex Fridman
(01:46:16)
I don’t know if there’s parallels there, but I could also say just on a human level, he has a good heart and I interact with a lot of powerful people, and that’s not always the case. So that’s a good thing there. If we can draw parallels to the Hobbit or whatever. Who gets to put on the ring?
Marc Andreessen
(01:46:36)
Frodo.
Lex Fridman
(01:46:36)
Frodo. Yeah.
Marc Andreessen
(01:46:37)
Yeah. Maybe one of the lessons of Lord of the Rings is even Frodo would’ve been, even Frodo would’ve been corrupted. But nevertheless, you had somebody who could do what it took at the time. The thing that I find just so amazing about the Elon phenomenon and all the critiques is the one thing that everybody in our societies universally agrees on because of our post-Christian egalitarian, so we live in this post secularized Christian context in the west now, and we consider Christianity backwards, but we still believe essentially all the same things. We just dress them up in fake science.

(01:47:12)
So the one thing that we’re all told, we’re all taught from early is that the best people in the world are the people who care about all of humanity. All of our figures are people who care about all of … Jesus cared about all of humanity. Gandhi cared about all of humanity. Martin Luther King cared about all of humanity. The person who cares the most about everybody. And with Elon, you have a guy who literally … He talks about this constantly, and he talks about exactly the same in private. He is literally, he is operating on behalf of all of humanity to try to get us … He goes through to get us through multi-planetary civilization so that we can survive a strike at any one planet so that we can extend the light of human consciousness into the world and into the universe and have it persist in the good of the whole thing. And literally the critique is, yeah, we want you to care about all of humanity, but not like that.
Lex Fridman
(01:47:56)
Yeah. All the critics. All the surface turmoil, the critics will be forgotten.
Marc Andreessen
(01:48:03)
Yeah. I think that’s clear.
Lex Fridman
(01:48:05)
You said that we always end up being ruled by the elites of some kind. Can you explain this law, this idea?
Marc Andreessen
(01:48:13)
So this comes from a Italian political philosopher from about a hundred years ago named Robert … I’m going to mangle … I’ll let you pronounce the Italian. Michels or Michels. I learned about it through a famous book on politics. Probably the best book on politics written in the 20th century called The Machiavellians by this guy James Burnham, who has had a big impact on me. But in The Machiavellians, he resurrects what he calls this Italian realist school of political philosophy from the ’10s and ’20s. To be clear, this was not like a Mussolini thing. These were people who were trying to understand the actual mechanics of how politics actually works. So to get to the actual mechanical substance of how the political machine operates.

(01:48:55)
And this guy, Michels had this concept he ended up with called the Iron Law of Oligarchy. And so what the Iron Law of Oligarchy … Take a step back to say what he meant by oligarchy because it has multiple meanings. So basically, in classic political theory, there’s basically three forms of government at core. There’s democracy, which is rule of many, there’s oligarchy, which is rule of the few, and there’s monarchy, which is rule of the one. And you can just use that as a general framework of any government going to be under is going to be one of those. Just mechanical observation. Without even saying which one’s good or bad, just a structural observation. And so the question that Michels asked was, is there such a thing as democracy? Is there actually such a thing as democracy? Is there ever actually direct government? And what he did was he mounted this incredible historical exploration of whether democracies had ever existed in the world. And the answer basically is almost never. And we could talk about that.

(01:49:45)
But the other thing he did was he sought out the most democratic private organization in the world that he could find at that point, which he concluded was some basically communist German autoworkers union that was wholly devoted to the workers of the world uniting back when that was the hot thing. And he went in there and he is like, okay, this is the organization out of all organizations on planet Earth that must be operating as a direct democracy. And he went in there and he’s like, “Oh, nope.” There’s a leadership class. There’s like six guys at the top and they control everything and they lead the rest of the membership along by the nose, which is of course the story of every union. The story of every union is always the story of there’s a Jimmy Hoffa in there running the thing. We just saw that with the dock worker’s union. There’s a guy and he’s in charge. And by the way, the number two is his son. That’s not an accident.

(01:50:34)
So the Iron Law of Oligarchy basically says democracy is fake. There’s always a ruling class. There’s always a ruling elite structurally. And he said, “The reason for that is because the masses can’t organize.” What’s the fundamental problem? Whether the mass is 25,000 people in a union or 250 million people in a country, the masses can’t organize, the majority cannot organize, only a minority can organize. And to be effective in politics, you must organize. And therefore, every political structure in human history has been some form of a small organized elite ruling a large and dispersed majority. Every single one. The Greeks and the Florentines had brief experiments in direct democracy, and they were total disasters. In Florence … I forget the name of it. It was called The Workers’ Revolt or something like that. There was a two-year period where they basically experimented with direct democracy during the Renaissance, and it was a complete disaster and they never tried it again.

(01:51:27)
In the state of California, we have our own experiment on this, which is the proposition system, which is an overlay on top of the legislature. Anybody who looks at it for two seconds concludes it’s been a complete disaster. It’s just a catastrophe, and it’s caused enormous damage to the state. And so basically the presumption that we are in a democracy is just by definition, fake. Now, good news for the US. It turns out the founders understood this. And so of course they didn’t give us a direct democracy. They gave us a representative democracy. And so they built the oligarchy into the system in the form of Congress and the executive branch and the judicial branch. So anyway, so as a consequence, democracy is always everywhere fake. There is always a ruling elite. And basically the lesson of the Machiavellians is you can deny that if you want, but you’re fooling yourself. The way to actually think about how to make a system work and maintain any shred of freedom is to actually understand that that is actually what’s happening.
Lex Fridman
(01:52:18)
And lucky for us, the founders saw this and figured out a way to, given that there’s going to be a ruling elite, how to create a balance of power among that elite so it doesn’t get out of hand.
Marc Andreessen
(01:52:33)
And it was very clever. Some of this was based on earlier experiments. By the way, these were very, very smart people. And so they knew tremendous amounts of Greek and Roman history. They knew the Renaissance history. The Federalist Papers, they argued this a great length. You can read it all. They ran one of the best seminars in world history trying to figure this out. And they went through all this. So they thought through it very carefully, but just, I’ll give you an example, which continues to be a hot topic. So one way they did it just through the three branches of government, executive, legislative, and judicial. Balance the powers. But the other way they did it was they echoing what had been done earlier I think in the UK Parliament, they created the two different bodies of the legislature. And so the House and the Senate. And as you know, the house is a portion on the basis of population, and the Senate is not. The small states have just as many senators as the big states. And then they made the deliberate decision to have the house get reelected every two years to make it very responsive to the will of the people. And they made the decision to have the Senate get reelected every six years so that it had more buffer from the passions of the moment.

(01:53:35)
But what’s interesting is they didn’t choose one or the other. They did them both. And then to get legislation passed, you have to get through both of them. And so they built in a second layer of checks and balances. And then there’s a thousand observations we could make about how well the system is working today and how much does it live up to the ideal, and how much are we actually complying with the constitution? And there’s lots of open questions there, but this system has survived for coming on 250 years with a country that has been spectacularly successful. But I don’t think, at least … I don’t think any of us would trade the system for any other one. And so it’s one of the great all time achievements.
Lex Fridman
(01:54:09)
Yeah. It’s incredible. And we should say they were all pretty young relative to our current set of leaders.
Marc Andreessen
(01:54:15)
They were. Many in their 20s at the time. And super geniuses. This is one of those things where it’s just like, all right, something happened where there was a group of people where nobody ever tested their IQs, but these are Einstein’s of politics. An amazing thing. But anyway, I go through all that, which is they were very keen students of the actual mechanical practice of democracy, not fixated on what was desirable. They were incredibly focused on what would actually work, which is I think the way to think about these things.
Lex Fridman
(01:54:40)
There were engineers of sort, not the fuzzy humanity students of sort.
Marc Andreessen
(01:54:45)
They were shape rotators, not word cells.
Lex Fridman
(01:54:48)
I remember that. Wow, that meme came and went. I think you were central to them. You’re central to a lot of memes.
Marc Andreessen
(01:54:54)
I was.
Lex Fridman
(01:54:55)
You’re the meme dealer and the meme popularizer.
Marc Andreessen
(01:54:59)
That meme I gets some credit for and then the current thing is the other one I get some credit for. I don’t know that I invented either one, but I popularized them.

Journalism

Lex Fridman
(01:55:05)
Take credit and run with it. If we can just linger on the Machiavellians. It’s a study of power and power dynamics, like you mentioned, looking at the actual reality of the machinery of power. From everything you’ve seen now in government, but also in companies, what are some interesting things you can continue to say about the dynamics of power, the jostling for power that happens inside these institutions?
Marc Andreessen
(01:55:34)
Yeah. A lot of it, we already talked about this a bit with the universities, which is you can apply a Machiavellian style lens to … It’s why I posed the question to you that I did, which is okay, who runs the university, the trustees, the administration, the students or the faculty? And the true answer is some combination of the three, of the four plus the donors. By the way, plus the government, plus the press, et cetera. And so there’s a mechanical interpretation of that. Companies operate under the exact same set of questions. Who runs a company? The CEO, but the CEO EO runs the company basically up to the day that either the shareholders or the management team revolt. If the shareholders revolt, it’s very hard for the CEO O to stay in the seat. If the management team revolts, it’s very hard for the CEO to stay in the seat.

(01:56:16)
By the way, if the employees revolt, it’s also hard to stay in the seat. By the way, if the New York Times comes at you, it’s also very hard to stay in the seat. If the Senate comes at you, it’s very hard to stay in the seat. So a reductionist version of this that is a good shorthand is who can get who fired? So who has more power? The newspaper columnist who makes $200,000 a year, or the CEO who makes $200 million a year/ and it’s like, well, I know for sure that the columnist can get the CEO fired. I’ve seen that happen before I have yet to see a CEO get a columnist fired.
Lex Fridman
(01:56:48)
Did anyone ever get from the Bill Ackman assault on journalism? So Bill really showed the bullshit that happens in journalism.
Marc Andreessen
(01:56:59)
No. Because what happens is they wear it with the … And I would say to their credit, they wear it as a badge of honor, and then to their shame, they wear it as a badge of honor, which is if they’re doing the right thing, then they are justifiably priding themselves for standing up under pressure. But it also means that they can’t respond to legitimate criticism and they’re obviously terrible at that now. As I recall, he went straight to the CEO of Axel Springer that owns Insider. I happen to know the CEO O, and I think he’s quite a good CEO. Well, there’s a good example. Does the CEO Axel Springer run his own company?

(01:57:32)
So there’s a fascinating thing playing out right now. Not to dwell on these fires. But you see the pressure reveals things, right? And so if you’ve been watching what’s happening with the LA Times recently. So this guy, biotech entrepreneur buys the LA Times, whatever, eight years ago. It is just like the most radical social revolutionary thing you can possibly imagine. It endorses every crazy left-wing radical you can imagine. It endorses Karen Bass, it endorses Gavin Newsom. It’s just a litany of all the people who are currently burning the city to the ground. It’s just like endorsed every single bad person every step of the way. He’s owned it the entire time. He for the first time, I think, put his foot down right before the November election and said, we’re not … He said, “We’re going to get out of this thing where we just always endorse the Democrat.” I think he said, “We’re not endorsing for the presidency.” And the paper flipped out. It’s like our billionaire backer who’s … And I don’t know what he spends, but he must be burning 50 or a hundred million dollars a year out of his pocket to keep this thing running.

(01:58:28)
He paid 500 million for it, which is amazing. Back when people still thought these things were businesses. And then he’s probably burned another 500 million over the last decade keeping it running. And he burns probably another 50, a hundred million a year to do this. And the journalists at the LA Times hate him with the fury of a thousand suns. They just absolutely freaking despise him, and they have been attacking him. The ones that can get jobs elsewhere quit and do it, and the rest just stay and say the worst, most horrible things about him. And they want to constantly run these stories attack him. And so he has had this reaction that a lot of people in LA are having right now to this fire and to this just incredibly vivid collapse of leadership. And all these people that his paper head endorsed are just disasters.

(01:59:11)
He’s on this tour. Basically he’s decided to be the boy who says the emperor has no clothes, but he’s doing it to his own newspaper. Very smart guy. He is on a press tour and he is basically saying, yes, we did all that and we endorsed all these people and it was a huge mistake and we’re going to completely change. And his paper is in a complete internal revolt. But I go through it, which is okay, now we have a very interesting question, which is who runs the LA Times? Because for the last eight years, it hasn’t been him. It’s been the reporters. Now for the first time, the owner is showing up saying, “Oh no, I’m actually in charge,” and the reporters are saying, “No, you’re not.” It is freaking on. And so again, the Machiavellian’s mindset on this is like, okay, how is power actually exercised here? Can a guy who’s even super rich and super powerful who even owns his own newspaper, can he stand up to a full scale assault, not only by his own reporters, but by every other journalism outlet who also now thinks he’s the Antichrist?
Lex Fridman
(02:00:08)
And he is trying to exercise power by speaking out publicly and so that’s the game of power there.
Marc Andreessen
(02:00:13)
And firing people.
Lex Fridman
(02:00:13)
Firing people. Yeah.
Marc Andreessen
(02:00:15)
He has removed people and he has set new rules. He’s now at long last actually exercising prerogatives of an owner of a business, which is decide on the policies and staffing of the business. There are certain other owners of these publications that are doing similar things right now. He’s the one I don’t know so he’s the one I can talk about. But there are others that are going through the same thing right now. And I think it’s a really interesting open question in a fight between the employees and the employer it’s not crystal clear that the employer wins that one.

Bill Ackman

Lex Fridman
(02:00:43)
And just to stay on journalism for a second, we mentioned Bill Ackman. I just want to say put him in the category we mentioned before of a really courageous person. I don’t think I’ve ever seen anybody so fearless in going after, in following what he believes in publicly. That’s courage. Several things he’s done publicly has been really inspiring. Just being courageous.
Marc Andreessen
(02:01:10)
What do you think is the most impressive example?
Lex Fridman
(02:01:12)
Where he went after journalists whose whole incentive is to … It’s like kicking the beehive or whatever. You know what’s going to follow and to do that. That’s why it’s difficult to challenge journalistic organizations because they’re going to … There’s just so many mechanisms they use, including writing articles and get cited by Wikipedia and then drive the narrative and then they can get you fired, all this stuff. Bill Ackman, like a bad MFer just tweets these essays and just goes after them legally and also in the public eye. I don’t know. That was truly inspiring. There’s not many people like that in public and hopefully that inspires not just me, but many others to be courageous themselves.
Marc Andreessen
(02:02:05)
Did you know of him before he started doing this in public?
Lex Fridman
(02:02:08)
I knew of Neri, his wife, who’s a brilliant researcher and scientist. And so I admire her. Looked up to her and think she’s amazing.
Marc Andreessen
(02:02:15)
Well, the reason I ask if you knew about Bill is because a lot of people had not heard of him before, especially before October 7th and before some of the campaigns he’s been running since in public with Harvard and so forth. But he was very well known in the investment world before that. He was a so-called activist investor for … Very successful and widely respected for probably 30 years before now. And I bring that up because it turns out they weren’t for the most part battles that happened in full public view. They weren’t national stories. But in the business and world, the activist investor is a very … It’s like in the movie Taken. It’s a very specific set of skills on how to really take control of situations and how to wreck the people who you’re going up against. There’s been controversy over the years on this topic, and there’s too much detail to go into. But the defense of activist investing, which I think is valid, is these are the guys who basically go in and take stakes in companies that are being poorly managed or under-optimized. And then generally what that means is, at least the theory is that means the existing management is become entrenched and lazy, mediocre, whatever. Not you’re responding to the needs of the shareholders. Often not responding to the customers. And the activists basically go in with a minority position and then they rally support among other investors who are not activists. And then they basically show up and they force change. But they are the aggressive version of this. I’ve been involved in companies that have been on the receiving end of these where it is amazing how much somebody like that can exert pressure on situations even when they don’t have formal control. It would be another chess piece on the mechanical board of how power gets exercised. And basically what happens is the effective analysts, a large amount of time they end up taking over control of companies even though they never own more than 5% of the stock. So anyway,

(02:04:02)
So it turns out with Bill’s … It’s such a fascinating case. He has that complete skill set. And he has now decided to bring it to bear in areas that are not just companies. And two interesting things for that. One is some of these places and some of these battles are still ongoing, but number one, a lot of people who run universities or newspapers are not used to being up against somebody like this. And by the way, also now with infinitely deep pockets and lots of experience in courtrooms and all the things that go with that. But the other is through example he is teaching a lot of the rest of us the activists playbook in real time. And so the Liam Neeson skill set is getting more broadly diffused just by being able to watch and learn from him. So I think he’s having a … I would put him up there with Elon in terms of somebody who’s really affecting how all this is playing out.
Lex Fridman
(02:04:48)
But even set aside just courage and-
Marc Andreessen
(02:04:50)
Yes. Including by the way, courage to go outside of his own zone. I’ll give you an example. My venture capital firm, we have LPs. There are things that I feel like I can’t do or say because I feel like I would be bringing embarrassment or other consequences to our LPs. He has investors also where he worries about that. So a couple of things. One, it’s his willingness to go out a bit and risk his relationship with his own investors. But I will tell you the other thing, which is his investors … I know this for a fact. His investors have been remarkably supportive of him doing that. Because as it turns out, a lot of them actually agree with him. It’s the same thing he does in his activism campaigns. He is able to be the tip of the spear on something that actually a lot more people agree with.
Lex Fridman
(02:05:33)
Yeah. It turns out if you have truth behind you, it helps.
Marc Andreessen
(02:05:37)
And just again, how I started is a lot of people are just fed up.

Trump administration

Lex Fridman
(02:05:41)
You’ve been spending a bunch of time in Mar-a-Lago, in Palm Beach helping the new administration in many ways, including interviewing people who might join. So what’s your general sense about the talent, about the people who are coming into the new administration?
Marc Andreessen
(02:05:56)
So I should start by saying I’m not a member of the new administration. I’m not in the room when a lot of these people are being selected.
Lex Fridman
(02:06:03)
I believe you said unpaid intern.
Marc Andreessen
(02:06:05)
I am an unpaid intern. So I’m a volunteer when helpful, but I’m not making the decisions, nor am I in a position to speak for the administration. I don’t want to say anything that would cause people to think I’m doing that. It’s a very unusual situation where you had an incumbent president and then you had a four-year gap where he is out of office, and then you have him coming back. And as you’ll recall, there was a fair amount of controversy over the end of the first term. The specific concern was the first Trump administration, they will all say this is they didn’t come in with a team. They didn’t come into the team. And most of the institutional base of the Republican Party were Bush Republicans. And many of them had become never Trumpers. And so they had a hard time putting the team together. And then by the way, they had a hard time getting people confirmed. And so if you talk to the people who were there in the first term, it took them two to three years to even get the government in place. And then they basically only had the government in place for basically like 18 months and then COVID hit. And then the aftermath and everything and all the drama and headlines and everything.

(02:07:02)
And so the concern, including from some very smart people in the last two years has been, boy, if Trump gets a second term, is he going to be able to get a team that is as good as the team he had last time or a team that is actually not as good? Because maybe people got burned out. Maybe they’re more cynical now. Maybe they’re not willing to go through the drama. By the way, a lot of people in the first term came under their own withering legal assaults, and some of them went to prison. A lot of stuff happened. Lots of investigations, lots of legal fees, lots of bad press, lots of debanking by the way. A lot of the officials in the first Trump term got debanked, including the president’s wife and son.
Lex Fridman
(02:07:39)
Yeah. I heard you tell that story. That’s insane. That’s just insane.
Marc Andreessen
(02:07:41)
In the wake of the first term, yes. We now take out spouses and children with our ring of power. And so there’s this legitimate question as to okay, what will the team for the second term look like? At least what I’ve seen and what you’re seeing with the appointments is it looks much, much better. First of all, it just looks better than the first term and not because the people in the first term were not necessarily good, but you just have this influx of incredibly capable people that have shown up that want to be part of this and you just didn’t have that the first time. And so they’re just drawing on a much deeper, richer talent pool than they had the first time. And they’re drawing on people who know what the game is. They’re drawing on people now who know what is going to happen, and they’re still willing to do it.

(02:08:20)
And so they’re going to get, I think, some of the best people from the first term, but they’re bringing in a lot of people who they couldn’t get the first time around. And then second is there’s a bunch of people, including people in the first term where they’re just 10 years older. And so they went through the first term and they just learned how everything works. Or there are young people who just had a different point of view and now they’re 10 years older and they’re ready to go serve in government. So there’s a generational shift happening. And actually one of the interesting things about the team that’s forming up is it’s remarkably young. Some of the cabinet members and then many of the second and third level people are in their 30s and 40s, which is a big change from the gerontocracy that we’ve been under for the last 30 years.

(02:08:59)
I think the caliber has been outstanding. And we could sit here and list tons and tons of people, but the people who are running. It’s everything from the people who are running all the different departments at HHS. The number two at the Pentagon is Steve Feinberg, who’s just an incredible legend of private equity, incredible capable guy. Actually two of my partners are going in who I both think are amazing. Many, many parts of the government the people are really impressive.
Lex Fridman
(02:09:25)
Well, I think one of the concerns is actually that given the human being of Donald Trump, that there would be more tendency towards, let’s say favoritism versus meritocracy. That there’s circles of sycophancy that form. And if you’re be able to be loyal and never oppose and just basically suck up to the president, that you’ll get a position. So that’s one of the concerns. And I think you’re in a good position to speak to the degree that’s happening versus hiring based on merit and just getting great teams.
Marc Andreessen
(02:10:06)
Yeah. So look, start by saying any leader at that level, by the way, any CEO, there’s always some risk of that. That’s like a natural reality warps around powerful leaders. And so there’s always some risk to that. Of course, the good powerful leaders are very aware of that. And Trump, at this point in his life, I think is highly aware of that, at least in my interactions with him. He definitely seems very aware of that. So that’s one thing. I would just say, I think the way to look at that … And look, like I said, I don’t want to predict what’s going to happen once this whole thing starts unfolding. I would just say again, the caliber of the people who are showing up and getting the jobs, and then the fact that these are some of the most accomplished people in the business world and in the medical field. Jay Bhattacharya coming in to run NIH. I was part of the interview team for a lot of the HHS folks.
Lex Fridman
(02:10:52)
Nice. Jay’s amazing. Oh, I was so happy to see that.
Marc Andreessen
(02:10:55)
So I literally got … This is the story. I got to the transition office for one of the days of the HHS interviews, and I was on one of the interview interviewing teams. I didn’t know who the candidates were, and they gave us the sheet in the beginning, and I go down the sheet and I saw Jay’s name. I almost physically fell on my chair. And I was just like … I happen to know Jay. I happen to know Jay, and I respect him enormously. And then he proved himself under this … Talk about a guy who proved himself under extraordinary pressure over the last five years.
Lex Fridman
(02:11:20)
And then go radical under the pressure. He maintained balance and thoughtfulness and depth. Incredibly-
Marc Andreessen
(02:11:28)
Very serious, very analytical, very applied. Yes. A hundred percent. Tested under pressure came out. The more people look back at what he said and did. None of us perfect, but overwhelmingly insightful throughout that whole period. We would all be much better off today had he been in charge of the response. And so just an incredibly capable guy. And look, and then he learned from all that. He learned a lot in the last five years. And so the idea that somebody that could be head of NIH as compared to the people we’ve had is just breathtakingly. It’s just a gigantic upgrade. And then Marty McAree coming.
Marc Andreessen
(02:12:00)
It is just a gigantic upgrade. And then Marty Makary coming in to run FDA, exact same thing. The guy coming to run a CDC, exact same thing. I’ve been spending time with Dr. Oz. So again, I’m not on these teams, I’m not in the room, but I’ve been spending enough time trying to help that his level of insight into the healthcare system, it’s astounding. And it comes from being a guy who’s been in the middle of the whole thing and been talking to people about this stuff, and working on it and serving as a doctor himself and in medical systems for his entire life. He’s like a walking encyclopedia on these things. And very dynamic, very charismatic, very smart, organized, effective. So to have somebody like that in there. And so anyway, I have 30 of these stories now across all these different positions. And then to be quite honest, you do the compare and contrast to the last four years, and these people are not in the same ballpark, they’re just wildly better. And so pound for pound this maybe the best team in the White House since, I don’t even know, maybe the 90s, maybe the 30s, maybe the 50s. Maybe Eisenhower had a team like this or something, but there’s a lot of really good people in there now.

DOGE

Lex Fridman
(02:13:16)
Yeah, the potential for change is certainly extremely high. Can you speak to DOGE? What’s the most wildly successful next two years for DOGE, can you imagine? Maybe also can you think about the trajectory that’s the most likely and what kind of challenges would it be facing?
Marc Andreessen
(02:13:36)
Yeah, so start by saying, again, disclaimer, I have to say, I’m not on DOGE, I’m not a member of DOGE.
Lex Fridman
(02:13:43)
We should say there’s about 10 lawyers in the room, they’re staring. No, I’m just kidding.
Marc Andreessen
(02:13:48)
Both the angels and the devils on my shoulder are literally [inaudible 02:13:51]. So I’m not speaking for DOGE, I’m not in charge of DOGE. Those guys are doing it, I’m not doing it. But again, I’m volunteering to help as much as I can and I’m 100% supportive. Yeah, so look, I think the way to think, the basic outlines are in public, which is it’s a time limited basically commission. It’s not a formal government agency. It’s a time limited, 18 month. In terms of implementation, it will advise the executive branch. And so the implementation will happen through the White House. And the president has total latitude on what he wants to implement. And then basically what I think about it is three streams, target sets, and they’re related but different. So money, people, and regulations. And so the headline number they put as the $2 trillion number, and there’s already disputes over that and whatever, and there’s a whole question there. But then there’s the people thing.

(02:14:44)
And the people thing is interesting, because you get into these very fascinating questions. And I’ve been doing this, I won’t do this for you as a pop quiz, but I do this for people in government as a pop quiz and I can stump them every time. Which is, A, how many federal agencies are there? And the answer is somewhere between 450 and 520, and nobody’s quite sure. And then the other is how many people work for the federal government? And the answer is something on the order, I forget, but like 4 million full-time employees and maybe up to 20 million contractors, and nobody’s quite sure. And so there’s a large people component to this. And then by the way, there’s a related component to that, which is how many of them are actually in the office? And the answer is not many, most of the federal buildings are still empty.

(02:15:27)
And then there’s questions of are people working from home or are we actually working from home? So there’s the people dimension, and of course the money and the people are connected. And then there’s the third, which is the regulation thing. And I described earlier how basically our system of government is much more now based on regulations than legislation. Most of the rules that we all live under are not from a bill that went through Congress, they’re from an agency that created a regulation. That turns out to be very, very important. So one is Elon had already described the DOGE wants to do broad-based regulatory relief, and Trump has talked about this, and basically get the government off of people’s backs and liberate the American people to be able to do things again. So that’s part of it. But there’s also something else that’s happened, which is very interesting, which was there were a set of Supreme Court decisions about two years ago that went directly after the idea that the executive branch can create regulatory agencies, and issue regulations and enforce those regulations without corresponding congressional legislation.

(02:16:20)
And most of the federal government that exists today, including most of the departments and most of the rules and most of the money and most of the people, most of it is not enforcing laws that Congress passed. Most of it is regulation. And the Supreme Court basically said, “Large parts, large to maybe all of that regulation that did not directly result from a bill that went through Congress, the way that the cartoon said that it should, may not actually be legal. Now, the previous White House, of course, was super in favor of big government. They did nothing based on this, they didn’t pull anything back in. But the new regime, if they choose to, could say, “Look, the thing that we’re doing here is not challenging the laws, we’re actually complying with the Supreme Court decision that basically says we have to unwind a lot of this and we have to unwind the regulations which are no longer legal, constitutional. We have to unwind the spend and we have to unwind the people.

(02:17:16)
And that’s how you get from basically you connect the thread from the regulation part back to the money part back to the people part. They have work going on all three of these threads. They have, I would say, incredibly creative ideas on how to deal with this. I know lots of former government people who 100% of them are super cynical on this topic, and they’re like, “This is impossible, this could never possibly work.” And I’m like, “Well, I can’t tell you what the secret plans are, but blow my mind.” And all three of those, they have ideas that are really quite amazing, as you’d expect from the people involved. And so over the course of the next few months, that’ll start to become visible. And then the final thing I would say is this is going to be very different than attempts, there have been other programs like this in the past. The Clinton-Gore administration had one and then there were others before that, Reagan had one. The difference is this time, their social media,

(02:18:13)
It’s interesting, one of the reasons people in Washington are so cynical is because they know all the bull shit. They know all the bad spending and all the bad rules. Look, we’re adding a trillion dollars to the national debt every 100 days right now. And that’s compounding, and it’s now passing the size of the defense department budget and it’s compounding, and pretty soon it’s going to be adding a trillion dollars every 90 days, and then it’s going to be adding a trillion dollars every 80 days, and then it’s going to be a trillion dollars every 70 days. And then if this doesn’t get fixed, at some point we enter a hyperinflationary spiral and we become Argentina or Brazil, and [inaudible 02:18:44]. And so everybody in D.C. knows that something has to be done, and then everybody in D.C. knows for a fact that it’s impossible to do anything.

(02:18:54)
They know all the problems and they also know the sheer impossibility of fixing it. But I think what they’re not taking into account, what the critics are not taking into account is these guys can do this in the full light of day. And they can do it on social media, they can completely bypass the press, they can completely bypass the cynicism. They can expose any element of unconstitutional or silly government spending. They can run victory laps every single day on what they’re doing. They can bring the people into the process. And again, if you think about it, this goes back to our Machiavellian structure, which is if you think about, again, you’ve got democracy, oligarchy, monarchy, rule of the many, rule of the few, rule of the one. You could think about what’s happening here as a little bit of a sandwich, which is we don’t have a monarch, but we have a president, rule of the one with some power.

(02:19:37)
And then we have the people who can’t organize, but they can be informed and they can be aware, and they can express themselves through voting and polling. So there’s a sandwich happening right now is the way to think about it, which is you’ve got basically rule of one combining with rule of many. And rule of many is they do get to vote, the people do get to vote basically, and then essentially Congress and this permanent bureaucratic class in Washington as the oligarchy in the middle. And so the White House plus the people I think have the power to do all kinds of things here, and I think that would be the way I would wash it.
Lex Fridman
(02:20:11)
The transparency. Elon, just by who he is is incentivized to be transparent, and show the bull shit in the system and to celebrate the victories. So it’s going to be so exciting. It honestly just makes government more exciting, which is a win for everybody.
Marc Andreessen
(02:20:31)
These people are spending our money. These people have enormous contempt for the taxpayer. Okay, here’s the thing you hear in Washington, here’s one of the things. So the first thing you hear is, “This is impossible, they’ll be able to do nothing.” And then, yeah, I walk them through this and it starts to dawn on them that this is a new kind of thing. And then they’re like, “Well, it doesn’t matter, because all the money is in entitlements and the debt and the military.” And so yeah, you’ve got this silly, fake whatever, NPR funding or whatever, and just it’s a rounding error and it doesn’t matter. And you look it up in the budget and it’s like, whatever, $500 million or $5 billion, or it’s the charging stations that don’t exist. It’s the $40 billion of charging stations and they build eight charging stations, or it’s the broadband internet plan that delivered broadband to nobody and cost you $30 billion, so these boondoggles. And what everybody in Washington says is that $30 billion is a rounding error on the federal budget, it doesn’t matter. Who cares if they make it go away? And of course, any taxpayer is like, “What the fuck?”
Lex Fridman
(02:21:32)
What do you mean?
Marc Andreessen
(02:21:33)
It’s $30 billion. And the press is in on this too, and then the experts are like, “Well, it doesn’t matter because it’s rounding error.” No, it’s $30 billion. And if you’re this cavalier about $30 billion, imagine how cavalier you are about the three trillion. Then there’s the, okay, $30 billion. Is $30 billion a lot of the federal budget in percentage? No, it’s not, but $30 billion divided by, do the math, $30 billion divided by let’s say 300 million taxpayers. What’s that math expert?
Lex Fridman
(02:21:34)
$100.
Marc Andreessen
(02:21:57)
$100 per taxpayer per year. Okay, so $100 to an ordinary person working hard every day to make money and provide for their kids. $100 is a meal out, it’s a trip to the amusement park. It’s the ability to buy additional educational materials. It’s the ability to have a babysitter to be able to have a romantic relationship with your wife. There’s 100 things that that person can do with $100 that they’re not doing because it’s going to some bull shit program that is being basically where the money’s being looted out in the form of just ridiculous ridiculousness and graft. And so the idea that that $30 billion program is not something that is a very important thing to go after, the level of contempt for the taxpayer is just off the charts.

(02:22:40)
And then that’s just one of those programs, there’s 100 of those programs. And they’re all just like that, it’s not like any of this stuff is running well. The one thing we know is that none of this stuff is running well, we know that for sure. And we know these people aren’t showing up to work, and we know that all this crazy stuff is happening. Do you remember Elon’s story of what got the Amish to turn out to vote in Pennsylvania? Oh, okay. So Pennsylvania is like a wonderful state, great history. It has these cities like Philadelphia that have descended other cities into just complete chaos, violent madness, and death. And the federal government has just let it happen, these incredibly violent places.

(02:23:16)
And so the Biden administration decided that the big pressing law enforcement thing that they needed to do in Pennsylvania was that they needed to start raiding Amish farms to prevent them from selling raw milk with armed raids. And it turns out it really pissed off the Amish. It turns out they weren’t willing to drive to the polling places because they don’t have cars, but if you came and got them, they would go and they would vote. And that’s one of the reasons why Trump won. Anyway, so the law enforcement agencies are off working on crazy things. The system’s not working. And so you add up, just pick $130 billion programs, all right, now you’re okay. Math major, 100 times 100.
Lex Fridman
(02:23:52)
10,000.
Marc Andreessen
(02:23:53)
$10,000, okay. $10,000 per tax payer per year.
Lex Fridman
(02:23:57)
But it’s also not just about money, obviously money is a hugely important thing, but it’s the cavalier attitude that then in the ripple effect of that, it makes it so nobody wants to work in government and be productive. It makes it so that it breeds corruption, it breeds laziness. It breeds secrecy because you don’t want to be transparent about having done nothing all year, all this kind of stuff. And you now want to reverse that so that it will be exciting for the future to work in government, because the amazing thing if you’re the steelman government is you can do shit at scale. You have money and you can directly impact people’s lives in a positive sense at scale. It’s super exciting. As long as there’s no bureaucracy that slows you down, or not huge amounts of bureaucracy that slows you down significantly.
Marc Andreessen
(02:24:53)
Yeah. So here’s the trick, this blew my mind. Because once you open the hellmouth of looking into the federal budget, you learn all kinds of things. So there is a term of art in government called impoundment. So if you’re like me, you’ve learned this the hard way when your car has been impounded. The government meaning of impoundment, the federal budget meaning is a different meaning. Impoundment is as follows. The constitution requires Congress to authorize money to be spent by the executive branch. So the executive branch goes to Congress, says, “We need money X.” Congress does their thing. They come back and they say, “You can have money Y.” The money’s appropriated from Congress, the executive branch spends it on the military or whatever they spend it on, or on roads to nowhere or charging stations to nowhere or whatever. And what’s in the constitution is the Congress appropriates the money. Over the last 60 years, there has been an additional interpretation of appropriations applied by the courts and by the system, which is the executive branch not only needs Congress to appropriate X amount of money, the executive branch is not allowed to underspend.
Lex Fridman
(02:25:56)
Yeah, I’m aware of this. I’m aware of this.
Marc Andreessen
(02:26:00)
And so there’s this thing that happens in Washington at the end of every fiscal year, which is September 30th, and it’s the great budget flush. And any remaining money that’s in the system that they don’t know how to productively spend, they deliberately spend it unproductively, to the tune of hundreds and hundreds of billions of dollars. A president that doesn’t want to spend the money can’t not spend it.

(02:26:20)
Like, okay, A, that’s not what’s in the constitution. And there’s actually quite a good Wikipedia page that goes through the great debate on this that’s played out in the legal world over the last 60 years. And basically, if you look at this with anything resembling I think an open mind, you’re like, “All right, this is not what the founders meant.” And then number two, again, we go back to this thing of contempt. Can you imagine showing up and running the government like that and thinking that you’re doing the right thing, and not going home at night and thinking that you’ve sold your soul? I actually think you headed it a really good point, which is it’s even unfair to the people who have to execute this because it makes them bad people, and they didn’t start out wanting to be bad people. And so there is stuff like this.
Lex Fridman
(02:27:01)
Yeah, everywhere.
Marc Andreessen
(02:27:01)
Everywhere. And so we’ll see how far these guys get. I am extremely encouraged, what I’ve seen so far.

H1B and immigration

Lex Fridman
(02:27:07)
It seems like a lot of people will try to slow them down, but yeah, I hope they get far. Another difficult topic, immigration. What’s your take on the, let’s say, heated H-1B visa debate that’s going on online and legal immigration in general?
Marc Andreessen
(02:27:22)
I should start by saying I am not involved in any aspect of government policy on this. I’m not planning to be, this is not an issue that I’m working on or that I’m going to work on. This is not part of the agenda of what the firm is doing, my firm is doing. So I’m not in the new administration or the government, I’m not planning to be, so purely just personal opinion. So I would say I would describe this as I have a complex or nuanced, hopefully nuanced view on this issue that’s maybe a little bit different than what a lot of my peers have. And I thought about this, I didn’t say anything about it all the way through the big debate over Christmas, but I thought about it a lot and read everything. I think what I realized is that I just have a very different perspective on some of these things, and the reason is because of the combination of where I came from and then where I ended up.

(02:28:09)
Let’s start with this, where I ended up, in Silicon Valley, and I have made the pro high-skilled immigration argument many, many times, the H-1B argument many times. In past lives, I’ve been in D.C. many times arguing with prior administrations about this, always on the side of trying to get more H-1B’s and trying to get more high-skilled immigration. And I think that argument is very strong and very solid, and has paid off for the US in many, many ways. And we can go through it, but I think it’s the argument everybody already knows, it’s like the stock. You take any Silicon Valley person, you press the button and they tell you why we need to brain drain the world to get more H-1B’s. So everybody gets that argument.
Lex Fridman
(02:28:46)
So it’s basically, just to summarize, it’s a mechanism by which you can get super smart people from the rest of the world, import them in, keep them here to increase the productivity of the US companies.
Marc Andreessen
(02:28:58)
And then it’s not just good for them and it’s not just good for Silicon Valley or the tech industry, it’s good for the country because they then create new companies and create new technologies and create new industries that then create many more jobs for Americans, native born Americans, than would’ve previously existed. And so it’s a positive, some flywheel thing where everybody wins. Everybody wins, there are no trade-offs, it’s all absolutely glorious in all directions. There cannot possibly be a moral argument against it under any circumstances. Anybody who argues against it is obviously doing so from a position of racism, is probably a fascist and a Nazi. That’s the thing, and like I said, I’ve made that argument many times. I’m very comfortable with that argument. And then I’d also say, look, I would say number one, I believe a lot of it, I’ll talk about the parts I don’t believe, but I believe a lot of it.

(02:29:43)
And then the other part is, look, I benefit every day. I always describe it as I work in the United Nations, my own firm and our founders and our companies and the industry and my friends are just this amazing panoply, cornucopia of people from all over the world. And I’ve worked, I don’t know, at this point with people from, it’s got to be, I don’t know, 80 countries or something, and hopefully over time it’ll be the rest as well. And it’s been amazing, and they’ve done many of the most important things in my industry and it’s been really remarkable. So that’s all good. And then there’s just the practical version of the argument, which is we are the main place these people get educated anyway. The best and the brightest tend to come here to get educated. And so this is the old Mitt Romney, staple a green card to maybe not every university degree, but every technical degree. Maybe the sociologists we could quibble about, but the roboticists for sure, for sure. For sure, we can all agree that-
Lex Fridman
(02:30:40)
At least I won you over on something today.
Marc Andreessen
(02:30:42)
Well, no, I’m exaggerating for effect.
Lex Fridman
(02:30:45)
And I lost you, I had you for half a second.
Marc Andreessen
(02:30:48)
I haven’t gotten to the other side of the argument yet.
Lex Fridman
(02:30:49)
Okay, thank you.
Marc Andreessen
(02:30:50)
So surely we can all agree that we need to staple a green card.
Lex Fridman
(02:30:54)
The rollercoaster is going up.
Marc Andreessen
(02:30:55)
The rollercoaster is ratcheting slowly up. So yeah, so surely we can all agree that the roboticists should all get green cards. And again, there’s a lot of merit to that, obviously. Look, we want the US to be the world leader in robotics. What’s step one to being the world leader in robotics is have all the great robotics people. Unlike the underpants, it’s like a very straightforward formula. All right, that’s all well and good, all right, but it gets a little bit more complicated because there is a argument that’s right underneath that that you also hear from these same people. And I have made this argument myself many times, which is we need to do this because we don’t have enough people in the US who can do it otherwise. We have all these unfilled jobs, we’ve got all these companies that wouldn’t exist.

(02:31:33)
We don’t have enough good founders, we don’t have enough engineers, we don’t have enough scientists. Or then the next version of the argument below that is our education system is not good enough to generate those people, which is a weird argument by the way. Because our education system is good enough for foreigners to be able to come here preferentially in a very large number of cases, but somehow not good enough to educate our own native foreign people. So there’s little cracks in the matrix that you can stick your fingernail into and wonder about and we’ll come back to that one. But at least, yes, our education system has its flaws. And then underneath that is the argument that Vivek made, which is we have cultural rot in the country and native-born people in the country don’t work hard enough, and spend too much time watching TV and TikTok and don’t spend enough time studying differential equations.

(02:32:19)
And again, it’s like, all right, yeah, there’s a fair amount to that. There’s a lot of American culture that is, there’s a lot of frivolity, we have well-documented social issues on many fronts, many things that cut against having a culture of just straightforward, high achievement and effort and striving. But anyway, those are the basic arguments. But then I have this other side of my personality and thought process, which is, well, I grew up in a small farming town of rural Wisconsin, the rural Midwest, and it’s interesting, there’s not a lot of people who make it from rural Wisconsin to high tech.

(02:32:54)
And so it’s like, all right, why is that exactly? And I know this, I’m an aberration. I was the only one from anybody I ever knew who ever did this. I know what an aberration I am and I know exactly how that aberration happened, and it’s a very unusual set of steps, including many that were just luck. But there is in no sense a talent flow from rural Wisconsin into high tech, like not at all. There is also in no sense a talent flow from the rest of the Midwest into high tech. There is no talent flow from the south into high tech. There is no flow from the Sunbelt into high tech. There’s no flow from the deep south into high tech. Literally, it’s like the blanks. There’s this whole section of the country where the people just for some reason don’t end up in tech.

(02:33:38)
Now, that’s a little bit strange, because these are the people who put a man on the moon. These are the people who built the World War II War Machine. These are the people, at least their ancestors are the people who built the second industrial revolution, and built the railroads and built the telephone network and built logistics and transportation in the auto industry. The auto industry was built in Cleveland and Detroit. And so at least these people’s parents and grandparents and great grandparents somehow had the wherewithal to build all of these amazing things, invent all these things.

(02:34:07)
And then there’s many, many, many, many stories in the history of American invention and innovation and capitalism, where you had people who grew up in the middle of nowhere, Philo Farnsworth who invented the television, and just tons and tons of others, endless stories like this. Now you have a puzzle and the conundrum, which is like, okay, what is happening on the blank spot of the map? And then of course, you also can’t help noticing that the blank spot on the map, the Midwest, the South, you’ve also just defined Trump country, the Trump voter base.

(02:34:35)
And it’s like, oh, well, that’s interesting. How did that happen? And so either you really, really, really have to believe the very, very strong version of the Vivek thesis or something, where you have to believe that that basically culture, the whole civilization in the middle of the country and the south of the country is so deeply flawed, either inherently flawed or culturally flawed, such that for whatever reason, they’re not able to do the things that their parents and grandparents were able to do, and that their peers are able to do. Or something else is happening. Would you care to guess on what else is happening?
Lex Fridman
(02:35:03)
You mean what, affirmative action?
Marc Andreessen
(02:35:05)
Affirmative action. Think about this, this is very entertaining. What are the three things that we know about affirmative action? It is absolutely 100% necessary, however, it cannot explain the success of any one individual, nor does it have any victims at all.
Lex Fridman
(02:35:25)
That could explain maybe disproportionate, but surely it doesn’t explain why you’re probably the only person in Silicon Valley from Wisconsin.
Marc Andreessen
(02:35:34)
What educational institution in the last 60 years has wanted farm boys from Wisconsin?
Lex Fridman
(02:35:38)
But what institution rejected farm boys from Wisconsin?
Marc Andreessen
(02:35:42)
All of them.
Lex Fridman
(02:35:43)
All of them.
Marc Andreessen
(02:35:43)
Of course. Okay, so we know this, we know this. The reason we know this is because of the Harvard and UNC Supreme Court cases. This was three years ago, these were big court cases. Because the idea of affirmative action has been litigated for many, many, many years and through many court cases, and the Supreme Court repeatedly in the past had upheld that it was a completely legitimate thing to do. And there’s basically two categories of affirmative action that really matter. One is the admissions into educational institutions and then the other is jobs, getting hired. Those are the two biggest areas. The education one is super potent, has been a super potent political issue for a very long time for all… People have written and talked about this for many decades, I don’t need to go through it. There’s many arguments for why it’s important, there’s many arguments as to how it could backfire. It’s been this thing.

(02:36:25)
But the Supreme Court upheld it for a very long time. The most recent ruling, I’m not a lawyer, I don’t have the exact reference in my head, but there was a case in 2003 that said that Sandra Day O’Connor famously wrote that although it had been 30 years of affirmative action and although it was not working remotely as it had been intended, she said that, well, basically we need to try it for another 25 years. But she said basically as a message to future Supreme Court justices, if it hasn’t resolved basically the issues it’s intended to resolve within 25 years, then we should probably call it off. By the way, we’re coming up on the 25 years, it’s a couple of years away. The Supreme Court just had these cases, it’s a Harvard case and I think a University of North Carolina case.

(02:37:07)
And what’s interesting about those cases is the lawyers in those cases put a tremendous amount of evidence into the record of how the admissions decisions actually happen at Harvard and happen at UNC. And it is like every bit as cartoonishly garish and racist as you could possibly imagine, because it’s a ring of power. And if you’re an admissions officer at a private university or an administrator, you have unlimited power to do what you want, and you can justify any of it under any of these rules or systems. And up until these cases, it had been a black box where you didn’t have to explain yourself and show your work. And what the Harvard and UNC cases did is they basically required showing the work. And there was all kinds of phenomenal detail, number one is there were text messages in there that will just curl your hair, of students being spoken of and just crude racial stereotypes that would just make you want to jump out the window. It’s horrible stuff.

(02:37:58)
But also, there was statistical information. And of course, the big statistical kicker to the whole thing is that at top institutions, it’s common for different ethnic groups to have different cutoffs for SAT that are as wide as 400 points. So different groups. So specifically Asians need to perform at 400 SAT points higher than other ethnicities in order to actually get admitted into these. White people are a part of this, but Asians are a very big part of this. And actually the Harvard case is actually brought by an activist on behalf of actually the Asian students who are being turned away. And it’s the cliche now in the valley and in the medical community, which is like, if you want a super genius you hire an Asian from Harvard, because they are guaranteed to be freaking Einstein. Because if they weren’t, they were never getting admitted. Almost all the qualified Asians get turned away.

(02:38:47)
So they’ve been running this, it’s a very, very explicit, very, very clear program. This, of course, has been a third rail of things that people are not supposed to discuss under any circumstances. The thing that has really changed the tenor on this is I think two things. Number one, those Supreme Court cases, the Supreme Court ruled that they can no longer do that. I will tell you, I don’t believe there’s a single education institution in America that is conforming with the Supreme Court ruling, I think they’re all flagrantly ignoring it. And we could talk about that.
Lex Fridman
(02:39:14)
Mostly because of momentum probably, or what?
Marc Andreessen
(02:39:16)
They are trying to make the world a better place. They’re trying to solve all these social problems, They are trying to have diverse student populations. They are trying to live up to the expectations of their donors. They’re trying to make their faculty happy. They are trying to have their friends and family think that they’re good people. They’re trying to have the press write nice things about them. It’s nearly impossible for them. And to be clear, nobody has been fired from an admissions office for 25 years and prior, what the Supreme Court now is ruled to be illegality. And so they’re all the same people under the exact same pressures. And so the numbers are moving a little bit, but I don’t know anybody in the system who thinks that they are complying with Supreme Court. Like who’s in charge, in the rank ordering of who rules who, the university’s rule the Supreme Court way more than the Supreme Court rules the universities.

(02:40:05)
Well, another example of that is I think that every sitting member of the Supreme Court right now went to either Harvard or Yale, the level of incestuousness here is… Anyway, so there’s that. And so this has been running for a very long time. So one is the Harvard and UNC cases gave up the game, number one, or at least showed what the mechanism was. And then number two, the other thing is obviously the aftermath of October 7th, and what we discovered was happening with Jewish applicants and what was happening at all the top institutions for Jewish applicants was they were being managed down, either being actively managed down as a percentage of the base. And let’s say I’ve heard reports of extremely explicit basically plans to manage the Jewish admissions down to their representative percentage of the US population, which is 2%. And there’s a whole backstory here, which is 100 years ago, Jews were not admitted into a lot of these institutions, and then there was a big campaign to get them in.

(02:40:57)
Once they could get in, they immediately became 30% of these institutions because there are so many smart, talented Jews. So it went from 0% to 30%, and then the most recent generation of leadership has been trying to get it done to 2%. And a lot of Jewish people, at least a lot of Jewish people I know, they kind of knew this was happening but they discovered it the hard way after October 7th. So basically the Supreme Court case meant that you could address this in terms of the Asian victims. The October 7th meant that you could address it in terms of the Jewish victims. And for sure, both of those groups are being systematically excluded. And then of course, there’s the thing that you basically can’t talk about, which is all the white people are being excluded. And then it turns out it’s also happening to black people, and this is the thing that blew my freaking mind when I found out about it.

(02:41:44)
So I just assumed that this was great news for American Blacks, because obviously if Whites, Asians, and Jews are being excluded, then the whole point of this in the beginning was to get the Black population up, and so this must be great for American Blacks. So then I discovered this New York Times article from 2004 called Blacks are Being Admitted into Top Schools at Greater Numbers, but which ones? And by the way, this is in the New York Times, this is not in, whatever, The National Review, this is New York Times, 2004. And the two authorities that were quoted in the story are Henry Louis Gates, who’s the dean of the African-American Studies community in the United States, super brilliant guy. And then Lani Guinier, she was a potential Supreme Court appointee under, I think she was a close friend of Hillary Clinton.

(02:42:32)
And there was for a long time, she was on the short list for Supreme Court. So one of the top jurists, lawyers in the country, but both Black, legendarily successful in the academic and legal worlds and Black. And they are quoted as the authorities in this story, and the story that they tell, it’s actually amazing. And by the way, it’s happening today in education institutions and it’s happening in companies, and you can see it all over the place, and the government. Which is at least at that time, the number was half of the Black admits into a place like Harvard were not American-born Blacks, they were foreign-born Blacks, specifically Northern African, generally Nigerian or West Indian.

(02:43:18)
And by the way, many Nigerians and Northern Africans have come to the US and have been very successful. Nigerian-Americans as a group way outperform, they’re just a super smart cohort of people. And then West Indian Blacks in the US are incredibly successful. Most recently, by the way, Kamala Harris, as well as Colin Powell, just two examples of that. And so basically what Henry Louis Gates and Lani Guinier said in the story is Harvard is basically struggling to either, whatever it was, identify, recruit, make successful, whatever it was, American-born native Blacks, and so therefore they were using high-skill immigration as an escape hatch to go get Blacks from other countries. And then this was 2004 when you could discuss such things, obviously that is a topic that nobody has discussed since, it has sailed on. All of the DEI programs of the last 20 years have had this exact characteristic.

(02:44:08)
There’s large numbers of Black people in America who are fully aware of this and are like, “It’s obviously not us that are getting these slots, we’re literally competing with people who are being imported.” And if you believe in the basis of affirmative action, you were trying to make up for historical injustice of American Black slavery. So the idea that you import somebody from Nigeria that never experienced that is tremendously insulting to Black Americans. Anyway, so you can see where I’m heading with this. We have been in a 60-year social engineering experiment to exclude native-born people from the educational slots and jobs that high-skill immigration has been funneling foreigners into. And so it turns out it’s not a victim-free thing, there’s 100%, there’s victims. Because why? There’s only so many, for sure there’s only so many education slots, and then for sure, there’s only so many of these jobs. Google only hires so many, whatever, level seven engineers. And so that’s the other side of it, and so you’re a farm boy in Wisconsin.
Lex Fridman
(02:45:00)
So, that’s the other side of it. And so, you’re a farm boy in Wisconsin, or a Black American whose ancestors arrived here on a slave ship, 300 years ago, in Louisiana, or a Cambodian immigrant in the Bronx, and you are a kid, or a Jewish immigrant, or from a very successful Jewish family, and for three generations, you and your parents and grandparents went to Harvard, and what all of those groups know is the system that has been created is not for them. It’s designed specifically to exclude them, and then what happens is all of these tech people show up in public and say, yeah, let’s bring in more foreigners. So, anyway, so the short version of it is, you can’t anymore, I don’t think, just have the “high-skilled immigration,” conversation for either education or for employment without also having the DEI conversation.

(02:45:53)
And then DEI is just another word for affirmative action, so it’s the affirmative action conversation. And you need to actually deal with this at substance and to see what’s actually happening to people, you needed to join these topics. And I think it is much harder to make the moral claim for high-skilled immigration given the extent to which DEI took over both the education process and the hiring process.
Marc Andreessen
(02:46:15)
So, first of all, that was brilliantly laid out, the nuance of it. So, just to understand, it’s not so much a criticism of H-1B, high-skilled immigration, it’s that there needs to be more people saying, yay, we need more American- born hires.
Lex Fridman
(02:46:31)
So, I spent the entire Christmas holiday reading every message on this and not saying anything, and what I was… Which you know me well enough to know that’s a serious level of-
Marc Andreessen
(02:46:40)
Yeah, that was very Zen.
Lex Fridman
(02:46:41)
Yes, thank you, thank you. No, it wasn’t, there was tremendous rage on the other side of it, but I suppressed it. So, I was waiting for the dog that didn’t bark, and the dog that didn’t bark was I did not… And tell me if you saw one. I did not see a single example of somebody pounding the table for more high-skilled immigration, who was also pounding the table to go get more smart kids who are already here into these educational institutions and into these jobs. I didn’t see a single one.
Marc Andreessen
(02:47:07)
That’s true, I think I agree with that. There really was a divide.
Lex Fridman
(02:47:12)
But it was literally, it was like the proponents of high-skilled immigrant… And again, this was me for a very long time. I kind of took myself by surprise on this because I had the much, say, simpler version of this story for a very long… Like I said, I’ve been in Washington many times under past presidents, lobbying for this. By the way, never made any progress, which we could talk about, it never actually worked. But I’ve been on the other side of this one. But I was literally sitting there being like, all right, which of these super geniuses, many of whom by the way are very successful, high-skilled immigrants or children of high-skilled immigrants, which of these super geniuses are going to say, actually we have this incredible talent source here in the country? Which again, to be clear, I’m not talking about white people, I’m talking about native-born Americans, whites, Asians, Jews, Blacks, for sure. For sure, for sure, those four groups,
Marc Andreessen
(02:47:55)
But also white people.
Lex Fridman
(02:47:57)
Yeah, and also white people.
Marc Andreessen
(02:47:59)
People that are making the case for American-born hires are usually not also supporting H-1B. It’s an extreme divide, and those people, they’re making that case are often not making it in a way that’s… Making it in quite a radical way, let’s put it this way.
Lex Fridman
(02:48:20)
Yeah, yeah. But you have this interesting thing, you have a split between the sides that I’ve noticed, which is one side has all of the experts. And I’m using air quote for people listening to audio, I’m making quotes in the air with my fingers as vigorously as I can. One side has all the certified experts, the other side just has a bunch of people who are like, they know that something is wrong and they don’t quite know how to explain it. And what was so unusual about the Harvard UNC cases, by the way, in front of the Supreme Court is they actually had sophisticated lawyers for the first time in a long time actually put all this evidence together and actually put it in the public record. They actually had experts, which is just really rare.

(02:48:51)
Generally what you get is you get… Because if you don’t have experts, what do you have? You know something is wrong, but you have primarily an emotional response. You feel it, but can you put it into the words and tables and charts that a certified expert can? No, you can’t, that’s not who you are. That doesn’t mean that you’re wrong, and it also doesn’t mean that you have less of a moral stance. And so, it’s just like, all right… Now, by the way, look, I think there are ways to square the circle, I think there’s a way to have our cake and eat it too, I think there’d be many ways to resolve this. I think, again, I think the way to do it is to look at these issues combined, look at DEI combined with high-skilled immigration. It so happens that DEI is under much more scrutiny today than it has been for probably 20 years, affirmative action is. The Supreme Court did just rule that it is not legal for universities to do that, they are still doing it, but they should stop.

(02:49:46)
And then, there are more and more, you’ve seen more companies now also ditching their DEI programs, in part… That’s happening for a bunch of reasons, but it’s happening in part because a lot of corporate lawyers will tell you that the Supreme Court rulings in education either already apply to businesses, or it just is a clear foreshadowing the Supreme Court will rule on new cases that will ban in businesses. And so, there is a moment here to be able to look at this on both sides. Let me add one more nuance to it though, that makes it even more complicated. So, the cliché is we’re going to brain drain the world, you’ve heard that? We’re going to take all the smart people from all over the world, we’re going to bring them here, we’re going to educate them, and then we’re going to keep them, and then they’re going to raise their families here, create businesses here, create jobs here, right?
Marc Andreessen
(02:50:28)
In the cliché, that’s a super positive thing.
Lex Fridman
(02:50:30)
Yeah. Okay, so what happens to the rest of the world?
Marc Andreessen
(02:50:35)
They lose?
Lex Fridman
(02:50:36)
Well, how fungible are people? How many highly ambitious, highly conscientious, highly energetic, high achieving, high IQ, super geniuses are there in the world? And if there’s a lot, that’s great, but if there just aren’t that many, and they all come here, and they aren’t where they would be otherwise, what happens to all those other places? So, it’s almost impossible for us here to have that conversation, in part because we become incredibly uncomfortable as a society talking about the fact that people aren’t just simply all the same, which is a whole thing we could talk about, but also we are purely the beneficiary of this effect. We are brain draining the world, not the other way around. There’s only four… So, if you look at the flow of high-skilled immigration over time, there’s only four permanent sinks of high-skilled immigration places people go. It’s the US, Canada, the UK, and Australia.
Marc Andreessen
(02:51:31)
Oh, Australia.
Lex Fridman
(02:51:32)
It’s four of the five, five eyes. It’s the major Anglosphere countries. And so, for those countries, this seems like a no-lose proposition, it’s all the other countries that, basically, what we four countries have been doing is draining all the smart people out. It’s actually much easier for people in Europe to talk about this I’ve discovered, because the Eurozone is, whatever, 28 countries, and within the Eurozone, the high-skilled people over time have been migrating to originally the UK, but also specifically I think it’s the Netherlands, Germany, and France. But specifically, they’ve been migrating out of the peripheral Eurozone countries. And the one where this really hit the fan was in Greece. So, Greece falls into chaos, disaster, and then you’re running the government in Greece and you’re trying to figure out how to put an economic development plan together, all of your smart young kids have left, what are you going to do?

(02:52:19)
By the way, this is a potential… I know you care a lot about Ukraine, this is a potential crisis for Ukraine. In part because of this, because we enthusiastically recruit Ukrainians, of course, and so we’ve been brain draining Ukraine for a long time, but also, of course, war does tend to cause people to migrate out. And so, when it comes time for Ukraine to rebuild as a peaceful country, is it going to have the talent base even that it had five years ago, is a very big and important question. By the way, Russia, we have brain drained a lot of really smart people out of Russia, a lot of them are here, over the last 30 years. And so, there’s this thing, it’s actually really funny if you think about it, the one thing that we know to be the height of absolute evil that the West ever did was colonization and resource extraction.

(02:53:03)
So, we know the height of absolute evil was when the Portuguese and the English and everybody else went and had these colonies, and then went in and we took all the oil, and we took all the diamonds, or we took all the whatever, lithium or whatever it is. Well, for some reason we realized that that’s a deeply evil thing to do when it’s a physical resource, when it’s a non-conscious physical matter, for some reason we think it’s completely morally acceptable to do it with human capital. In fact, we think it’s glorious and beautiful and wonderful and the great flowering of peace and harmony and moral justice of our time to do it, and we don’t think for one second what we’re doing to the countries that we’re pulling all these people out of.

(02:53:38)
And this is one of these things, I don’t know, maybe we’re just going to live in this delusional state forever, and we’ll just keep doing it, and it’ll keep benefiting us, and we just won’t care what happens, but I think there may come… This is one of these submarines 10 feet under the water line, I think it’s just a matter of time until people suddenly realize, oh my God, what are we doing? We need the rest of the world to succeed too, we need these other countries to flourish. We don’t want to be the only successful country in the middle of just complete chaos and disaster, and we just extract and we extract and we extract, and we don’t think twice about it.
Marc Andreessen
(02:54:11)
This is so deeply profound, actually. So, what is the cost of “winning” if these countries are drained in terms of human capital, on the level of geopolitics, what does that lead to? Even if we talk about wars and conflict and all of this, we actually want them to be strong, in the way we understand strong, not just in every way, so that cooperation and competition can build a better world for all of humanity. It’s interesting, this is one of those truths where you just speak and it resonates, and I didn’t even think about it.
Lex Fridman
(02:54:51)
Yeah, exactly.
Marc Andreessen
(02:54:53)
So, you were sitting during the holiday season, just boiling over. So, all that said, there’s still to use some good to the H-1B?
Lex Fridman
(02:55:03)
Okay, so then you get this other… Okay, so then there’s-
Marc Andreessen
(02:55:03)
Come all the way around.
Lex Fridman
(02:55:06)
… there’s another nuance. So there’s another nuance, there’s another nuance, which is mostly in the valley we don’t use H-1Bs anymore, mostly we use O1s. So, there’s a separate class of these, and the O1 is like this… It turns out the O1 is the super genius visa. So, the O1 is basically our founder… When we have somebody from anywhere in the world, and they’ve invented a breakthrough new technology, and they want to come to the US to start a company, they come in through an O1 Visa. And that actually, it’s a fairly high bar, it’s a high acceptance rate, but it’s a pretty high bar, and they do a lot of work, and you have to put real work into it, really prove your case. Mostly what’s happened with the H-1B Visa program is that it has gone to basically two categories of employers.

(02:55:47)
One is basically a small set of big tech companies that hire in volume, which is exactly the companies that you would think, and then the other is it goes to these, what they call the mills, the consulting mills. And so, there’s these set of companies with names, I don’t want to pick on companies, but names like Cognizant, that hire, basically have their business model is bringing in primarily Indians in large numbers, and they often have offices next to company-owned housing, and they’ll have organizations that are literally thousands of Indians living and working in the US, and they do basically call it mid-tier IT consulting. So, these folks, they’re making good wages, but they’re making 60 or 80 a or 100,000 a year, not the 300,000 that you’d make in the Valley.

(02:56:34)
And so, in practice, the startups, basically little tech as we call it, or the startup world, mainly doesn’t use H-1Bs at this point, and mainly can’t, because the system is kind of rigged in a way that we really can’t. And then, again, you get to the underlying morality here, which is, it’s like, well, Amazon, Amazon’s… I love Amazon. But they’re a big powerful company, they’ve got more money than God, they’ve got resources, they’ve got long-term planning horizon, they do big profound things over decades at a time, they could, or any of these other companies, could launch massively effective programs to go recruit the best and brightest from all throughout the country. And you’ll notice they don’t do that, they bring in 10,000, 20,000 H1Bs a year. And so, you’ve got a question there, and then these mills, there’s lots of questions around them, and whether that’s even an ethical way… I don’t want to say they’re unethical, but there’s questions around exactly what the trade-offs are there. And this like is a Pandora’s box that really nobody really wanted to be opened. To play devil’s advocate in all this, in terms of national immigration issues, none of this is a top end issue, just because the numbers are small, and so I don’t think, the administration has said this is not a priority of theirs for right now. But I guess what I would say is, there is actually a lot of complexity and nuance here. Like I said, I have a lot of friends and colleagues who came over on H-1Bs or O-1s, green cards, many are now citizens, and every single one of them was… Not every single one. A lot of them were enthusiastic to defend the honor of immigrants throughout this whole period. And they said to me, it’s like, well, Marc, how can we more clearly express the importance of high school immigration to the US?

(02:58:14)
And I was like, I think you can do it by advocating for also developing our native-born talent. Do you want to inflame the issue or do you want to diffuse the issue? I think the answer is to diffuse the issue. Let me give you one more positive scenario, and then I’ll also beat up on the university some more. Do you know about the National Merit Scholarship System, have you heard about this?
Marc Andreessen
(02:58:39)
Not really, can you explain?
Lex Fridman
(02:58:40)
So, there’s a system that was created during the Cold War called the National Merit Scholars, and it is a, basically, it was created, I forget, in the 50s or 60s when… It was people in government actually wanted to identify the best and the brightest, as heretical an idea as that sounds today. And so, it’s basically a national talent search for, basically, IQ. Its goal is to identify basically the top 0.5% of the IQ in the country. By the way, completely regardless of other characteristics. So, there’s no race, gender, or any other aspect to it, it’s just going for straight intelligence. It uses, first, the PSAT, which is the preparatory SAT that you take, and then the SAT. So, it uses those scores, that is the scoring, it’s a straight PSAT/SAT scoring system. They use the SAT as a proxy for IQ, which it is. They run this every year, they identify, they get down to 1% of the population of the kids, of 18 year olds in an given year, who score highest on the PSAT, and then they further qualify down to the 0.5% that also replicate on the SAT. And then it’s like, the scholarship amount is like $2,500. So, it was a lot of money 50 years ago, not as much today. But it’s a national system being run, literally, to find the best and the brightest. How many of our great and powerful universities use this as a scouting system? Our universities all have sports teams, they all have national scouting, they have full-time scouts who go out and they go to every high school and they try to find all the great basketball players and bring them into the NCAA, into all these leagues. How many of our great and powerful and enlightened universities use the National Merit System to go do a talent search for the smartest kids and just bring them in?
Marc Andreessen
(03:00:21)
Let me guess, very few. Zero.
Lex Fridman
(03:00:23)
Zero. As you say it, that’s brilliant, there should be that same level of scouting for talent internally.
Marc Andreessen
(03:00:30)
Go get the smartest ones. I’ll give you one more kicker on this topic, if I haven’t beaten it to death. The SAT has changed. So, the SAT used to be a highly accurate proxy for IQ that caused a bunch of problems, people really don’t like the whole idea of IQ. And so, the SAT has been actively managed over the last 50 years by the college board that runs it, and it has been, essentially, like everything else, it’s been dumbed down, in two ways. Number one, it’s been dumbed down where an 800 from 40 years ago does not mean what an 800 means today. And 40 years ago, it was almost impossible to get an 800. Today, there’s so many 800s that you could stock the entire Ivy League with 800s, and so it’s been deliberately dumbed down. And then, two is, they have tried to pull out a lot of what’s called the g-loading.

(03:01:21)
And so they, they’ve tried to detach it from being an IQ proxy because IQ is such an inflammatory concept. And the consequence of that is, and this is sort of perverse, they’ve made it more coachable, right? So, the SAT 40 years ago, coaching didn’t really work, and more recently it has really started to work. And one of the things you see is that the Asian spike, you see this giant leap upward in Asian performance over the last decade, and I think, looking at the data, I think a lot of that is because it’s more coachable now, and the Asians do the most coaching. So, there’s a bunch of issues with this. And so, the coaching thing is really difficult because the coaching thing is a subsidy then to the kids whose parents can afford coaching, and I don’t know about you, but where I grew up, there was no SAT coaching. So, there’s an issue there. I didn’t even know what the SAT was until the day I took it, much less that there was coaching, much less that it could work, so much less we could afford it.

(03:02:08)
So, number one, there’s issues there. But the other issue there is think about what’s happened by the dumbing down, 800 no longer captures all the smart, 800 is too crude of a test. It’s like the AI benchmarking problem. It’s the same problem they have AI benchmarking right now, 800 is too low of a threshold. There are too many kids scoring 800. Because what you want is you want, whatever, if it’s going to be 100,000 kids, I don’t know what it is, if it’s going to be 50,000 kids a year scoring 800, you also then want kids to be able to score 900 and 1000, and 1100, and 1200, and you want to ultimately get to, you’d like to ultimately identify the top 100 kids, and make sure that you get them in MIT. And the resolution of the test has been reduced so that it actually is not useful for doing that.

(03:02:49)
And again, I would say this is part of the generalized corruption that’s taken place throughout this entire system, where we have been heading in the reverse direction from wanting to actually go get the best and brightest and actually put them in the places where they should be. And then, just the final comment would be, the great thing about standardized testing and the National Merit System is it’s, like I said, it’s completely race blind, it’s gender blind, it’s blind on every other characteristic, it’s only done on test scores. And you can make an argument about whether that’s good or bad, but it is, for sure, it’s the closest thing that we had to get to merit. It was the thing that they did when they thought they needed merit to win the Cold War.

(03:03:23)
And of course, we could choose to do that anytime we want. And I just say, I find it incredibly striking, and an enormous moral indictment of the current system that there are no universities that do this today. So, back to the immigration thing, just real quick, it’s like, okay, we aren’t even trying to go get the smart kids out of [inaudible 03:03:39], and even if they think that they can get into these places, they get turned down. And the same thing for the smart Asians, and the same thing for the smart Jews, and the same thing for the smart Black people. And I don’t know how that’s moral, I don’t get it at all.
Lex Fridman
(03:03:54)
As you said about the 800, so I took the SAT and the ACT many times, and I’ve always gotten perfect on math, 800. And I’m not special, it doesn’t identify genius, I think you want to search for genius. And you want to create measures that find genius of all different kinds, speaking of diversity. And I guess we should reiterate and say over and over and over, defend immigrants, yes, but say we should hire more and more native-born.
Marc Andreessen
(03:04:32)
Well, you asked me in the beginning what’s the most optimistic forecast that we could have? And the most optimistic forecast would be, my God, what if we did both?
Lex Fridman
(03:04:44)
So, that’s the reasonable, the rational, the smart thing to say here. In fact, we don’t have to have a war.
Marc Andreessen
(03:04:50)
Well, it would defuse, it would defuse the entire issue.
Lex Fridman
(03:04:52)
Yeah.
Marc Andreessen
(03:04:53)
If everybody in the center and the South of the country, and every Jewish family, Asian family, Black family knew they were getting a fair shake, it would defuse the issue. How about defusing the issue? What a crazy radical… Sorry, I don’t mean to really get out over my skis here, but…

Little tech

Lex Fridman
(03:05:06)
I think your profile on X states, it’s time to build. It feels like 2025 is a good year to build. So, I wanted to ask your advice, and maybe for advice for anybody who’s trying to build, who is trying to build something useful in the world. Maybe launch a startup, or maybe just launch apps, services, whatever, ship software products. So, maybe, by way of advice, how do you actually get to shipping?
Marc Andreessen
(03:05:44)
So, a big part of the answer I think is we’re in the middle of a legit revolution, and I know you’ve been talking about this on your show. But AI coding, this is the biggest earthquake to hit software in certainly my life, maybe since the invention of software. And we’re involved in various of these companies, but these tools, from a variety of companies, are absolutely revolutionary, and they’re getting better by leaps and bounds every day. And you know all this. But the thing with coding, there’s open questions of whether AI can get better at, I don’t know, understanding philosophy, or whatever, creative writing or whatever, but for sure we can make it much better at coding, because you can validate the results of coding. And so, there’s all these methods of synthetic data and self-training and reinforcement learning that, for sure, you can do with coding.

(03:06:30)
And so, everybody I know who works in the field says AI coding is going to get to be phenomenally good. And it’s already great. And anybody who wants to see this, just go on YouTube and look at AI coding demos, little kids making apps in 10 minutes, working with an AI coding system. And so, I think it’s the golden age… I think this is an area where it’s clearly the golden age. The tool set is extraordinary. In a day as a coder, for sure, in a day you can retrain yourself, start using these things, get a huge boost in productivity, as a non-coder, you can learn much more quickly than you could before.
Lex Fridman
(03:07:00)
That’s actually a tricky one in terms of learning as a non-coder to build stuff, I feel like you still need to learn how to code. It becomes a superpower, it helps you be much more productive. You could legitimately be a one person company and get quite far.
Marc Andreessen
(03:07:19)
I agree with that, up to a point. So, I think, for sure, for quite a long time, the people who are good at coding are going to be the best at actually having AIs code things, because they’re going to understand what, very basic, they’re going to understand what’s happening. And they’re going to be able to evaluate the work, and they’re going to be able to literally manage AIs better, even if they’re not literally handwriting the code, they’re just going to have a much better sense of what’s going on. So, I definitely think, 100% my nine-year-old is doing all kinds of coding classes, and he’ll keep doing that for, certainly through 18, we’ll see after that. And so, for sure that’s the case. But look, having said that, one of the things you can do with an AI is say, teach me how to code.

(03:07:58)
And there’s a whole bunch of, I’ll name names, Khan Academy… There’s a whole bunch of work that they’re doing at Khan Academy for free, and then we have this company, Replit, which was originally specifically built for kids for coding, that has AI built in, that’s just absolutely extraordinary now. And then, there’s a variety of other systems like this. Yeah, the AI is going to be able to teach to code… AI, by the way, is, as you know, is spectacularly good at explaining code. And so, the tools have these features now where you can talk to the code base, and so you can literally ask the code base questions about itself. And you can also just do the simple form, which is you can copy and paste code into a ChatGPT and just ask it to explain it, what’s going on, rewrite it, improve it, make recommendations. And so, there’s dozens of ways to do this.

(03:08:46)
By the way, you can also, even more broadly than code, like, okay, you want to make a video game, okay, now you can do AI, art generation, sound generation, dialogue generation, voice generation, all of a sudden you don’t need designers, you don’t need voice actors. Yeah, there’s just unlimited…And then a big part of coding is so-called glue, it’s interfacing into other systems. So, it’s interfacing into Stripe, to take payments, or something like that, and AI is fantastic at writing glue code. So, really, really good at making sure that you can plug everything together, really good at helping you figure out how to deploy. It’ll even write a business plan for you. So, it’s just this, it’s like everything happening with AI right now, it’s like this latent superpower, and there’s this incredible spectrum of people who have really figured out massive performance increases, productivity increases with it already, there’s other people who aren’t even aware it’s happening.

(03:09:39)
And there’s some gearing to whether you’re a coder or not, but I think there are lots of non-coders that are off to the races, and I think there are lots of professional coders who are still like, eh… The blacksmiths were not necessarily in favor of the car business. So, there’s the old William Gibson quote, “The future is here, it’s just not evenly distributed yet,” and this is maybe the most potent version of that that I’ve ever seen.
Lex Fridman
(03:10:04)
Yeah, there’s the old meme with the bell curve, the people on both extremes say, “AI coding is the future.” It’s very common, the programmers to say, if you’re any good of a programmer, you’re not going to be using it, that’s just not true. I consider myself a reasonably good programmer and my productivity has been just skyrocketed, and the joy of programming skyrocketed, every aspect of programming is more efficient, more productive, more fun, all of that kind of stuff.
Marc Andreessen
(03:10:38)
I would also say code has, of anything in industrial society, code has the highest elasticity, which is to say the easier it is to make it, the more of it gets made. I think effectively there’s unlimited demand for code. In other words, there’s always some other idea for a thing that you can do, a feature that you can add, or a thing that you can optimize. And so, overwhelmingly, the amount of code that exists in the world is a fraction of even the ideas we have today, and then we come up with new ideas all the time. And so, I think that… I was, in the late 80s, early 90s, when automated coding systems started to come out, expert systems, a big deal in those days, and there was a famous book called The Decline and Fall of the American Programmer, that predicted that these new coding systems were going to mean we wouldn’t have programmers in the future, and of course, the number of programming jobs exploded by a factor of 100.

(03:11:27)
My guess is we’ll have more coding jobs probably by an order of magnitude 10 years from now. That will be different, they’ll be different jobs, they’ll involve orchestrating AI, but we will be creating so much more software that the whole industry will just explode in size.
Lex Fridman
(03:11:44)
Are you seeing the size of companies decrease in terms of startups? What’s the landscapes of little tech?
Marc Andreessen
(03:11:51)
All we’re seeing right now is the AI hiring boom of all time.
Lex Fridman
(03:11:55)
Oh, for the big tech?
Marc Andreessen
(03:11:57)
And little tech.
Lex Fridman
(03:11:57)
And little tech.
Marc Andreessen
(03:11:58)
Everybody’s trying to hire as many engineers as they can to build AI systems, it’s 100%… There’s a handful of company… There’s a little bit, in customer service, we have some companies and others, I think it’s Klarna that’s publicizing a lot of this, in Europe, where… There are jobs that can be optimized, and jobs that can be automated. But for engineering jobs, it’s just an explosion of hiring, that at least, so far, there’s no trace of any sort of diminishing effect. Now, having said that, I am looking forward to the day, I am waiting for the first company to walk in saying, yes… The more radical form of it. So, basically, the companies that we see are basically one of two kinds, we see the companies that are basically… Sometimes we use weak form, strong form. So, the weak form companies, I sometimes use the term, it’s call it the sixth bullet point, AI is the sixth bullet point on whatever they’re doing.
Lex Fridman
(03:12:51)
Sure.
Marc Andreessen
(03:12:52)
Right? And it’s on the slide. So, they’ve got the whatever, da, da, da, da, da… And then AI is the sixth thing. And the reason AI is the sixth thing is because they had already previously written the slide before the AI revolution started, and so they just added the six bullet point in the slide. Which is how you’re getting all these products that have the AI button up in the corner, the little sparkly button. And all of a sudden Gmail is offering to summarize your email, which I’m like, I don’t need that, I need you to answer my email, not summarize it. What the hell? Okay, so we see those, and that’s fine, that’s like, I don’t know, putting sugar on the cake or something. But then, we see the strong form, which is the companies that are building from scratch for AI, and they’re building it… I actually just met with a company that is building literally an AI email system, as an example, so just-
Lex Fridman
(03:13:32)
Oh, nice, I can’t wait.
Marc Andreessen
(03:13:34)
Yeah, they’re going to completely… So, very obvious idea, very smart team, it’s going to be great. And then, Notion, just another, not one of our companies, but just came out with a product. So now companies are going to basically come through, sweep through, and they’re going to do basically AI first versions of basically everything. And those are, companies built… AI is the first bullet point, it’s the strong form of the argument.
Lex Fridman
(03:13:55)
Cursor is an example of that, they basically said, okay, we’re going to rebuild the thing with AI as the first citizen.
Marc Andreessen
(03:14:01)
What [inaudible 03:14:02] from scratch that we could build on this? And again, this is part of the Full Employment Act for startups and VCs is, if a technology transformation is sufficiently powerful, then you actually need to start the product development process over from scratch because you need to reconceptualize the product, and then usually what that means is you need a new company because most incumbents just won’t do that. So, yeah, that’s underway across many categories. What I’m waiting for is the company where it’s like, no, our org chart is redesigned as a result of AI. So, I’m waiting for the company where it’s like, no, we’re going to have… And the cliché, here’s a thought experiment, the cliché would be we’re going to have the human executive team, and then we’re going to have the AIs be the workers. So, we’ll have a VP of engineering supervising 100 instances of coding agents. Okay, maybe… By the way, or maybe the VP of engineering should be the AI, maybe supervising human coders who are supervising AIs.

(03:14:57)
Because one of the things that AI should be pretty good at is managing because it’s process-driven, it’s the kind of thing that AI is actually pretty good at, right? Performance evaluation, coaching. And so, should it be an AI executive team? And then, of course, the ultimate question, which is AI CEO. And then, maybe the most futuristic version of it would be an actual AI agent that actually goes fully autonomous. Yeah, what if you really set one of these things loose and let it basically build itself a business? And so, I will say, we’re not yet seeing those, and I think there’s a little bit of the systems aren’t quite ready for that yet, and then I think it’s a little bit of, you really do need, at that point, a founder who’s really willing to break all the rules, and really willing to take the swing, and those people exist, and so I’m sure we’ll see that.
Lex Fridman
(03:15:46)
And some of it is, as you know with all the startups, this is the execution. The idea that you have a AI first email client seems like an obvious idea, but actually creating one, executing it, and then taking on Gmail is really difficult. Gmail, it’s fascinating to see Google can’t do it, because why? Because momentum, because it’s hard to re-engineer the entirety of the system, because feels like Google is perfectly positioned to do it. Same with you have Perplexity, which I love, Google could technically take on Perplexity and do it much better, but they haven’t, not yet. So, it’s fascinating why that is for large companies, that is an advantage for little tech, they could be agile.
Marc Andreessen
(03:16:33)
Yeah, that’s right.
Lex Fridman
(03:16:34)
They can move fast.
Marc Andreessen
(03:16:34)
Yeah. Little companies can break glass in a way big companies can’t.
Lex Fridman
(03:16:37)
Right.
Marc Andreessen
(03:16:38)
This is sort of the big breakthrough that Clay Christensen had in the Innovator’s Dilemma, which is sometimes when big companies don’t do things, it’s because they’re screwing up. And that certainly happens. But a lot of times they don’t do things because it would break too much glass. Specifically, it would interfere with their existing customers and their existing businesses, and they just simply won’t do that. And by the way, responsibly, they shouldn’t do that. And so, they just get, this is Clay Christensen’s big thing, is they often don’t adapt because they’re well-run, not because they’re poorly run. But they’re optimizing machines, they’re optimizing against the existing business. And as you just said, this is a permanent state of affairs for large organizations. Every once in a while, one breaks the pattern and actually does it, but for the most part, this is a very predictable form of human behavior, and this fundamentally is why startups exist.

AI race

Lex Fridman
(03:17:26)
It feels like 2025 is when the race for dominance in AI will see some winners. It’s a big year. So, who do you think wins the race? OpenAI, Meta, Google, xAI… Who do you think wins the AI race?
Marc Andreessen
(03:17:39)
I would say, I’m not going to predict, I’m going to say there’s questions all over the place. And we have this category of question we call the trillion-dollar question, which is literally, depending on how it’s answered, people make or lose a trillion dollars, and I think there’s, I don’t know, five or six trillion questions right now, that are hanging out there, which is an unusually large number. And I’ll just hit a few of them and we can talk about them. So, one is big models versus small models, another is open models versus closed models, another…
Marc Andreessen
(03:18:00)
… Small models. Another is open models versus closed models. Another is whether you can use synthetic data or not. Another is chain of thought. How far can you push that? And reinforcement learning. And then another one is political trillion dollar questions, policy questions, which the US and the EU have both been flunking dramatically and the US hopefully is about to really succeed at. Yeah. And then there’s probably another half dozen big important questions after that. And so these are all just like, say, this is an industry that’s in flux in a way that I even more dramatic, I think, than the ones I’ve seen before.

(03:18:35)
And look, the most obvious example of the flux is sitting here less than three years ago, sitting here in December of ’22, we would’ve said that Open Ai is just running away with everything. And sitting here today, it’s like there’s at least six world-class God model companies and teams that are, by the way, generating remarkably similar results. That’s actually been one of the most shocking things to me is it turns out that once you know that it’s possible to build one incredibly smart Turing Test-passing large language model, which was a complete shock and surprise to the world, it turns out within a year you can have five more. There’s also a money component thing to it, which is to get the money to scale one of these things into the billions of dollars. There’s basically right now only two sources of money that will do that for you. One is the hyperscalers giving you the money which you turn around and round trip back to them, or foreign sovereigns, other country sovereign wealth funds, which can be difficult in some cases, for companies to access.

(03:19:33)
So there’s maybe another trillion-dollar question is the financing question. Here’s one. So Sam Altman has been public about the fact that he wants to transition OpenAI from being a non-profit to being a for-profit. The way that that is legally done is that … And there is a way to do it, there is a way in US law to do it. The IRS and other legal entities, government entities, scrutinizes this very carefully because the US takes foundation non-profit law very seriously because of the tax exemption.

(03:19:59)
And so historically, the way that you do it is you start a for-profit and then you raise money with the for-profit to buy the assets of the non-profit at fair market value. And the last financing round at OpenAI was 150 some billion dollars. And so logically, if the flip is going to happen, the for-profit has to go raise 150 billion out of the chute to buy the assets. Raising 150 billion is a challenge. So is that even possible? If that is possible, then OpenAI maybe is off to the races as a for-profit company. If not, I don’t know. And then obviously the Elon lawsuit. So just because they’re the market leader today, there’s big important questions there. Microsoft has this kind of love-hate relationship with them. Where does that go? Apple’s lagging badly behind, but they’re very good at catching up. Amazon is primarily hyperscalar, but they now have their own models.
Lex Fridman
(03:20:52)
And then there’s the other questions like you laid out brilliantly, briefly and brilliantly, open versus closed, big versus little models, synthetic data. That’s a huge, huge question. And then test on compute with a chain of thought. They’re all of that. And it’s just fascinating. And these are, I think it’s fair to say, trillion-dollar questions.
Marc Andreessen
(03:21:11)
Yeah, these are big. Look here’s a trillion-dollar question, which is kind of embedded in that, which is just hallucinations. So if you are trying to use these tools creatively, you’re thrilled because they can draw new images and they can make new music and they can do all this incredible stuff. They’re creative. The flip side of that is if you need them to be correct, they can’t be creative. And that’s the term hallucination. And these things do hallucinate. And there have been court cases already where lawyers have submitted legal briefs that contain made-up court citations, case citations. The judge is like, “Wait a minute, this doesn’t exist.” And the very next question is, “Did you write this yourself?” And the lawyer goes, “Er…”
Lex Fridman
(03:21:49)
I mean, that’s why with Elon, with Grok, looking for truth. I mean, that’s an open technical question. How close can you get to truth with LLMs?
Marc Andreessen
(03:21:58)
Yeah, that’s right. And my sense, this is a very contentious topic at the industry, my sense is to the extent that there is a domain in which there is a definitive and checkable and provable answer, and you might say, math satisfies that, coding satisfies that, and maybe some other fields, then you should be able to generate synthetic data. You should be able to do chain of thought reasoning. You should be able to do reinforcement learning and you should be able to ultimately eliminate hallucinations. But by the way, that’s a trillion-dollar question right there as to whether that’s true. But then there’s questions like, okay, is that going to work in the more general domain? So for example, one possibility is these things are going to get truly superhuman at math and coding.

(03:22:36)
But at discussing philosophy, they’re basically as smart as they’re ever going to be. And they’re going to be kind of say mid-wit grad student level. And the theory there would just be they’re already out of training data. They literally, you talk to these people, literally the big models, the big models are within a factor of 2X of consuming all the human-generated training data, to the point that some of these big companies are literally hiring people like doctors and lawyers to sit and write new training data by hand. And so does this mean that you have to, if you want your model to get better philosophy, you have to go hire a thousand philosophers and have them write new content, and is anybody going to do that? And so maybe these things are topping out in certain ways and they’re going to leap way ahead in other ways.

(03:23:16)
Anyway, so we just don’t … Actually, maybe my main conclusion is anybody telling you these big sweeping conclusions, this whole, all of these abstract generalized super intelligence AGI stuff, maybe it’s the engineer in me, but no, that’s too abstract. It’s got to actually work. And then by the way, it has to actually have to be able to pay for it. I mean, this is a problem right now with the big models, the big models that are really good at coding and math. They’re actually very expensive to run. They’re quite slow.

(03:23:51)
Another trillion-dollar question, future chips, which I know you’ve talked a lot about. Another trillion-dollar question, yeah, I mean all the global issues. Oh, another trillion-dollar question, censorship. And, as they say, all the human feedback training process. Exactly what are you training these things to do? What are they allowed to talk about? How long do they give you these … How often do they give these incredibly preaching moral lectures?

(03:24:21)
Here’s a trillion-dollar question. How many other countries want their country to run its education system, healthcare system, new system, political system, on the basis of an AI that’s been trained according to the most extreme left-wing California politics? Because what they have on offer right now. And I think the answer to that is not very many. So there’s massive open questions there about, and by the way, what morality of these things are going to get trained on as a …
Lex Fridman
(03:24:48)
And now [inaudible 03:24:50], we’re cracking wide open with what’s been happening over the past few months. Censorship on every level of these companies, and just the very idea what truth means and what it means to expand the Overton window of LLMs or the Overton window of human discourse.
Marc Andreessen
(03:25:08)
So what I experienced, going back to how we started, what I experienced was, all right, social media censorship regime from hell, debanking at large scale, and then the war on the crypto industry, trying to kill it. And then basically declared intent to do the same thing to AI and to put AI under the same kind of censorship and control regime as social media and the banks. And I think this election tipped, in America, I think this election tipped us from a timeline in which things were going to get really bad on that front to a timeline in which I think things are going to be quite good.

(03:25:40)
But look, those same questions also apply outside the US and the EU is doing their thing. They’re being extremely draconian and they’re trying to lock in a political censorship regime on AI right now that’s so harsh that even American AI companies are not even willing to launch new products in the EU right now. That’s not going to last. But what happens there and what are the trade-offs? What levels of censorship are American companies going to have to sign up for if they want to operate in the EU? Or is the EU still capable of generating its own AI companies or have we brain drained them so that they can’t? So big questions.

X

Lex Fridman
(03:26:15)
Quick question. So you’re very active on X. A very unique character: flamboyant, exciting, bold. You post a lot. I think there’s a meme, I don’t remember it exactly, but that Elon posted something like inside Elon, there are two wolves. One is please be kind or more positive. And the other one is, I think doing the, I take a big step back and fuck yourself in the face guy. How many wolves are inside your mind when you’re tweeting?
Marc Andreessen
(03:26:51)
To be clear, a reference from the comedy classic Tropic Thunder.
Lex Fridman
(03:26:54)
Tropic Thunder, yeah. Legendary movie.
Marc Andreessen
(03:26:56)
Yes. Any Zoomers listening to this who haven’t seen that movie, go watch it immediately.
Lex Fridman
(03:27:02)
Yeah, there’s nothing offensive about it.
Marc Andreessen
(03:27:04)
Nothing offensive about it at all. So Tom Cruise’s greatest performance. So yeah, no, look, I should start by saying I’m not supposed to be tweeting at all.
Lex Fridman
(03:27:19)
Yeah.
Marc Andreessen
(03:27:19)
Yes, yes, yes. But you know.
Lex Fridman
(03:27:22)
So how do you approach that? How do you approach what to tweet?
Marc Andreessen
(03:27:25)
I mean, I don’t. I don’t well enough. It’s mostly an exercise in frustration. Look, there’s a glory to it and there’s an issue with it, and the glory of it is instantaneous global communication. X in particular is the town square on all these social issues, political issues, everything else, current events. But I mean, look, there’s no question of the format. The format of at least the original tweet is prone to be inflammatory. I’m the guy who at one point, the entire nation of India hated me because I once tweeted something. It turned out that it’s still politically sensitive in the entire continent. I stayed up all night that night as I became front page headline and leading television news in each time zone in India for a single tweet. So the single tweet out of context is a very dangerous thing. Obviously X now has the middle ground where they now have the longer form essays. And so probably the most productive thing I can do is longer form things.
Lex Fridman
(03:28:26)
You’re not going to do it though are you?
Marc Andreessen
(03:28:28)
I do. I do. From time-to-time. I do.
Lex Fridman
(03:28:28)
Sometimes.
Marc Andreessen
(03:28:29)
I should do more of them. Yeah. Look, obviously X is doing great. And then like I said, Substack know has become the center for a lot of them. I think the best deeply thought through, certainly intellectual content, tons of current events stuff there as well. And then, yeah, then there’s a bunch of new systems that are very exciting. So I think one of the things we can look forward to in the next four years is number one, just a massive reinvigoration of social media as a consequence of the changes that are happening right now. I’m very excited to see what’s going to happen with that. And then it’s happened on X, but it’s now going to happen on other platforms.

(03:29:05)
And then the other is crypto’s going to come right back to life. And actually that’s very exciting. Actually, that’s worth noting is that’s another trillion-dollar question on AI, which is in a world of pervasive AI, and especially in a world of AI agents, and imagine a world of billions or trillions of AI agents running around, they need an economy. And crypto, in our view, happens to be the ideal economic system for that, because it’s a programmable money. It’s a very easy way to plug in and do that. And there’s this transaction processing system that can do that. And so I think the crypto AI intersection is potentially a very, very big deal. And so that was going to be impossible under the prior regime, and I think under the new regime, hopefully, it’ll be something we can do.

Yann LeCun

Lex Fridman
(03:29:48)
Almost for fun. Let me ask a friend of yours, Yann LeCun, what are your top 10 favorite things about Yann LeCun? I think he’s a brilliant guy. I think he’s important to the world. I think you guys disagree on a lot of things, but I personally like vigorous disagreement, I, as a person in the stands, like to watch the gladiators go at it. And-
Marc Andreessen
(03:30:12)
No, he’s a super genius. I mean, look, I wouldn’t say we’re super close, but casual friends. I worked with him at Meta. He was the chief scientist at Meta for a long time and still works with us. And obviously is a legendary figure in the field and one of the main people responsible for what’s happening. My serious observation would be it’s the thing I’ve talked to him about for a long time, and I keep trying to read and follow everything he does, is he’s probably, he is the, I think, see if you agree with this, he is the smartest and most credible critic of LLMs as the path for AI. And he’s not, there’s certain, I would say, troll-like characters who are just cropping everything but Yann has very deeply thought through, basically, theories as to why LLMs are an evolutionary dead end.

(03:30:58)
And I actually, I try to do this thing where I try to model, try to have a mental model of the two different sides of a serious argument. And so I’ve tried to internalize that argument as much as I can. Which is difficult because we’re investing it behind LLMs as aggressively as we can. And so if he’s right, that could be a big problem, but we should also know that. And then I sort of use his ideas to challenge all the bullish people to really test their level of knowledge. So I like to grill people.

(03:31:28)
I got my CS degree 35 years ago, so I’m not deep in the technology, but to the extent I can understand Yann’s points, I can use them to really surface a lot of the questions for the people who are more bullish. And that’s been, I think, very, very productive. So it is very striking that you have somebody who is that central in the space, who is actually a full-on skeptic. And again, this could go different ways. He could end up being very wrong. He could end up being totally right, or it could be that he will provoke the evolution of these systems to be much better than they would’ve been.
Lex Fridman
(03:32:02)
He could be both right and wrong. First of all, I do agree with that. He’s one of the most legit and rigorous and deep critics of the LLM path to AGI know. His basic notion is that there AI needs to have some physical understanding of the physical world, and that’s very difficult to achieve with LLMs. And that is a really good way to challenge the limitations of LLMs and so on. He’s also been a vocal and a huge proponent of open source, which is a whole nother, which you have been as well.
Marc Andreessen
(03:32:35)
Which is very useful.
Lex Fridman
(03:32:36)
And that’s been just fascinating to watch.
Marc Andreessen
(03:32:40)
And anti-doomer.
Lex Fridman
(03:32:40)
Anti-doomer?
Marc Andreessen
(03:32:42)
He’s very anti-doomer.
Lex Fridman
(03:32:43)
He embodies … he also has many wolves inside.
Marc Andreessen
(03:32:47)
Yes, he does. Yes, does. Yes he does. So it’s been really, really fun to watch.
Marc Andreessen
(03:32:50)
The other two. Okay, here’s my other wolf coming out. The other two of the three godfathers of AI are radicals. Full-on far left, I would say either Marxists or borderline Marxists. And they’re, I think, quite extreme in their social political views. And I think that feeds into their doomerism, and I think they are lobbying for draconian government. I think what would be ruinously destructive government legislation and regulation. And so it’s actually super helpful, super, super helpful to have you on as a counterpoint to those two.

Andrew Huberman

Lex Fridman
(03:33:22)
Another fun question, our mutual friend Andrew Huberman. First maybe, what do you love most about Andrew? And second, what score on a scale of one to 10 do you think he would give you on your approach to health?
Marc Andreessen
(03:33:34)
Oh, three.
Lex Fridman
(03:33:36)
Physical three. You think you’d score that high, huh? Okay.
Marc Andreessen
(03:33:39)
Exactly.
Lex Fridman
(03:33:40)
That’s good.
Marc Andreessen
(03:33:41)
Exactly. Well, so he convinced me to stop drinking alcohol, which was a big-
Lex Fridman
(03:33:46)
Successfully?
Marc Andreessen
(03:33:47)
Well, other than my family, it was my favorite thing in the world. And so it was a major, major reduction. Having a glass of scotch at night, it was the thing I would do to relax. And so he has profoundly negatively impacted my emotional health. I blame him for making me much less happy as a person, but much, much, much healthier, physically healthier. So that I credit him with that. I’m glad I did that. But then his sleep stuff like, yeah, I’m not doing any of that.
Lex Fridman
(03:34:14)
Yeah.
Marc Andreessen
(03:34:14)
I have no interest in his sleep shit. No. This whole light, natural light, no, we’re not doing.
Lex Fridman
(03:34:20)
You’re too hardcore for this?
Marc Andreessen
(03:34:21)
I don’t see any natural … I don’t see any natural light in here.
Lex Fridman
(03:34:24)
It’s all covered. It’s all horrible.
Marc Andreessen
(03:34:27)
And I’m very happy. I would be very happy living and working here because I’m totally happy without natural light.
Lex Fridman
(03:34:32)
In darkness.
Marc Andreessen
(03:34:34)
Yes.
Lex Fridman
(03:34:34)
It must be a metaphor for something.
Marc Andreessen
(03:34:35)
Yes, it’s a test. Look, it’s a test of manhood as to whether you can have a blue screen in your face for three hours and then go right to sleep. I don’t understand why you shouldn’t want to take shortcuts.

Success

Lex Fridman
(03:34:45)
I now understand what they mean by toxic masculinity. All right. So let’s see. You’re exceptionally successful by most measures, but what to you is the definition of success?
Marc Andreessen
(03:35:02)
I would probably say it is a combination of two things. I think it is contribution. So have you done something that mattered ultimately and specifically mattered to people? And then the other thing is, I think, happiness is either overrated or almost a complete myth. And in fact, Interesting, Thomas Jefferson did not mean happiness the way that we understand it. When he said “Pursuit of happiness” in the Declaration of Independence, he meant it more of the Greek meaning, which is closer to satisfaction or fulfillment. So I think about happiness as the first ice cream cone makes you super happy. The first mile of the walk in the park during sunset makes super happy. The first kiss makes you super happy. The thousandth ice cream cone, not so much. The thousandth mile of the walk through the park. The thousandth kiss can still be good, but maybe just not right in a row. And so happiness is this very fleeting concept, and the people who anchor on happiness seem to go off the rails pretty often. So the deep sense of having been, I don’t know how to put it, useful.
Lex Fridman
(03:36:20)
So that’s a good place to arrive at in life.
Marc Andreessen
(03:36:23)
Yeah, I think so. Yeah. I mean, who was it who said, the source of all the ills in the world is man’s inability to sit in a room by himself doing nothing. But if you’re sitting in a room by yourself and you’re like, all right, four in the morning, it’s like, all right, have I lived up to my expectation of myself? If you have, the people I know who feel that way are pretty centered and generally seem very, I don’t know how to put it, pleased, proud, calm, at peace. The people who are sensation seekers … Some of the sensations, … By the way, there’s certain entrepreneurs, for example, who are into every form of extreme sport and they get huge satisfaction out to that, or they’re sensation seeking in useful and productive ways. Larry Ellison was always like that. Zuckerberg is like that. And then there’s a lot of entrepreneurs who end up, drugs, like sexual escapades that seem like they’ll be fun at first and then backfire.
Lex Fridman
(03:37:26)
Yeah. But at the end of the day, if you’re able to be at peace by yourself in a room at 4:00 AM and I would even say happy, but I know, I understand Thomas Jefferson didn’t mean it the way, maybe I mean it, but I can be happy by myself at 4:00 AM with a blue screen.
Marc Andreessen
(03:37:43)
That’s good. Exactly.
Lex Fridman
(03:37:44)
Staring at a cursor.
Marc Andreessen
(03:37:46)
Exactly.

God and humanity

Lex Fridman
(03:37:49)
As a small tangent, a quick shout out to an amazing interview you did with Bari Weiss and just to her in general, Bari Weiss of the Free Press. She has a podcast called, Honestly, with Bari Weiss. She’s great. People should go listen. You were asked if you believe in God. One of the joys … See, we talked about happiness. One of the things that makes me happy is making you uncomfortable.
Marc Andreessen
(03:38:13)
Thank you.
Lex Fridman
(03:38:13)
So this question is designed for … Many of the questions today were designed for that. You were asked if you believe in God, and you said after a pause, that you’re not sure. So it felt like the pause, the uncertainty there was some kind of ongoing search for wisdom and meaning. Are you, in fact, searching for wisdom and meaning?
Marc Andreessen
(03:38:37)
I guess I’d put it this way. There’s a lot to just understand about people that I feel like I’m only starting to understand. And that’s certainly a simpler concept than God. So that’s what I’ve spent a lot of the last 15 years trying to figure out. I feel like I spent my first whatever, 30 years figuring out machines, and then now I’m spending 30 years figuring out people, which turns out to be quite a bit more complicated. And then, I don’t know, maybe God’s the last 30 years or something. And then look, I mean just like Elon, it’s just like, okay, the known universe is very complicated and mystifying. I mean, every time I pull up an astronomy, my kid super in astronomy, and it’s like, Daddy, how many galaxies are there in the universe? And how many galaxies are there in the universe?
Lex Fridman
(03:39:26)
A hundred billion?
Marc Andreessen
(03:39:27)
Okay, how?
Lex Fridman
(03:39:28)
Yeah, yeah.
Marc Andreessen
(03:39:33)
How is that freaking possible? It’s such a staggering concept that I-
Lex Fridman
(03:39:39)
I actually wanted to show you a tweet that blew my mind from Elon from a while back. Elon said, “As a friend called it, this is the ultimate skill tree. This is a wall of galaxies a billion light years across.” So these are all galaxies.
Marc Andreessen
(03:39:55)
Yeah. How is it that big? How the hell? I’m like, I can read the textbook and the this and the that and the whatever, 8 billion years and the Big Bang and the whole thing. And then it’s just like, all right, wow. And then it’s like, all right, the Big Bang. All right, what was before the Big Bang?
Lex Fridman
(03:40:13)
Do you think we humans will ever colonize like a galaxy and maybe even go beyond?
Marc Andreessen
(03:40:19)
Sure. I mean, yeah, in the fullness of time. Yeah.
Lex Fridman
(03:40:22)
So you have that kind of optimism. You have that kind of hope that extends across thousand of [inaudible 03:40:26]?
Marc Andreessen
(03:40:26)
In the fullness of time. I mean, all the challenges with it that I do, but yeah, why not? I mean, again, in the fullness of time, it’ll take a long time.
Lex Fridman
(03:40:33)
You don’t think we’ll destroy ourselves?
Marc Andreessen
(03:40:34)
No, I doubt it. I doubt it. And fortunately we have Elon giving us the backup plan. So I don’t know. I grew up real Midwest, just conventionally Protestant Christian. It never made that much sense to me. Got trained as an engineer and a scientist. I’m like, “Oh, that definitely doesn’t make sense.” I’m like, “I know I’ll spend my life as an empirical rationalist and I’ll figure everything out.” And then again, you walk up against these things, you bump up against these things and you’re just like, “All right, okay. I guess there’s a scientific explanation for this, but wow.” Then there’s like, “All right, where did that come from?” Then how far back can you go on the causality chain? Yeah. Then even just experiences that we all have on earth, it’s hard to rationally explain it all. And then, so yeah, I guess I’d just say I’m kind of radically open-minded, at peace with the fact that I’ll probably never know.

(03:41:27)
The other thing though, that’s happened, and maybe the more practical answer to the question is I think I have a much better understanding now of the role that religion plays in society that I didn’t have when I was younger. And my partner, Ben has a great … I think he quotes his father on this. He’s like, “If a man does not have a real religion, he makes up a fake one, and the fake ones go very, very badly.”

(03:41:48)
And so there’s this, it’s actually really funny, there’s this class of intellectual … There’s this class of intellectual that has what appears to be a very patronizing point of view, which is, “Yes, I’m an atheist, but it’s very important that the people believe in something.” And Marx had the negative view on that, which is religion is the opiate of the masses. But there’s a lot of right-wing intellectuals who are themselves, I think, pretty atheist or agnostic, that are like, it’s deeply important that the people be Christian or something like that. And on the one hand it’s like, wow, that’s arrogant and presumptive. But on the other hand, maybe it’s right because what have we learned in the last hundred years is in the absence of a real religion, people will make up fake ones.

(03:42:27)
There’s this writer, there’s this political philosopher who’s super interesting on this named Eric Voegelin. And he wrote in the mid-part of the century, mid-late-part of the 20th century, he was born in, I think, 1900, died in ’85. So he saw the complete run of communism and Nazism and himself fled, I think he fled Europe and the whole thing. His big conclusion was basically that both communism and Nazism, fascism, were basically religions, but in the deep way of religions. We call them political religions, but they were like actual religions. And they were what Nietzsche forecasted when he said, “God is dead. We’ve killed him, and we won’t wash the blood off our hands for a thousand years.” Is we will come up with new religions that will just cause just mass murder and death. And you read his stuff now and you’re like, “Yep, that happened.”

(03:43:20)
And then of course, as fully elite moderates, of course, we couldn’t possibly be doing that for ourselves right now, but, of course, we are. And I would argue that Eric Voegelin, for sure, would argue that the last 10 years we have been in a religious frenzy, that woke has been a full scale religious frenzy and has had all of the characteristics of a religion, including everything from patron saints to holy texts, to sin. Wokeness has, I think, has had every single aspect of an actual religion other than redemption, which is maybe the most dangerous religion you could ever come up with, is the one where there’s no forgiveness. And so I think if Voegelin were alive, I think he would’ve zeroed right in on that, would’ve said that. And we just sailed right off. I mentioned earlier we somehow rediscover the religions of the Indo-Europeans. We’re all into identity politics and environmentalism. I don’t think that’s an accident.

(03:44:15)
So anyway, there is something very deep going on in the human psyche, on religion, that is not dismissible and needs to be taken seriously. Even if one struggles with the specifics of it.
Lex Fridman
(03:44:33)
I think I speak for a lot of people that it has been a real joy and, for me, an honor to get to watch you seek to understand the human psyche as you described. You’re in that thirty-year part of your life, and it’s been an honor to talk with you today. Thank you, Marc.
Marc Andreessen
(03:44:50)
Thank you, Lex. Is that it? That’s only, how long is that?
Lex Fridman
(03:44:54)
Four hours with Marc Andreessen is like 40 hours of actual content so …
Marc Andreessen
(03:45:00)
I’ll accept being one of the short ones.
Lex Fridman
(03:45:01)
For the listener. Marc looks like he’s ready to go for 20 more hours, and I need a nap. Thank you, Marc.
Marc Andreessen
(03:45:11)
Thank you, Lex.
Lex Fridman
(03:45:12)
Thanks for listening to this conversation with Marc Andreessen. To support this podcast, please check out our sponsors in the description. And now let me leave you with some words from Thomas Sowell. “It takes considerable knowledge just to realize the extent of your own ignorance.” Thank you for listening and I hope to see you next time.

Transcript for Jennifer Burns: Milton Friedman, Ayn Rand, Economics, Capitalism, Freedom | Lex Fridman Podcast #457

This is a transcript of Lex Fridman Podcast #457 with Jennifer Burns.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Lex Fridman
(00:00:00)
The following is a conversation with Jennifer Burns, a historian of ideas including the evolution of economic, political, and social ideas in the United States in the 20th century to today. She wrote two biographies, one on Milton Friedman and the other on Ayn Rand, both of which I highly recommend. This was a super technical and super fascinating conversation. At the end, I make a few comments about my previous conversation with President Zelensky for those of you who may be interested. This is the Lex Friedman podcast. To support it, please check out our sponsors in the description. Now, dear friends, here’s Jennifer Burns.

Milton Friedman

Lex Fridman
(00:00:48)
You have written two biographies, one on Milton Friedman and one on Ayn Rand. So if we can, we will focus on each one separately, but first, let’s talk about the ideas that two of them held in common, the value of individual freedom, skepticism of collectivism, and the ethics of capitalism. Can you talk about the big picture ideas they converge on?
Jennifer Burns
(00:01:11)
Yeah. So, Milton Friedman and Ayn Rand, in the biggest picture, they’re both individualists, and they’re skeptical of collectivities and collectivism. So, their unit of analysis is the individual. What’s good for the individual? What works for the individual? Their understanding of society flows from that. They also both use this focus on individualism to justify and to support capitalism as a social and economic system. So, we can put them in a similar category. We can call them individualists. We could call them libertarians of a sort. They’re also really different in how they approach capitalism, how they approach thinking.

(00:01:52)
Ayn Rand developed her own moral and philosophical system to justify individualism, and to connect the individual to capitalism, and to support capitalism as a social and economic system. Friedman struggles a bit more with how to justify capitalism, and he’ll ultimately come down to freedom as his core value, like his God, as he says. So, freedom does connect back to the individual, but he’s not justifying capitalism for his own sake. He’s justifying it for its ability to underwrite freedom in a social sense and also in the individual sense
Lex Fridman
(00:02:28)
At a high level, are there interesting differences between them? You already mentioned a few, maybe in terms of who they are personally, maybe in terms of how they approach the justification for capitalism, maybe other ways.
Jennifer Burns
(00:02:39)
Yeah, for sure. So beyond this idea that Milton Friedman takes a while to come to his justification of capitalism, whereas Ayn Rand has it from the start. She really focuses on the core quality of rationalism and rationality. Rationality is the defining feature of human beings, and so she works from there, whereas Milton Friedman eventually converges on this idea of freedom. So, that’s one part of it. The other is their intellectual styles are really, really different. Their interpersonal styles are really different.

(00:03:12)
So, Friedman has big ideas, big principles that guide him, but he’s also deeply empirical. He spends most of his career doing historical research, economic research, pulling data from how people actually make economic decisions, and live in the world, and using them to test and refine his theories. Where Rand, to some degree, we could say she’s empirical and that she lives through the Russian Revolution, and takes a very big lesson from that, but her style of thinking is really first principles, an axiomatic approach going from the basic idea of rationality, and then playing it out in different spheres.

(00:03:50)
So, those are just very different intellectual approaches, and then they lead in some ways to really different ways of thinking about how you get things done in the world. Ayn Rand is a purist. She wants to start with the pure belief. She doesn’t want it to be diluted. One of her favorite sayings was, “It’s earlier than you think.” In other words, we’re still moving towards a place where we can really hold and express these ideals purely. Friedman, although he didn’t use this terminology, he was much more half a loaf guy like, “I’ll take what I can get, and then I’ll try to move to where I really want to be.” But he is able to compromise, especially when he moves from being an economist into being more of a political thinker.

(00:04:35)
So, that’s a really different intellectual style, and then it also plays out in their lives, and that Ayn Rand is incredibly schismatic. I mean, she wants her friends to believe what she believes and support what she supports. She’s willing to break a relationship if it doesn’t match. Milton Friedman, he also does tend to have friends who agree with him, yet he’s always willing to debate his opponents, and he’s willing to do so with a smile on his face. He’s a happy warrior, and he actually will win a lot of debates simply by his emotional affect and his cheerfulness and his confidence, where Rand will lose debates because she gets so angry in the face of disagreement.

(00:05:19)
So, they have a lot of similarities and a lot of differences, and it’s been really fascinating to dive deep into both of them.
Lex Fridman
(00:05:26)
I just re-listened to Ayn Rand’s, I think, last lecture, or at least it’s called that, and just the confrontational nature of how she answers questions or how she addresses critics and so on. There is a kind of charisma to that. So, I think both of them are very effective at winning over a popular support, but in very different styles. It seems like Ayn Rand is very cranky, but I mean, is the most charismatic, cranky person I think I’ve ever listened to.
Jennifer Burns
(00:06:00)
I mean, people talked about her, meeting her, and coming to believe in her ideas in a similar way as they did with Marxism, and that suddenly everything made sense, and that when they came to believe in objectivism, they felt they had this engine for understanding the entire world. Now after a while for most people, that then became confining, but that certainty. Friedman had some of that as well. He clothed it differently. He clothed it in happiness where Rand closed it, as you said, in crankiness or anger. I mean, there’s also an arc to Rand.

(00:06:35)
She gets angrier and angrier and crankier and crankier over the course of her life. What I enjoyed about my research is I was able to get into this early moment when she was different and a little more open, and then I watched her close and her harden over time.
Lex Fridman
(00:06:51)
Would it be fair to say that Milton Friedman had a bit more intellectual humility where he would be able to evolve over time, and be convinced by the reality of the world to change the nuances of policies, the nuances of how he thought about economics or about the world?
Jennifer Burns
(00:07:10)
Yeah, absolutely. Friedman believed in being able to say, “I was wrong.”
Lex Fridman
(00:07:15)
Right.
Jennifer Burns
(00:07:15)
There are some things he said he was wrong about. We’ll delve more into monetarism and monetary policy, but he was able to talk about the ways his ideas hadn’t mapped onto the world the way he thought they would. He does a really interesting interview at the end of his life where he’s beginning to voice some doubts about globalization, which was he was sort of a prophet of globalization, a cheerleader of globalization. He really thought it would lead to a better world in all respects. Towards the end of his life, it’s about two years before he dies, there’s a note of doubt about how globalization unfolded, and what it would mean particularly for the American worker.

(00:07:52)
So, you can see him still thinking, and that to me, I had assumed he became crankier and crankier and more and more set in his ways. Of course, there’s a phase where he does become that way, especially as he’s in the public eye, and there’s not room for nuance. But to find in the last years of his life him being so reflective, that was absolutely not something Rand could do.
Lex Fridman
(00:08:13)
I think there’s a thread throughout this conversation where we should actually also say that you’re a historian of ideas. How’s that?
Jennifer Burns
(00:08:21)
I am a historian of ideas, yes.
Lex Fridman
(00:08:23)
So, we’re talking about today in part about two people who fought for ideas, for an idea, like we’ve mentioned freedom for capitalism. They did it in very different ways. It’s so interesting to see the impact they both had and how their elucidation explanation of those ideas reverberated throughout society, and how we together as a society figure out what works, the degree to which they have influence on the public, the degree to which they have influence on individual administrations like the Reagan administration, Nixon and so on, and how it might return, fade away, and then come back in the modern times.

(00:09:07)
It’s so interesting if you just see this whole world as a game of ideas where we were pushing and pulling and trying to figure stuff out. A bunch of people got real excited over 100 years ago about communism, and then they try stuff out, and then the implementation broke down, and we keep playing with ideas. So, these are the two greats of playing with ideas. I think that’s a thread that just runs through this.
Jennifer Burns
(00:09:37)
Yeah, and of pushing back against that movement towards communism, social democracy. But one difference that I really should emphasize, Rand is a writer of fiction. She’s a philosopher, but she’s also a writer of fiction. So, she is working almost in the mythic register, much more in the psychological register. She’s creating characters that people identify with and people relate to experiences they’ve had, and that’s one of the reasons she hits so deep. She’s also offering people… I read all the fan letters to her. People would say things like, “I read The Fountainhead, and now I’m getting a divorce,” having just these incredible realizations.
Lex Fridman
(00:10:23)
Milton Friedman didn’t get such letters.
Jennifer Burns
(00:10:24)
Milton Friedman didn’t get such things, or I’ll meet someone, and they’ll say to me, “Ayn Rand is the reason I went to medical school.” A woman said this to me a few years back, “It never even occurred to me that I could be a doctor until I read Ayn Rand, and I said, “I’m going to go to medical school.” So, she has that really intense impact on people. So, she thought of herself as rational. She thought of rationality as what she was doing, but she was actually doing a kind of mythopoetic psychological work as well.

(00:10:55)
Whereas Friedman, on the one hand, was much more rational, and there’s a whole set of economic thinking. He provides a rational framework for understanding the world, and it’s the framework of neoclassical economics. At the same time, he does pull on mythologies of the idea of America and the Gilded Age, the frontier mythology, the individual immigrant, the settler mythology. He pulls on these, but he doesn’t create them, and he’s more kind of playing a tune he already has. Whereas I think Rand really does something a little bit deeper in her ability to reach into people’s psyche, and then take that emotional psychological experience, and fuse it to an intellectual world and a political world.

(00:11:43)
That’s really what makes her so powerful. So, I think she comes back in to relevancy in a different way than Friedman does, because I think in some ways she’s tapped into a more universal human longing for independence and autonomy and self-creation and self-discovery.
Lex Fridman
(00:12:03)
Nevertheless, there’s still pragmatic ideas that are still important today from Milton Friedman, even just on the economics level. So, let’s dig in. Let me try. I took some notes. Let me try to summarize who Milton Friedman is, and then you can correct me. So, he is widely considered to be one of the greatest and most influential economists in history, not just the 20th century, I think, ever. He was an advocate of economic freedom, like we said, and just individual freedom in general. He strongly advocated for free market capitalism and limited government intervention in the economy, though you do give…

(00:12:44)
I’ve listened to basically everything you have on the internet. You give some more depth and nuance on his views on this in your books. He led the famed Chicago School of Economics, and he won the Nobel Prize in economics in 1976. He greatly influenced economic policies during the Reagan administration and other administrations. He was an influential public intellectual, highly influential, not just among economists. He lived 1912 to 2006. So, that means he lived and worked through some major world events where his ideas were really important, the Great Depression with the New Deal, World War Two with post-war reconstruction, the Rise and Fall of the Bretton Woods Monetary System as we may talk about, the Cold War and all the conflicts involved in that, the tensions around communism and so on, so the fall of the Soviet Union.

(00:13:44)
Also, he has some interesting relationships to China’s economic transformation since the 1970s, the stagflation of the 1970s, and I’m sure there’s a lot more. So, can you maybe continue this thread, and give a big picture overview of the ideas he’s known for?
Jennifer Burns
(00:14:03)
Yeah, sure. That’s a great summary. You learn fast. So, let me start with the economics, and then I can transition to how he used those economic ideas to become a real voice in the American conservative movement, in the American political realm. So, I’ll highlight for ideas or contributions or episodes. One was his work with Anna Schwartz in revising our understanding of the Great Depression. That’s tightly related to the second, which is the School of Monetarism that he and Schwartz really become founders of.

(00:14:40)
Then there is the prediction of stagflation and the explanation of that in the 1970s, which really is one of these career-making predictions, and we can dig into that. Then in terms of technical economics, he’s known for the permanent income hypothesis, which he develops with a group of female collaborators that I can talk about. So, those are four technical pieces end up being really brought together in what becomes the Chicago School of Economics. He’s undoubtedly the head and the leader of the Chicago School of Economics. There’s an earlier generation that he learns from. There’s his generation.

The Great Depression


(00:15:23)
There’s also a Chicago School of Law and Economics that’s really profoundly influential, and then there’ll be a third generation that he’s somewhat distinct from, but that goes on to really shape economics. But let me go back to these four pieces, and let me start with Great Depression. So, Milton Friedman actually lives through the Great Depression. He’s in college when it hits. So, he is in college just 1928 to 1932. He’s aware of the Depression, and he’s deciding, “Should I study mathematics, or should I study economics?” He’s had some good economics teachers, but it’s really the context.

(00:16:06)
It’s looking around at the slow dissolving of economic prosperity. So, he decides to go to Chicago. He decides to study economics. What’s really interesting is that the Great Depression is so unexpected. It’s unpredicted. It’s unprecedented, and economists are really struggling to know how to respond to it. So, he’s going to arrive at the University of Chicago when the field is struggling to know what to do. So, he’s in this really open space where the institutional economics of the 1920s has failed to predict, which was focused on business cycles.

(00:16:44)
This is the irony. Their big thing was charting and understanding business cycles, and then we have the biggest business cycle of all time. They haven’t seen it coming, and they don’t have a good explanation for it. What he will get at Chicago is the remnants of a monetary understanding of the economy. So his teachers, they don’t know exactly what’s going on, but they look first to the banking crisis. They look first to 1933. It’s bank runs failures of maybe it’s up to a third of American banks. Thousands of banks are failing per week. So, they’re focused on that. So, that’s the first imprint he will have.

(00:17:26)
The Great Depression has something to do with a banking system. The second imprint he will have is that all of his professors are profoundly concerned about the social crisis. They want relief programs. They want them now. They want bank regulation and financial reform. They’re very active. This is not laissez-faire by any stretch of the imagination. So, Friedman has that imprinting. So, he gets there in ’32, ’36, ’37. The ideas of John Maynard Keynes from Britain, which has a different explanation, Keynes has a different explanation of the Great Depression, will make landfall in American economics, and be very profoundly influential. Most American economists, but Friedman already…

(00:18:10)
It’s too late for Friedman. He already has a different perspective. So, Keynesianism unfolds. I can say more about that, but it basically leads to more active federal government participation in the economy. What underlies a lot of that, its adaptation in America particularly, is the idea that capitalism has failed. Capitalism has revealed itself to have a profound flaw, and that its cycles of boom and bust creates social instability, chaos. It needs to be tamed. It needs to be regulated. So, that becomes the baseline of politics in the United States, the understanding of the New Deal, the understanding of the Democratic Party, even to some extent the understanding of the Republican Party.

(00:19:01)
Friedman never quite sure about that. He has a hunch that there’s something else going on, and he does not buy that capitalism has ground to a halt, or the other idea is that capitalism has gone through some phase transition. It worked great maybe while we had a frontier. This is a very serious argument that people were making. United States used to have a frontier, a place where Europeans hadn’t fully settled. Of course, they’re pushing out the native tribes. That’s another story, but that this frontier is the engine of economic growth, and the frontier is now over. It’s closed, and we’re going to stagnate. There’s a theory of secular stagnation.

(00:19:40)
So, to deal with secular stagnation, we’re just going to have to have a more active state. So, Friedman is suspicious of all these assumptions, and he has this idea that it’s something to do with money. Money is somehow important, and so he joins together with Anna Schwartz, who is an economist. She doesn’t at this time hold a PhD. She’s working for the National Bureau of Economic Research. They come together to do this study of money in the U.S. economy. It takes them 12 years to write the book. They’re releasing their ideas, and they’re arguing, and Friedman is writing papers, giving talks, saying, “Money’s really important,” and nobody’s really believing him.

(00:20:21)
He’s a crank. He’s at Chicago. Chicago is a well-known university, but he’s considered a crank. Then in ’63, he and Anna Schwartz published this book, and it’s 800 pages. It’s a reinterpretation of the history of the United States through money, like the central character is money, whether it’s specie, greenback or the U.S. currency. They have a whole chapter on the Great Depression, and what they’ve literally done, Schwartz has done most of this, they’ve gone… Schwartz has gone to banks, and said, “Show me your books.” Then she’s added up column by column, “How much money is in your vault? How much money is on deposit? How much money is circulating?”

(00:20:59)
So, they literally have graphs. You can see them in the book of how much money has been circulating in the U.S. at various different points in time. When they get to the Great Depression, they find the quantity of money available in the economy goes down by a third. In some ways, this is completely obvious because so many banks have failed, and we don’t have any type of bank insurance at that point. So if your bank goes under, your savings are there, the money essentially vanishes, and it’s fractional reserve banking. So, you’ve put in… They can loan up to 90% on their deposits.

(00:21:35)
So, Friedman and Schwartz present this argument that what really made the Great Depression so bad was this drop in the amount of money, the 30% drop in the money, they called the Great Contraction. Then they go further, and they say, “Well, how did this happen and why?” They pinpoint the Federal Reserve, which is a fairly new institution at that time. They say, “What did the Federal Reserve do? The lender of last resort, what did it do in the face of what they’re depicting as a massive, unprecedented liquidity crisis?” They find it’s not really doing much.

(00:22:09)
They really dig into the details, and they find that the Federal Reserve has gone through a sort of personnel change. Some of the key leaders in the 1920s, Benjamin Strong is one of them. He’s now deceased, and the dominance of the New York Federal Reserve, which in their telling is global, it’s interconnected. It’s seen a lot of financial things come and go. They believe that the New York Fed had the understanding to recognize this is a liquidity crisis. We should be very generous. We should support all the banks. Their influence has diminished for the kind of banks that are more… They don’t say the Rubes and the Hicks, but it basically is.

(00:22:49)
It’s like, “The people in charge don’t know what they’re doing.” So, the Fed pursues this policy of masterly inactivity. They don’t see it as a problem. They don’t do much. There’s an enormous liquidity crisis, and that’s their version of what the Great Depression is all about, that it’s a financial system meltdown. It’s a liquidity crisis, and that it in some ways, well, in many ways, they argue very strong counterfactual argument. The Federal Reserve could have prevented it, and it did not. So, it becomes then an institutional failure and a political failure, not a failure of capitalism as a system. So, this book comes out. It’s a blockbuster.

(00:23:34)
Even those economists who’ve been like, “Friedman is a crank. I don’t buy it,” are like, “Friedman and Schwartz are onto something. Milton Friedman and Anna Schwartz are onto something.” So, that really changes the game. This is also one of his most influential contributions, because Friedman and Schwartz becomes the playbook for the Federal Reserve. We have lived through this, the financial crisis. The Federal Reserve is ready to loan. Covid, the Federal Reserve does all kinds of new things, because no Federal Reserve chair wants to be in Friedman-Schwartz 2.0 that somebody writes, where they’re the bad guy who let the economy meltdown.

(00:24:16)
So, the specifics of what they say to do have obviously evolved as the system has changed, but this is a playbook for how to deal with economic crisis. It’s Friedman and Schwartz. So, it’s absolutely fundamental, and that is really going to be the place he makes his mark.
Lex Fridman
(00:24:34)
There’s a lot of things to say here. So first, the book we’re talking about is a monetary history of the United States, in part for which Milton Friedman won the Nobel Prize. You’ve also mentioned the influence of the Great Depression, if you could even just rewind to that.
Jennifer Burns
(00:24:34)
Yes.
Lex Fridman
(00:24:48)
So, he went to, I guess, college in Rutgers.
Jennifer Burns
(00:24:50)
That’s right.
Lex Fridman
(00:24:51)
He was mathematical proclivities, so he was wanted to be a mathematician. So, it’s a cool crossroads. It’s interesting how the right time, the right person arrives. So, you described this really well, that he had this choice to be a mathematician or an economist. The economist is the University of Chicago. Mathematician is Brown University, whichever, and then this is also the beginnings, as you’ve described, of mathematical economics. So, he fits in nicely into this using… I think you said the number of equations started going up per paper-
Jennifer Burns
(00:25:35)
Yes.
Lex Fridman
(00:25:36)
… which is a really nice way to put it, so really the right person at the right time to try to solve this puzzle of the economy melting down. It’s so interesting, just one human. It’s just from just zooming in on a single human making a decision about life. It’s hard to know when you’re in it that the world is melting down from an economics perspective, and that I could do something about this to figure out what it is. Also, I’m going to reject the mainstream narrative about why this happened.
Jennifer Burns
(00:26:11)
So, the other piece of the puzzle, when he goes to Rutgers, he thinks he’ll be an actuary. So, Milton Friedman’s family, his parents are immigrants, Jewish immigrants from Eastern Europe. They’re pretty atypical and that they don’t stay in New York, and they moved to Rahway, New Jersey. They put together a fairly middle-class life as kind of… They have a shop. They do some wholesale buying and selling, and then his father dies when he’s 16. His life becomes more precarious, but it’s never as precarious as he makes it up be. He’s got three older sisters. They earn a good living.

(00:26:45)
Incidentally, they all have better grades in high school than he does, but he’s the one that goes to college. But it’s actually really important that he loses his father figure, because he’s then looking for other father figures, and he meets two at Rutgers. One is Arthur Burns, who will go on to have a huge influence in his career. No relation to me, by the way. But, Arthur Burns is like him, a fellow Jewish immigrant boy on the make. He’s older, and he’s making a career as an economist. Then there’s Homer Jones who has gone to the University of Chicago, and is studying with Frank Knight at Chicago, and says, “You have to go to Chicago.”

(00:27:25)
So, he has these two mentors. Burns in particular suggests, ” Oh, I could be an economist. That could be my career path.” The idea to be an actuary for an insurance company, I’m not sure where he got that idea, but he just thought that was something he could do as someone who was good at math. So, the college really opens the perspective, opens the door. Then I think it’s really key that, again, he doesn’t get an explanation that he buys for the Great Depression. So then he’s looking for one. The math part is really interesting aspect of his career. Now, he actually comes to Chicago to study with the mathematical economist, Henry Schultz, but he gets there, and he thinks Schultz is kind of dumb. He really does. He’s incredibly arrogant, and he just thinks, “This guy’s not that smart,” and it seems that… I mean, Schultz did some really important work in the early stages of mathematical economics, but a lot of the oral histories about him are like, “Yeah, he wasn’t that bright.” So, Friedman’s may be onto something. So, he falls into the set of students who are really enthralled with this other professor, Frank Knight. Frank Knight is against math and economics. Frank Knight is a neoclassical economist, but not a mathematical economist.

(00:28:45)
He’s an old school liberal. He’s really concerned about liberal democracy, economic liberalism. Friedman is very deeply influenced by Knight. He continues to pursue mathematical economics, so he’ll go… For part of his graduate career, he goes to Columbia University where he actually gets his PhD from. He works with a mathematical economist there. So, he comes out trained in what will eventually be econometrics, statistics and economics. His early publications are in statistics, but it’s not really where his intellectual heart and soul are. Eventually, he will turn very profoundly against mathematics and economics, and become a heterodox strain throughout 20th century economics.

(00:29:29)
It says, ” Simple models are better. We need to work off empirical data, not construct elegant models,” and becomes really counter-cultural within economics in that way.
Lex Fridman
(00:29:45)
The test of a good model is it should actually predict stuff that happened.
Jennifer Burns
(00:29:49)
It should predict stuff that happened. It should tie back to what’s going on.
Lex Fridman
(00:29:53)
I’m wondering which direction to go. So first, actually, if we could zoom out on the different schools of economics.

Schools of economic thought

Jennifer Burns
(00:29:59)
Yeah.
Lex Fridman
(00:29:59)
Just the basics. You mentioned Neoclassical. We mentioned Keynesian economics. We mentioned… What else did we mention? Well, the Chicago School of Economics.
Jennifer Burns
(00:29:59)
Right.
Lex Fridman
(00:30:08)
Where does Austrian economics fit into that pile, and Marxian economics? Can we just even just linger and try to redefine Keynesian economics and Chicago School of Economics and Neoclassical economics and Austrian economics, because there’s some overlap and contention?
Jennifer Burns
(00:30:27)
For sure. So, schools of economics, so we could start with classical economics. Classical economics, we could think of Adam Smith is your classic classical economist, the founder of the discipline. Classical economics does not really use math, is very close to political economy. It’s concerned with, as Smith puts it, the wealth of nations. It’s concerned to some degree with distribution. It’s concerned to some degree with what makes a good political system. What tends to really define classical economics when you’re looking from a great distance is what’s called the labor theory of value.

(00:31:08)
So, where does value come from in classical economics? It comes from the labor that a person puts into it. So, maybe this in some ways comes from Locke’s notion of property that you mingle your labor with the natural world. We can say labor theory of value. So, classical economics concerned with… Smith is arguing against mercantilism for more free trade, often goes by the name of political economy to show it’s more capacious. It’s thinking of politics and economics. You can still read these books today. The sentences are long. The words are different, but you can still follow along.

(00:31:49)
So, the real big transition from classical economics and political economy to economics as it’s understood today, comes with the marginal revolution. The marginal revolution is a scientific revolution that happens in a couple of different places simultaneously. This is one of these things that you see in the history of science. There’ll be some breakthrough like, “Darwin has a breakthrough,” but somebody else has sort of the same breakthrough at the same time, totally differently. So, there’s a version of marginalism that’s continental. There’s a version in the German-speaking lands, in the French-speaking lands and in Britain. They all come together, and the shift is in the theory of value. So, the theory of value in marginalism is on the margin.

(00:32:37)
So, say you have one apple, and you want a second one. How much is going from one apple to two apple worth for you? Probably quite a bit. If you had 10 apples, maybe going to 11 apples doesn’t matter that much. The marginal value is less. So, what marginalism does though, most importantly, is it opens the door to math and economics, because it means you can graph this now. You can depict this relationship graphically. There’s some really interesting work in the history of economics that shows a lot of the people who developed marginalism were looking to physics as a model, Physics, the queen of the sciences.

(00:33:22)
So, they were thinking… They imported terms from the natural world to describe the social world through the lens of economics terms like equilibrium. So, the idea being that if you looked at a market, a market would reach equilibrium when everybody has bought and sold all that they want, or the price will settle at an equilibrium price when it’s really the demand and supply are matching up.
Lex Fridman
(00:33:50)
Some of these ideas are things we would pick up at a microeconomics class.
Jennifer Burns
(00:33:55)
Oh yes, this is still out there. This is the basic foundation of microeconomics, marginal analysis, and so-
Jennifer Burns
(00:34:00)
… of microeconomics, marginal analysis. And so, in the German-speaking intellectual tradition, this is the root of Austrian economics and people picking up the marginal revolution in the German-speaking lands are opposed to the historicists who are thinking in a more evolutionary way about how societies grow and change and they have a vision of economic ideas as applying differently to different types of social arrangements. Where the marginalists, remember, inspired by physics and this is a set of natural laws that applies anywhere to any human society. So, that’s his first really big fissure that we’ll see again and again. Are you historically minded? Do certain traits of economic life in here adhere and become expressed in certain types of societies or are there universal economic laws that flow through any type of society?

(00:35:03)
So, that’s a juncture, a break and so marginalism … First, people start using, really, geometry to graph things but marginalism is also opening up to the possibility of calculus and the possibility of creating models. But at that point in time, late 19th century, a model is something like a physicist does. Think of an inclined plane and how fast does the ball roll from one to the other, it’s a physical representation of the world and eventually economists will start to create mathematical representations of the world but we’re not quite there yet. So, we’re late 19th century and we have this fissure, we have this introduction of marginal analysis that marks the juncture from classical economics to economics. So, let’s say now we have economics but we still have this fissure between historical thinking and let’s call it natural law thinking, that’s not quite right, but physical laws versus contingency. And then, in the United States, this ends up mapping onto debates about capitalism and so more historically minded economists tend to be interested in the progressive movement which is invested in taming and regulating industrial capitalism and changing its excesses, factory safety laws, wage laws, working conditions laws. Yet, in general, American economists all use marginal analysis just in different ways. The ones who are more drawn to marginal analysis become known as neoclassical economists. They’re neoclassical, the neo is because they’re using marginal analysis, the classical is because they don’t think we need to change the way the economy operates or the government operates, they’re not progressive. Whereas the progressives are saying things like we need to use social control. The state and the people, collectively and democratically, need to control the way economics unfolds and make sure things are fair and equal.

(00:37:06)
So, that school of thought becomes known as institutional economics in the United States by the 20th century. So, it’s part of the progressive movement late 19th century, into the 20th century, it really becomes institutional economics and it’s quite dominant and the neoclassical economists are still there but they’re very much a minority. And Frank Knight, Milton Friedman’s teacher, is one of the minority neoclassical economists and the institutionalists are much more progressive still.
Lex Fridman
(00:37:38)
Is it fair to say that the neoclassical folks and even the classical folks versus the institutional economics folks, they have a disagreement about how much government intervention that should be in the economy? So, neoclassical is less intervention and then an institutional economists, the progressive folks is more intervention?
Jennifer Burns
(00:37:58)
Yes, yes, exactly right. So, this is a situation in the 1920s but the other piece I should mention is that first generation of progressive economists were very radical, they were closely allied with the socialist movement, with labor radicalism and many of them lost their jobs at universities. This connects to the dawn of academic freedom, this is before academic freedom and they were chastened, they became much more mainstream. By the time we get to the 1920s, we don’t really have radical critiques of society coming from economists, much smaller profession, much less important than it is today and fairly peaceful because the 1920s are a fairly peaceful decade in the United States.

(00:38:47)
So, this is a situation when the Great Depression hits and, as I mentioned before, the most important institutional economist is Wesley Mitchell and he has said … He’s written a whole book on business cycles. He doesn’t see this business cycle coming and it hits and he doesn’t have a good explanation for it. Now, perhaps the preeminent neoclassical economist was Irving Fisher. Now, Irving Fisher is big into the stock market and Irving Fisher says sometime in late summer 1929 stocks are going ever higher and will continue to go ever higher forever. And so, he loses his reputation after the stock market crash. So, Milton Friedman is stepping into a field in which the greats have been discredited and there’s an enormous economic crisis all around.
Lex Fridman
(00:39:38)
And everybody’s struggling to figure out why the crisis happened, right?
Jennifer Burns
(00:39:42)
Yes. And the other thing he’s stepping into is a world where, in the United States, there’s a great deal of anger at capitalism, at the system, unemployed people on the street in Europe, there’s rising fascist movements, in Asia there’s rising fascist movements. And so, everyone’s very concerned about this and Friedman is seeing a lot of this through the lens of Frank Knight who feels like we are maybe reaching the end of what he calls liberalism, he calls himself an old-fashioned liberalism. We’re reaching the end of representative Democratic government because representative Democratic government cannot solve these social problems. And capitalism as it has developed, Knight is very pro-capitalist but he says it’s generating inequality and this is putting too many strains on the system.

(00:40:31)
So, Knight will become one of the people who helps Friedman think how do I develop a new theory of capitalism that works in an era of mass democracy where people can vote and people can express at the ballot box their unhappiness with what’s happening economically. So, this larger movement will generate, of which F.A. Hayek is a part, Friedman is a part that becomes the very early stirrings of trying to think about a new liberalism which will eventually be called neoliberalism.

Keynesian economics

Lex Fridman
(00:41:05)
Okay. So, if we can just linger on the definitions of things. So, we mentioned what neoclassical is and the institutional economics is, what’s Keynesian economics? And the Chicago School of Economics, I guess, is a branch of neoclassical that’s a little bit more empirical versus maybe model-based and Keynesian is very model heavy, more intervention of government. So, the real battle is Keynesian versus everybody else.
Jennifer Burns
(00:41:36)
That is what eventually comes to pass in the United States and in the overall developed profession of economics. The other piece of the puzzle here is the introduction of mathematics and it’s been around the edges but it will pick up speed. In the 1930s, the Econometrics Society is founded, they start publishing, people start using more statistical and mathematical tools to think about economics and they’re given a boost inadvertently by the rise of Keynesian economics. So, Keynes is trained in the neoclassical tradition, he’s an absolutely fascinating figure, he’s been there in the peace negotiations at Versailles, he basically calls World War II. He’s like, “Hey, we’re going to have another war here caused by Germany because this peace treaty has been done in such a vindictive way and people have made such bad decisions.” He’s there, he sees it happening.

(00:42:32)
And so, when the Great Depression unfolds, he basically comes up with a new theory for explaining what’s going on and the previous neoclassical understanding as things go up and things go down and, when they go down, there’s a natural mechanism to bring them back up. So, when the economy’s going down, prices are going down, wages are going down, everybody’s losing money but eventually firms are going to realize, hey, I can hire people cheap, hey, I can buy stuff cheap, I don’t have a lot of competition maybe I should get in the game here and then others will start to get in and then you regenerate prosperity in that way. And so, Keynes says, “Sure, that’s one theory but something different is happening right now. Part of why it’s happening is because we have … The working class is more empowered now, they’re not simply going to just take low wages and ride them down to the floor, we might not hit the floor.”

(00:43:28)
But also he says people might become too anxious to spend, they might not want to invest. And Keynes has these discussions of animal spirits, he’s still enough of a political economist to think, not just in terms of human rationality, but what are some other things going on in human beings. And people might decide to sit on their money, they might not invest it and so what happens then is you could get stuck in a bad equilibrium. So, in the neoclassical model of equilibrium, it restarts and resets itself and he says, “No, we could get stuck here, we get stuck in the Depression. And in that case, what has to happen?” He says, “The government stimulates investment and the government itself invests.” And then he argues that, this is a student of his, Richard Kahn says, “As the government invests a dollar, it has a multiplier effect. The dollar spent by the government ramifies out throughout the economy.”

(00:44:26)
So, it takes the government and puts it in the center as opposed to, say, the banking system or the financial system which would be the more Friedman analysis. And for many economists of Friedman’s generation, and he’s a weird generation because, the generation that becomes dominant, it’s just four years older, the men who become Keynesian economics but that four years is really important because they come in to grad school in economics and they get exposed to the new ideas of John Maynard Keynes and they … I think it’s Paul Samuelson calls it, it was like a South Sea virus that attacked, all of the younger economists immediately succumbed and no one under 50 ever got the disease because their thinking is already set. And so, Keynesianism, Keynes himself is very suspicious of math and economics and he and Friedman is fascinating.

(00:45:20)
One of the first books by Jan Tinbergen, a Dutch economist, to use math and economics use huge volumes. Volume one, Keynes pans it. Volume two, Friedman pans it. So, they’re in the same page but what happens is, as Keynesianism arrives in the United States, Franklin Roosevelt is not really a Keynesian, he’s an accidental or experimental Keynesian and there’s a bunch of different ideas in the United States that are very similar to Keynesianism. They’re not theorized but there are similar ideas that the government has to do something. So, this all comes together and American economists realize that you can construct models in the Keynesian perspective and, if you can use numbers in these models, you can go to Washington, D.C. with numbers and you seem like you have a lot more authority and so math becomes really twinned into Keynesian economics.
Lex Fridman
(00:46:24)
So, the numbers are used as a symbol of expertise. We really know what the hell is going on because we have some numbers, right?
Jennifer Burns
(00:46:32)
Right. And we can create a model and so we can say, okay, in the model, the interest rate is here and taxes are here so let’s play with government spending. Let’s make it up, let’s make it down and then we can get an estimation it’ll spit out, here’s predicted GDP. So, the other piece of the Keynesian revolution is it really gets people thinking holistically about the economy as one conceptual unit and you then have what Paul Samuelson will end up calling the neoclassical synthesis and this is still in economics today. If you take micro, you’re going to get supply and demand, scarcity, marginal analysis. If you take macro, you’re going to get a very different approach and that’s more Keynesian based. And so, the idea is that, and this makes sense, you can think of this from statistics, the way things act individually versus when they’re all added together, it can be very different.

(00:47:27)
So, there’s this uneasy piece where economists are using neoclassical tools to analyze individual behavior and individual market behavior and they’re shifting to a different paradigm when they think about the economy as a whole. And in this paradigm of the economy as a whole, the federal budget, the taxing and spending power of the federal government become paramount and that is called the fiscal revolution and that’s really the essence of Keynesianism. But the key thing to remember is that Keynesianism and Keynes are different and there’s this famous episode where John Maynard Keynes comes to D.C. and he goes to dinner and he comes back and he says to one of his friends in London, he said, “Yeah, it was really interesting, I was the only non-Keynesian there.”
Lex Fridman
(00:48:13)
Yeah.
Jennifer Burns
(00:48:15)
Yeah.
Lex Fridman
(00:48:16)
So, Keynesianism is more government intervention, fiscal policy, so put the government at the center of influencing the economy. And then the different flavors of whether it’s Austrian Economics or Chicago School of Economics is saying, no, we have to put less government intervention and trust the market more. And the formulation of that for Milton Friedman is trust the money more, not trust but the money supply is the thing that should be focused on.

Laissez-faire

Jennifer Burns
(00:48:52)
Yes. So, the Austrians and the Chicago School see economic prosperity and growth comes from individual initiative, individual entrepreneurship, private sources. The private market is what drives economic growth, not the public sector. And so, for Friedman then the question is what is the government’s role. And because he’s lived through the Great Depression, he’s not laissez-faire and he won’t ever be laissez-faire. Now, interestingly, Hayek, living through the Great Depression, at first is laissez-faire and he’s like, “Sure, let it rip.” And things get so bad that Hayek’s like, “Okay, that’s not going to work.”
Lex Fridman
(00:49:33)
Can we actually define laissez-faire? So, what do we mean? What’s the free market? What’s laissez-faire? What’s the extreme version here?
Jennifer Burns
(00:49:40)
So, yeah, laissez-faire means leave it be in France. It’s more often used as an insult than as an actual. Very few people are completely and totally laissez-faire, that would be the pure laissez-faire would be maybe pure anarchist position like the state does nothing or the state isn’t even there. But it tends to, if I could maybe make it more precise, it would be focused on freedom of contract would be essential and that means the buyer of labor and the seller of labor must have absolute freedom to contract. So, that means no minimum wage law, no working hours law, no employment law, things like that. This is all pre-progressive movement, a lot of things are that way. Imagine you’re in 19th century America and you have a farm and you hire someone to help you on the farm, you offer the money, they take it. If they fall off a ladder and break their back, maybe you help them out, maybe you don’t but there’s not a whole apparatus of legal liability and safety and things like that.

(00:50:47)
So, that would be one piece. Another piece of laissez-faire would be free trade amongst nations, so no regulation of who can invest in a nation or who can take money out of a nation. So, Nippon Steel could come and invest in US Steel and there would be no grounds in which to reject that. Or you could, as a billionaire in the United States, relocate you and all your money to another country and the United States couldn’t try to keep you and nobody else could stop you from coming in. And then in the context of economic crisis, laissez-faire would not encompass centrally provided relief because, in the pure theory, again, very seldom applied purely but in the pure theory, the wages need to come down far enough and people need to be desperate enough to start taking work and to start the machine again. So, the theory would be, if you give people relief, they might not go back to work.

(00:51:51)
Now, almost nobody says that in the Great Depression because the situation is so bad and people are starving on the street and feel, for humanitarian and ethical reasons, it’s not okay to say that. The Austrians though at first, Hayek and Lionel Robbins are like, “This is a business cycle and it needs to run its course and it will be detrimental if we intervene,” and then pretty soon Hayek has to change his tune.
Lex Fridman
(00:52:18)
So, the Austrians are the most hard core in terms of laissez-faire, right?
Jennifer Burns
(00:52:21)
Absolutely. And so, Hayek will make the turn towards accepting more of a state and then will come to talk about how the state needs to support what he calls a competitive order. But his mentor Ludwig von Mises still remains very hard core and is not really open to things like unemployment insurance or other state-based interventions.
Lex Fridman
(00:52:46)
What does von Mises say about human suffering that’s witnessed in the Great Depression, for example? What are we supposed to, as economists, as humans that define policy, what are we supposed to see when people are suffering at scale?
Jennifer Burns
(00:53:01)
Yeah, I wish I knew an answer to that question, I don’t know enough about von Mises and his reaction in the Great Depression. I think I would hazard that he would look more to the down the road and say, “Well, if you start here, you’re going to go places that are bad,” but I don’t factually know what he said in response. I do know that Hayek’s position doesn’t last very long, it’s not a position you can hold to. Maybe you could hold to it in other cycles. The other thing that was interesting is I found very few Americans saying this, most who were were small town electeds or the most famous is Andrew Mellon quoted by Herbert Hoover. So, not directly, we don’t have him on record saying this but apparently Hoover records in his memoirs that Mellon said something like liquidate real estate, liquidate stocks, purge the rottenness out of the system, people will live a healthier life.

(00:54:06)
And certainly, there were members of the Federal Reserve who felt like it would create, they didn’t say moral hazard, but it would create what we now call moral hazard, bad habits. Were we to intervene and to save failing banks because failing banks need to be taught a lesson, whey need to be taught discipline? And so, a lot of people, I think, saw it in the context of discipline, this is discipline and, if you remove the discipline, you’ll be taking away something fundamental in society.
Lex Fridman
(00:54:34)
So, Milton Friedman never quite went all the way to laissez-faire?
Jennifer Burns
(00:54:39)
No, no, he didn’t see that and what’s really interesting is the number of incredibly radical proposals that he and his teachers were floating. So, I’ve mentioned Frank Knight, another really important influence on Friedman was Henry Simons who was a junior professor at Chicago. And Simons had this idea for what he called 100% money which would be a law that says banks have to hold 100% of the deposits they receive, they can’t loan them out on the margin. This would completely and totally have overhauled the US banking system and he would’ve said there’s a category of things called banks where you get deposits and then there’s going to be a category of, he didn’t say investment banks, but investment vehicles that will invest.

(00:55:24)
So, similar to what did happen, in some ways, in the banking reforms in that, in the 1930s, the investment banks were split from the deposit banks and the banks that took deposits were much more highly regulated and they were supported by the FDIC. But the point being, the Chicago School had these very radical proposals for reform. Go off the gold standard, restrict the currency, change the banks, immediately relief payments now. What is important to note though is that they thought of all of those as emergency measures to get through the emergency not as permanent alterations in the state of what had to be and not permanent alterations between state and market where the Keynesian assumption is things have changed, times have changed, we’re in a new dispensation and we need a new relationship.

(00:56:17)
So, Friedman is very, Milton Friedman is very open to doing things differently in a state of emergency. He will have different ideas during World War II than any other time and that’s why I argue I think he would have been supportive of at least the first rounds of coronavirus relief because I think he would have put his emergency thinking hat on. So, in that way, he was definitely more flexible.

Friedrich Hayek

Lex Fridman
(00:56:43)
You mentioned Hayek. Who is this guy? What’s his relationship to Milton Friedman in the space of ideas and in the context of the Great Depression? Can we talk about that a little bit?
Jennifer Burns
(00:56:55)
Sure. So, F.A. Hayek is an Austrian economist who takes up a posting in London and he’s a mentor, a mentee rather of Ludwig von Mises. He’s writing about business cycles, Austrian Capital Theory and the Depression hits and he’s one of the few economists who, in the beginning, really is not calling for much intervention. Although, as he realizes how politically unpalatable that is, he will develop a more softened version of Austrian economics that has room for a whole range of social services. What’s significant about Hayek is that he is also watching what’s happening in Austria, what’s happening in Germany and he’s really worried the same thing is going to happen to the Western democracies. And he sees the root cause of this is socialism, the shift towards an expanded role for government which, we’ve been talking about, is happening in the United States, it’s also happening in Britain.

(00:57:54)
And so, he writes this book that becomes incredibly famous, The Road to Serfdom, basically saying taking these steps towards a planned economy or an economy that’s a modified form of capitalism is going to … Could, he’s very clear that this is not an inevitability but, if the same steps are taken and people follow the same line of thinking, we may end up in a coercive, totalitarian state. So, this becomes enormously popular in the United States. First of all, he’s in good touch with Friedman’s teachers even before this book comes out, they see them as kindred spirits. Frank Knight is in touch with him, Henry Simons is in touch with him, they all see themselves as liberals, they call themselves old fashioned unreconstructed liberals. And so, even before he becomes famous, Hayek will be trying to organize thinkers and intellectuals who he believe shares his values of what we would call today classical liberalism and to create a counter consensus to the one that’s gathering.

(00:58:53)
Now, Hayek also chooses not to argue against Keynes and he feels that this is a huge missed opportunity that he should have staked out the case against Keynes and that, because he did not, people come to believe there is no case against Keynes, Keynes is literally unanswerable. So, Hayek will have this great regret, he will channel some of his regrets into community building, specifically developing The Mont Pelerin Society, and it will fall to Friedman to really make that case against Keynes. But Hayek will end up at Chicago and Hayek really influences Friedman to think about what Hayek calls the competitive order and how the state can and must maintain a competitive order that is the system of laws, of norms, of practices that makes it possible for markets to function.

(00:59:51)
And this is one of these key differentiators between the older philosophy of laissez-faire and the newer reconceptualization of liberalism which says, yes, we need a state, we need a state that’s not intervening in markets under social democratic auspices but is structuring and supporting markets so that they can function with maximum freedom keeping in mind that, if there aren’t basic social supports needed, the market is apt to generate the type of either inequality or social instability that will call the whole system into question.

(01:00:28)
So, Hayek is really key in promoting this modified liberalism. But, from being a very prominent economist in the 1920s and 1930s as mathematics becomes the language of economics, Hayek is completely left out in the cold. Now, Friedman, to some degree, is left out in the cold but Friedman at least has proved the mathematical economists and he knows what they’re up to and he’s rejecting it from a position of expertise and knowledge and he literally drives the mathematical economists out of Chicago. They’re clustered in a group called the Cowles Commission and he makes their life hell, they flee, they flee the Friedman onslaught. But then when Hayek arrives at the University of Chicago, he would like to be considered for a position in the economics department and Friedman, Milton Friedman says, “No way, you’re not really an economist because you’re not empirical, because you just developed these theories.”

(01:01:26)
So, he has an appreciation for Hayek as a social thinker but not as an economist so what Friedman decides to do, his answer to Keynes will be deeply empirical but it will also be theoretical and it will create an alternative intellectual world and approach for economists who aren’t satisfied with Keynesianism. And almost single-handedly, Friedman will introduce political and ideological diversity into the field of economics because, from his beachhead in Chicago, he will develop the theory of monetarism. So, what is monetarism? The easy way to summarize it is this famous dictum of Milton Friedman’s, inflation is always and everywhere, a monetary phenomenon. And it’s fascinating that he becomes an expert in inflation because the first research and the first major research product of monetarism is that theory of the Great Depression in a monetary history of the United States and that is a theory of a deflation, all price is going down.

Money and monetarism


(01:02:33)
And he will go back to an idea that Irving Fisher had popularized but a very old idea, almost a truism, the Quantity Theory of Money which says the level of the price level is related to the amount of money circulating in an economy. So, if you have more money, prices go up, if you have less money, prices go down. Now, this seems very basic and almost too basic to bear repeating but Friedman is saying this very basic relationship holds true even in an advanced industrial economy and that is what people have started to doubt. And if you think about money, you think about banks, you don’t think necessarily about the federal budget spending and taxation.

(01:03:18)
And what you see happens in American economics, the textbooks previous to the Keynesian Revolution, they spent a lot of time on money, they spent a lot of time on interest rates. You can do word counts and other scholars have done the word counts and the word count for money after World War II just plummets and you start seeing things like taxation, budget, those things go up. So, what happens is the economics profession shifts its attention, it just looks away from money to other things and Friedman is one of the few who’s saying, no, money still matters, money still counts and it’s a very counterintuitive argument to make, it’s a very historical argument to make and this is absolutely fascinating to me.

(01:04:02)
With Anna Schwartz, he develops this 150-year timeframe, he also has students working on episodes of hyperinflation in different periods of time, he’s also looking back to ancient history, inflationary episodes there and he’s saying this is a law of economics. This is something that recurs throughout time, it’s not historical, it’s not contingent, it’s a law of economics. And his Keynesian counterpoints are saying, no, that’s not relevant any longer, maybe once it was relevant but it’s not relevant today. Now, in some ways, they have a point because, in order to pay for World War II, the federal government sells a lot of bonds, it issues a lot of debt and it wants to pay this debt back a low interest rate and it wants people to keep buying it, it wants the low interest rate to be competitive with other interest rates. So, wants, in general, low interest rates throughout the economy.

(01:05:03)
And the Federal Reserve has been so discredited by the Great Depression that the Treasury basically runs the Federal Reserve and says keep interest rates low. And so, that’s what it’s doing and so the Federal Reserve has stopped being an independent entity, it’s just a sub-department of the Treasury. But in 1951, they negotiate what’s called the Treasury Fed Accord and the Federal Reserve gets its independence but it doesn’t really use it but, statutorily, it now has it. And so, most economists are just observing a regime in which the Federal Reserve has no power, a regime in which there is really little inflation, the inflation that is seen is post … There’s a little burst of inflation in the Korean War. And they’re saying inflation’s not really important, it’s not really relevant and money’s not really relevant and important. And so, to break through and to make the argument, that’s why Friedman and Schwartz go to history and they’re able to make that argument for history.

(01:06:03)
So, then Friedman is coming out with a variety of papers that are saying, when I look at economic fluctuations, he maps them side by side to fluctuations in the money supply and says, look, they fit. And other economists, remember, they’re building complicated mathematical models and Friedman’s doing extremely simple stuff and they just think it’s dumb, it’s not interesting, it’s not true, they don’t buy it at all. But after A Monetary History of the United States, they have to pay attention. So, it’s really in those years Friedman is hammering this idea of monetarism and it starts to become something respectable, bordering on respectable for other economists to look to and think about and that’s really the beginning of the Keynesian Monetarist split where, if you start to give Friedman any credence, you’re heading towards a monetarist position.

(01:06:58)
Now, at the same time, Friedman comes out very publicly in 1964 as a supporter of Barry Goldwater and Keynesian economics has found a home in the Democratic Party. Its probably brightest moment in the sun is the administration of John F. Kennedy who brings in a lot of Harvard and Yale professors to the Council of Economic Advisors, he proposes a series of spending programs that are really guided by the Keynesian philosophy. And the Barry Goldwater is tremendously controversial, part for his votes against civil rights which Friedman really supports and part because he’s a hard core libertarian in an age when that’s not in the political mainstream or not discussed in the political mainstream and he’s just tremendously unpopular particularly in all the educated precincts where Friedman lives. So, Friedman is like an outcast and a pariah for his support of Goldwater.

(01:07:53)
And so, that actually really affects monetarism because people feel that this is now becoming a package deal and so there’s a great reluctance to embrace Friedman’s-
Jennifer Burns
(01:08:00)
And so there’s a great reluctance to embrace Friedman’s ideas because it seems like you would then have to embrace his politics.
Lex Fridman
(01:08:09)
So it’s associated with conservatism.
Jennifer Burns
(01:08:12)
So this is the years when conservatism… There is a movement that calls itself conservatism. And Friedman is very tightly allied with this movement from the beginning, partly through his friendship with William F. Buckley. And a lot of people say to me, yeah, but Friedman’s not conservative. And this is like a bigger… You have a whole separate podcast on this.

(01:08:33)
But for now, I’ll just say that conservative in the United States becomes a political brand that contains elements of conservatism that are recognizable across time and space, embrace of tradition or comfort with hierarchy, et cetera. And it also has something new and different, which is Friedman’s ideas about… Milton Friedman’s advocacy of more free markets, less government regulation, and the benefits of capitalism and the benefits of freedom.

(01:09:03)
And that gets folded into American conservatism, in part because Milton Friedman is such a powerful intellectual figure. And after his advocacy of Goldwater the media realizes this guy’s really smart. He has really interesting things to say. He makes great copy, he makes a great guest, and he starts writing a column for Newsweek magazine, which is a very big deal in a much more consolidated media environment. And he’s quoted in all the newspapers. And so his public profile really starts to rise right as he’s pushing monetarism as an alternative to the Keynesian synthesis.
Lex Fridman
(01:09:39)
Can we just linger on what is monetarism? Once again-
Jennifer Burns
(01:09:44)
Yes. Okay, I didn’t go into it.
Lex Fridman
(01:09:45)
Okay. The money supply.
Jennifer Burns
(01:09:47)
Yes.
Lex Fridman
(01:09:47)
So money is this thing that you can think of it as a notion where people buy and sell stuff, and there’s this fascinating complex dynamical system of people contracting with each other in this beautiful way. There’s so many pothead questions I want to ask here, about the nature of money. Money is fascinating in that way, and I think for Milton Friedman trusting the flow of money is really important and the signals that pricing and money in general provides is really important.
Jennifer Burns
(01:10:25)
Yeah. And I could take some of this back again to Frank Knight. So one thing Frank Knight said to all his students was the market is the best allocation mechanism we have. The market is what allocates resources. In a situation of scarcity the market allocates them the best. And Hayek will add to that by saying, “prices are information signals, and a price sends information to buyers and sellers about how they should act.” And these are two of the strongest arguments for why the government should not intervene in the price system because it will blur information or because it will allocate less efficiently than market allocation will.

(01:11:09)
And so what Friedman is really going to add to that is maybe going up a level and thinking in the macro about the whole economy and how money circulates through that economy as a whole. And so what he and Anna Schwartz do is they construct what are called monetary aggregates. This is adding together say all the money that’s on deposit in banks and all the money that’s believed to be circulating in people’s wallets.

(01:11:38)
And you also have to really go back in time. We don’t have credit cards. There is a stock market, but it’s tiny in terms of the number of people who invest. There aren’t mutual funds. When travelers checks are introduced, this is a big deal. So we have a very simple monetary system. And so Schwartz and Milton Friedman start measuring what they call the monetary aggregates. They focus on M1 and M2 and their favorite aggregate is M2, which I believe is encompassing sort of deposits and circulating medium.

(01:12:16)
The other thing to recall, there’s some fine distinctions between money in savings accounts and money in checking accounts, and money in savings accounts can earn interest and is generally believed not to circulate or money in checking accounts does not at that time bear interest and cannot legally bear interest. And so is thought of as circulating. And then there’s different institutional architectures of postal savings, banks and credit unions.

(01:12:48)
But Friedman is one, taking the focus to these aggregate amounts of money and saying, these really have a lot to do with economic booms and busts. When we have an expansion in the amount of available money, we see an expansion in economic activity. When we have a contraction in available money, we have a contraction. And so he says at this stage, the government through the mechanism of the Federal Reserve and its influence on interest rates can either make money more cheaply available and more freely available in the economy or can make money more expensive and slow things down.

(01:13:32)
But the central core idea of monetarism is this is potentially very bad if the government can hit the gas and then hit the brake, and hit the gas and hit the brake based on say what a politician wants or what somebody at the Federal Reserve wants. You have a lot of instability in the system. And so one of the core policy proposals of monetarism is let’s grow the money supply at a steady rate. And in the beginning, Friedman just says K%, he doesn’t even put a number on it because he says the number doesn’t matter. What matters is the steadiness in the growth rate. Because if it’s a steady growth rate, it will fade away. And then people will make economic decisions based on the fundamentals, not based on what they think is going to happen, not based on hedging against inflation or hedging against deflation. They’ll just be able to function.

(01:14:33)
So this is sort of the paradox of monetary policy. When it’s happening right you don’t see it, you don’t notice it. When it’s happening wrong, Friedman argues, it can just fundamentally destabilize everything. It can cause a great depression, can cause an artificial boom. And so he’s taking monetary policy at a time when most economists think it’s completely irrelevant and saying, this is the central game of the economy. Now we live in a world where we believe this and the Federal Reserve chair can’t open their mouth without headlines being generated. But Friedman is saying this at a time when the Federal Reserve is a mysterious and secretive organization. It’s not well-known, it’s not deeply appreciated. Some of the only people who appreciate the Fed’s power are hardcore rural populists, who have constituents who think the banks and money power are the problem, who are throwbacks from the frontier days.

(01:15:31)
So Friedman in the beginning has no constituency for this policy. He has no constituency for this analysis. And so just going back to summarize monetarism, it’s looking, it’s using the quantity theory of money to analyze the macroeconomy. It’s proposing a policy of slow and steady growth in the money supply. And then it is arguing that inflationary episodes when they emerge are profoundly driven by changes in the money supply, not by anything else.
Lex Fridman
(01:16:07)
And going even up a level as we started, how epic is it to develop this idea, to hold this idea and then to convince the United States of this idea that money matters, that today we believe is mostly correct for now? And so just this idea that goes against the experts and then eventually wins out and drives so much of the economy, the biggest, the most powerful economy in the world. So fascinating.

Stagflation

Jennifer Burns
(01:16:43)
Yeah. So that’s a fascinating story. And so what happens is Friedman has advanced all these ideas, he’s ruled the economics profession, he’s built a political profile and then he becomes the head of the American Economics Association. And he is asked in that role to give a presidential address. And so he gives his presidential address December 1967, and he says, “I’m going to talk about inflation and I’m going to talk about the trade-off between inflation and unemployment.” And this is what’s generally known as the Phillips Curve and the Phillips Curve in its original form is derived of post-World War II data. So it’s derived of about 12 years of data, and it shows that when inflation goes up, unemployment goes down. And the idea would make sense that as the economy’s heating up and lots of things are happening, more and more people are getting hired.

(01:17:41)
And so this relationship has led policymakers to think that sometimes inflation is good, and if you want to lower unemployment, you could let inflation go a little bit. And in the crude forms, it becomes to seem like a menu. Like you could take your model and you could plug in, I want this much unemployment. And it would say, well, great, this is how much inflation you should do. And so then you would target that inflation rate.

(01:18:10)
So Friedman gets up and he says, “this is wrong. This might work in the short term, but it’s not going to work in the long term.” Because in the long term inflation has… First of all, it has a momentum of its own. Once it gets going, it tends to build on itself. The accelerationist thesis, it accelerates. And once inflation gets going, and the reason it gets going is because workers go to the store and they see the price level has gone up, things have cost more. They ask for their wages to go up, then eventually the wages will go up too high and they will no longer be hireable or companies will decide at these high wages I can’t hire as many workers, I’d better lay off. So if inflation keeps going eventually over the long term it will result in high unemployment.

(01:19:02)
So he says, “theoretically, you could end up in a situation where you have high inflation and high unemployment. This hasn’t been seen,” but he says, “theoretically this could happen.” And then he goes and he says, “and the government has started expanding the money supply in 1966, so we’re going to get a bunch of inflation, and then we’re going to get a bunch of unemployment.” And he estimates about how long it will take. And then he says, “once this all happens, it will take about 20 years to get back to normal.” And-
Lex Fridman
(01:19:33)
And he predicts the stagflation of the 1970s.
Jennifer Burns
(01:19:37)
Stagflation of the 1970s.
Lex Fridman
(01:19:39)
[inaudible 01:19:39] for an economy. Again against the mainstream belief represented by the Phillips Curve.
Jennifer Burns
(01:19:47)
Yeah. And what really makes it happen is that many of the economists who most deeply dislike Friedman and most deeply dislike his politics in the 1970s, as they’re running their models, they start to say Friedman’s, right? They start to see in the data that he’s right.

(01:20:04)
And a very parallel process happens in Britain. Britain is going through a very similar burst of spending, burst of inflation. And so Friedman is vindicated in this very profound way, in the way that he himself said would be the ultimate vindication, which is, my theory should predict. So that prediction of stagflation is really the sort of final breakthrough of his ideas and also their importance to policy and to thinking about how we should intervene or not in the economy and what the role of the Federal Reserve is. Because he’s saying the Federal Reserve is incredibly powerful. And finally people start to believe him.
Lex Fridman
(01:20:43)
And I don’t know if we said, but to make clear stagflation means high unemployment and high inflation, which is a thing like you mentioned, was not seen before and he predicted accurately. And it also disproves the relationship, the inverse relationship between unemployment and inflation.
Jennifer Burns
(01:21:05)
Yeah. Now I should say the Phillips Curve is still out there. It’s been expectations augmented and it is relevant in the short term, but Friedman’s warning is still very much apt, that if you get too focused on unemployment you can let inflation out of the bag. And so until very recently, the Federal Reserve’s tradition has been focusing on inflation, believing that’s fundamental and that will keep unemployment low, rather than trying to lower unemployment at the cost of raising inflation.

Moral case for capitalism

Lex Fridman
(01:21:39)
Can we go back to Frank Knight and the big picture thing we started with, which is the justification of capitalism?
Jennifer Burns
(01:21:46)
Yes.
Lex Fridman
(01:21:46)
So as you mentioned, Milton Friedman searched for a moral justification of capitalism. Frank Knight was a big influence on Milton Friedman and including on this topic of understanding the moral justification of capitalism. I think you spoke about Knight’s Case for capitalism was grounded in the idea that the ability to act in the face of uncertainty creates profit. And it should, because taking risks should be rewarded. So this idea that taking risks in the face of uncertainty should create profit, and that becomes a justification, the ethics of capitalism. Can you just speak to that?
Jennifer Burns
(01:22:26)
Yeah. So Knight is talking about where does profit come from? And to his mind, it comes from the entrepreneurial function and the risk-taking function. And so he kind of weaves that into why capitalism works best and why it’s the most effective allocation machine and why it assigns responsibility in a way he believes that a socialist system never could.

(01:22:51)
Now, Knight though is not a booster of capitalism. It could be in part because he’s just a darkly pessimistic kind of depressive guy. And so he’s afraid capitalism is going to collapse and socialism or fascism is going to take over, or communism. And so he kind of descends into darkness there. Friedman as the more optimist believes with Hayek, that you can develop a different approach to capitalism that would preserve the price system, preserve allocation, but build in social supports, build in a social minimum, things like this.

(01:23:25)
But there’s a moment in his career where he’s really struggling to figure out, how do I make this case for capitalism? And basically the whole conservative movement or people who we later call the conservative movement are struggling to make this case. And he starts thinking about what makes capitalism work is that if you put forth effort, you get a reward. So then you could say, well, people get what they deserve under capitalism.

(01:23:47)
But then he kind of stops and he says, “that’s not really true, because we’re born with such different endowments and there’s a huge quotient of luck. So some people are just in the right position and some people aren’t. So if I say capitalism is moral because people get what they deserve, that’s not really true.” And he also kind of has an ethical reaction, which he ends up calling an aesthetic reaction. He’s kind of like, it just doesn’t feel right to say that. And so he struggles for a while with, what do I say? And then he basically says, “capitalism, it can’t be the core. Discipline of the market, can’t be the core to your ethics, it has to be something else.” So that’s when he will decide it’s freedom, it’s individual freedom. That’s really the ethical core. And capitalism makes individual freedom possible, because capitalism is dedicated to maximizing that.

(01:24:38)
And so the defense of capitalism comes through freedom. And at his stage in history, he’s able to set aside nice worry about inequality and say, when I look at the data, and this is true for the macro data at mid-century, incomes are actually converging. And also if you look historically, if a country goes from say, a more feudal, agrarian society to a more market-based society, incomes will converge. Now and then they might start to diverge. But Friedman’s in the moment when he’s seeing the convergence.

(01:25:11)
And so that’s what he’s really focused on. So he believes he can justify capitalism through the ethic of freedom. And he also believes that inequality is a problem that can be addressed through specific policies, and it’s not a fundamental feature of capitalism. In other words, he doesn’t see capitalism as an engine of inequality the way that Frank Knight did and the way that maybe some critics on the left would.

Freedom

Lex Fridman
(01:25:36)
How did he conceive of freedom? So individual freedom, economic freedom, political freedom, civil freedom, what was the tension, the dynamic between those different freedoms for him?
Jennifer Burns
(01:25:47)
So he really begins focusing on economic freedom. And he says it’s really important to focus on economic freedom because in the United States we don’t value it enough. By economic freedom he means the ability to keep what you’ve earned, the ability to make decisions about your business, the ability to make decisions about the work that you do. So this will translate into things like, there shouldn’t be a minimum wage. He believes the minimum wage has bad social effects, but he also believes you should be free to accept a job at a wage that you yourself have determined is acceptable to you. And there should be very minimal regulation questions around safety and other things, because the market will ultimately, if you create an unsafe product, it won’t sell. And that’s sort of your incentive.

(01:26:35)
So he really centers economic freedom because he thinks especially, and he’s really speaking from his vantage point in the universities and speaking to the kind of liberal consensus of the ’50s and ’60s, he thinks economic freedom has been undervalued in the American context. So he really wants to push that forward. He’s really kind of taking political freedom for granted.

(01:26:55)
Now later in his career when he becomes famous, he’s traveling the world, he spends time in Chile, and this country is now being ruled by a dictator, Augusto Pinochet, who starts introducing economic freedom, but there’s no political freedom. And Milton Friedman believes eventually these two things are going to go together. He tells Pinochet, “you’ve got economic freedom, and eventually it’s going to mean political freedom.” Pinochet is like, “okay, fine, not really interested in that. I want to know what I should do about inflation.”

(01:27:24)
But then when Milton Friedman leaves Chile, he is attacked and vilified for having been a supporter. It’s interpreted that he’s a supporter of the regime, which he’s not, but he realizes he has talked too much about economic freedom and he hasn’t talked enough about political freedom. He’s kind of assumed political freedom, because he’s come from the American context. So then he starts recalibrating them and saying, you know what? If you don’t have political freedom, you’re never going to be able to hold on to economic freedom. So he sees that they need to go together and they don’t naturally go together. And so he starts to become more clear in talking about political freedom.

(01:28:01)
Now let’s fast-forward to the end of his life, and he’s witnessing the emergence of what we call the Asian tiger. So capitalist economies that are doing very well, but they don’t have political freedom. But then he observes, you don’t have political freedom, in that you can’t vote in a free and fair election, but they also don’t have a Stasi, they don’t have a KGB. They’re not hauling people off for their wrong opinions. So then he says they have something called civic freedom. And so he kind of defines this third sphere, civic freedom, of debate, discussion, interpersonal relations, but you can’t be political.

(01:28:38)
So this is a late in life edition. I don’t think it’s fully theorized. I think what it shows is that during the Cold War, he very much believed economic and political freedom, capitalism and freedom, democracy, the United States, capitalism, this all went together. And he starts to see at the end of his life the emergence of different social systems that are using market trading and allocation, but aren’t giving people similar freedoms. And he’s kind of puzzling over that.

(01:29:07)
Now, he always believes that China will democratize, and he thinks China’s on the path to democratization, in part because Chile does democratize. Eventually Pinochet has voted out and it’s become a democratic capitalist and very prosperous country. And he thinks that’s exactly what’s happening in China. He sees Tiananmen and he doesn’t live long enough to get to where we are now in which doesn’t look like political or civic freedom is coming to China anytime soon.
Lex Fridman
(01:29:35)
And he did oppose the dual-track system of China, meaning the market is bottom-up, the government and China’s top-down and you can’t have both.
Jennifer Burns
(01:29:46)
He thought you couldn’t have both.
Lex Fridman
(01:29:48)
Yeah.
Jennifer Burns
(01:29:49)
He thought eventually the market would triumph.
Lex Fridman
(01:29:51)
Well, that’s a really powerful idea to say, okay, maybe there’s not political freedom, but just hold on to the economic freedom and eventually that’s going to give political freedom. Is that correct to say, start to work on the economic freedom and the political freedom piece will take care of itself?
Jennifer Burns
(01:30:10)
That’s what he believed. That’s what he believed, yeah. I think it’s more complicated than that. The people who gain out of a system of economic freedom could decide to collude in a system where there isn’t political freedom. That’s certainly a scenario. But again, that’s that core idea of freedom and that core belief that people want freedom and that people are drawn to freedom.

Ethics of competition

Lex Fridman
(01:30:34)
Just to go back to Frank Knight a little bit, he wrote an essay called The Ethics of Competition, he had the metaphor that economic life is a game. And then maybe that extends to society as a whole, like the entirety of it is a competitive game. And Milton Friedman, I think, adapted some of this, appreciated some of this. Can you speak to this metaphor?
Jennifer Burns
(01:30:55)
Yeah. I think what the metaphor of the game does is it asks you, okay, well what are the rules then? And let’s focus on the rules that keep the game going. So he didn’t use the concept of an infinite game, but I think that’s an interesting one, a game that all the players are in and keep going again and again and again.

(01:31:13)
And so that helped Knight along with Hayek shift from the allocation question, who’s getting what? Are things allocated fairly? To the more structural question of what are the rules of the game that we need to keep this system going? And so for a while that led to the discussion of monopoly. Well, we need rules against concentration or we need the rule of law. Everyone needs to be treated equally. People need to know what they’re up against. And then going back to monetarism, the core of monetarism is a rule. Friedman called it a monetary growth rule.

(01:31:55)
And so again, what keeps the economic game going is a rule about how much the money grows, that everybody knows. Nobody’s guessing, nobody’s changing the rules to help their side or to help the people they’re friendly with. We all know it’s there, it’s clear, it’s easy. And so that emphasis on rules I think really has a through line. It goes into Hayek’s competitive order, and then it goes into the monetary growth rule. And then today, monetary policy makes use of monetary policy rules. We have not abandoned discretion, but rules are used as a heuristic or a check, and those come out of freedman’s thinking.

(01:32:38)
And so it’s something… It’s really profound, and it was always counterposed to discretion, which Friedman worried would be subject to capture or political corruption. If you had discretion in policymaking or if you had discretion in these very big areas, then people would stop competing against each other in a market and they would turn their attention to getting control of the rules or the rule makers.
Lex Fridman
(01:33:07)
So if there’s clear transparent rules, then you’re free to play the game.
Jennifer Burns
(01:33:13)
Yes, exactly.
Lex Fridman
(01:33:15)
But then depending on the rules, the game can turn out the equilibrium that it arrives at might be different. So that speaks to the mechanism design, the design of the rules.
Jennifer Burns
(01:33:26)
Yeah. And that was again, to go back to the idea separating new liberalism or neoliberalism from classical liberalism was more of a focus on what are the rules that are needed? What is the competitive order that we want to set out? How do we design in social safeguards? How do we think about it?

(01:33:45)
And so that shift towards monetary policy and focusing on stable monetary growth, that becomes really important in the post-’70s era, is one of the basic rules of how capitalist economies should function. And it becomes really important because they see the example of say countries most notably in Latin America where monetary rules weren’t followed and different governments played politics with their currencies, and that created huge upheaval and huge social loss, economic loss, just economic disaster.

Win-win solutions

Lex Fridman
(01:34:21)
So my friend, she’s a poker player, philosopher of sorts, great human being, she has a podcast called Win-Win that everybody should listen to. And the whole purpose of the podcast and her whole way of being in spirit is to find win-win solutions. So do you think of economic life as having such win-win solutions? So being able to find rules where everybody wins, or is it always going to be zero-sum?
Jennifer Burns
(01:34:47)
I definitely believe in win-win, but with a big asterisk, you can have win-win, but it can feel like win-lose, which is, it’s not just are people getting more, it has a lot to do with do people feel they’re getting more, and do people feel they’re getting what’s fair and equal?

(01:35:06)
So you could have a situation, for instance, if you look at the history of going back to Chile, it has steady growth, steady income growth, steady diminution of inequality, and a high level of discontent within the society and a high level of belief that the society is corrupt and unfair. And that’s what matters, how people feel about it, how people perceive it, it matters. And we saw this recently, you can’t just come out with a bunch of statistics and tell people you’re winning in this game if they feel like they’re losing.

(01:35:44)
So that goes to all the non-rational factors and all the comparative factors that people have when they think about where they are vis-a-vis other people in society. So we’re just incredibly social creatures. We’re incredibly attuned to our status to rising and falling to where we sit vis-a-vis others. And so that absolutely has to be attended to. It can’t just be an economic analysis.

Corruption

Lex Fridman
(01:36:09)
That’s so interesting that the experience of the economy is different than the reality of the economy. On the topic of corruption, I think the reality of corruption versus the perception of corruption is really important in a lot of these nations. You take Ukraine for example, the perception of corruption has a big impact on the economy. You don’t want to invest. You’re very cautious as a business person. The reality of corruption can be way different than the actual perception. But if narratives stay cold, it’s a self-fulfilling prophecy, that it has a big effect on the psychology of the people involved. It’s interesting.
Jennifer Burns
(01:36:48)
Yeah. This goes back to Keynes’s analysis of the great depression. If people won’t invest, if they’re spooked, if the investing classes are spooked, you could be in real trouble. And in some ways, this simple analysis of the problem and proposal of a solution was enough to restore eventually the path to academic prosperity, right? That’s Franklin Roosevelt, “nothing to fear, but fear itself.” The sense of we know we have a future, we have optimism, then you believe in it.

(01:37:19)
And to go back to thinking about money, money works because we all believe in it. It’s a form of social trust, and it’s a form of belief and faith in our society and in the other people in it. And when that breaks down, the money system will break down as well.
Lex Fridman
(01:37:34)
Is there something that Milton Friedman said and thought about how to control the psychology of humans at scale?
Jennifer Burns
(01:37:42)
No. What’s interesting is he does talk, especially in his later work, he says, we have fiat currency. And this is an experiment. And we don’t know how it’s going to turn out and it’s turning out okay right now, but we’ve always had a commodity-based or backed currency of some form or another. And this is the first time. And so who really knows? So far so good.

(01:38:09)
And he also is very attuned, it’s interesting in his later writings when he’s thinking about this too, sure, I could design a monetary system that would be different, but when I look at history, I see that monetary systems have always say incorporated the role of the state. Because it’s so important to people. And so therefore, my theoretical designs really have to be tempered by what I’ve actually seen happen in history.

Government intervention

Lex Fridman
(01:38:33)
So maybe we could speak to this tension between how much government intervention is okay for Milton Friedman. So he was against minimum wage, but he was for guaranteed minimum income. Can you explain actually the difference between the two?
Jennifer Burns
(01:38:48)
Yeah. So this was one of the discoveries I made in my research. I found a paper from 1938. He wrote advocating what we would call today a universal basic income, a minimum income. And he basically sees this as part of the effort to create a new liberalism. And he basically says, we have advanced societies, we have prosperous societies. We have decided in keeping with our morals and our ethics that people should not be starving in an advanced society like this. The question is how are we going to make that happen?

(01:39:19)
And he ended up believing the best thing to do was to put a floor under everybody. And he said, you can get that based on your income. If you have a lot of income, you don’t get it. If you have a little income, you might get a little bit of it. If you have no income, you get enough of it. And he believed in the beginning, you should base that on what was required to buy food. That would be kind of an objective. You could objectively determine the nutrition and the price of food.

(01:39:47)
And so that for him… It’s important, he says it’s keeping with a liberal polity because it’s not intervening in the price system, it’s not intervening in economic relations. And it does not, in his view, require a bureaucracy to administer. It is not, in his view, require that you qualify for it by virtue of being in a protected class. You just get it as kind of part of your membership in this general citizenship body.

(01:40:15)
And so that to him was really different than a minimum wage because it did not interfere with the work bargain. His belief about minimum wages was specifically that it priced out unskilled labor. That what an unskilled laborer had to offer was a willingness to work for a very low wage. And if you set the minimum wage too high, businesses instead of hiring that higher-priced labor would not hire. Or we could think of today, they put in an electronic checkout or something like this, where you don’t actually need the labor.

(01:40:50)
So he really believed the minimum wage had that perverse incentive. Now, this is a live debate on what minimum wages do, and there seems to be a level at which you can set them that they can not have that perverse effect, and in fact can kind of create people with more spending money, that then powers the economy. So he had a very sort of clinical analysis of that rather than an empirical one or a really abstract analysis.

(01:41:16)
But the minimum income is fascinating because it seems very leftist to us. But what it is it’s purely individualistic. And it never really happened because it was so purely individualistic. Because American social policy typically identifies this group of people is deserving and we’ll give them benefits. So the classic example is soldiers, veterans. Another example is mothers raising dependent children. These people deserve money. The rest of you, you better go out and work. And so Friedman’s proposal, it really caught on in the ’60s. It ultimately went nowhere, but it was no litmus test, no income analysis, just we’re going to give you this much, everyone’s going to get this much. And he decided once mass…
Jennifer Burns
(01:42:00)
… this much. Everyone’s going to get this much. And he decided once mass taxation had come in, you could do it through taxes, and you could just rebate people who didn’t pay income taxes got a rebate that actually came to pass. It’s the earned income tax credit, and it’s considered extremely successful by policy analysts. It does what it’s supposed to do. It’s not that expensive. And so I see that as a paradigm of his thinking in that instead of creating a bureaucracy that does some form of redistribution or instead of trying to intervene in the market for labor or the market for something else, the market for housing, you provide a cash grant that people spend for themselves.

(01:42:42)
And so interestingly, that’s what happened in the emergency situation of COVID, right? That’s exactly what people did. They followed that model. We just get money out quick. And there’s a lot of discussions still about UBI is something that should be done. And I think it’s always going to be hard to pull off because I think Americans and their elected representatives don’t want to provide a universal benefit. They want to provide a targeted benefit because they believe there’s a moral component here. And Friedman advanced a policy that was really abstract and really it was devoid of judgment. It was pure and beautiful in that way but utterly impractical.
Lex Fridman
(01:43:23)
And it really focused on not interfering with the market and the signals that the market provides. He was really against price controls for the same kind of reason.
Jennifer Burns
(01:43:32)
Yeah, exactly. You could say, “Okay, but how does this not interfere with the market, right? If you provide people with a minimum income, won’t that change their incentives to work, et cetera?” There’s a big body of research on this. Most of it seems to show, that one, it’s way better than the current benefits cliff where you have to not work to get your benefits. And any incentive impact on working seems to be much lower than would be expected. But I’ll let the economists and the social scientists fight that one out and figure it out empirically. Hopefully, we should be able to.
Lex Fridman
(01:44:08)
Yeah, there’s been a bunch of studies. It’s interesting even just how you conduct studies like this, how you do these kinds of experiments, especially if you’re empirically minded because a lot of the studies I saw is they’re pretty small, so how do you make big conclusions about how to run the world, how to run the economies from such small studies? It’s all a fascinating experiment of ideas, and it’s also inspiring to see individuals and maybe small groups of individuals like the Chicago School of Economics to shake out what we believe and how we run the world.

Conservatism

Jennifer Burns
(01:44:51)
Yeah.
Lex Fridman
(01:44:51)
Inspiring. You call Milton Friedman, The Last Great Conservative, maybe to be a little bit controversial and make bold statements that get everybody excited, but what do you mean by that, and what makes it great conservative?
Jennifer Burns
(01:45:09)
So I was really thinking of that in terms of American political identities and particularly the 20th-century conservative movement, which people are always saying, “This isn’t conservatism.” And I’m saying, “Yes, in America, conservatism is different. It looks different. It feels different.” Conservatism in America builds in a big component of what we could call libertarianism, anti-government ideas and critics will say, but conservatism is about conserving institutions and practices, and it has a role for the state and an organic community. But in the United States, it’s always had since the 20th century also this anti-statist, let’s let the market rip. It’s not worry about what the market does to establish traditions. The market is our tradition. Capitalism is our tradition, so that was really synthesized.

(01:46:03)
Many people were there, but Friedman and the importance of his books Free to Choose Capitalism, and Freedom, the television series he did. All of these were core components of this American conservative synthesis as it evolved. And I really see that as having broken down. It is scattered into different pieces and we don’t know where they’re going to come back together again. But Friedman’s push for open global markets, unfettered free trade. It’s getting pushback on both the left and the right. That I think is just a major sign that both parties have turned away from this vision. I don’t know what they’ve turned to, but the way that Friedman brought these pieces together, I think that political moment has passed. So that’s what I was trying to talk about with the book title.

(01:46:57)
There’s another way though in which I think of him also as a conservative, which is that within the field of economics, he went back to this older idea, the quantity theory of money, and said, “This still has value. This can be applied in the modern day. It is something to teach us.” And he pushed back against this trend towards mathematicization, so he kept writing books. You can still pick up a Friedman book and read it, where’s lots of economics articles and output, it’s unreadable unless you’re in the field.

(01:47:26)
And so I think in that way, he was trying to conserve methodologically and intellectually the traditions of the field, the work that he and particularly Anna Schwartz did, that literal counting of things and deep analysis of data from the field that was the completely unfashionable and his time. Now, we’ve gone back to it with big data and with computers, but he helped bring that forward and preserve that tradition. So I think of him intellectually as a conservative, if you think of the mode of his thought. And so, what makes a great conservative is one who takes those older ideas and makes them fresh for a new time period. I think that’s exactly what he did.
Lex Fridman
(01:48:09)
You’ve also spoken about the fact that the times when he was out in public, there was more of an open battle of ideas where conservatism often had William F. Buckley. He had a more vibrant, deep debate over ideas where it seems less deep now.
Jennifer Burns
(01:48:36)
That is the thing that it’s hard, especially for the students I teach today, to be like. There were arguments about ideas, and conservatives won a bunch of them. And that happened in the ’70s and late 1960s to 1970s when one set of arguments was about economics. Like, “Okay, this idea of stimulating the economy by spending more. It has a downside. The downside’s called inflation, and the downside’s called too much regulation. And you’ve gone too far in bottling up the actual sources of economic growth and dynamism, and we have to let those free.”

(01:49:14)
In social policy, there was also a critique. The Great Society had all these ideas of ending poverty, and people came and analyzed them and said, “The programs aren’t helping. In some ways, you’ve actually created engines to trap people in poverty because you’ve given them a benefit and said, if they actually start to work, they lose the benefit. You’ve created all these perverse incentives,” and these ideas were fought out. They were empirical. They were controversial, and they were based on really deep research and really deep argumentation. And so it seems that era has passed. It seems we’re driven much more quickly by moods rather than thought-through ideas. Right now, it seems the ideas they follow the political mood and try to put together the underpinning of it where it really was the opposite for much of the 20th century.
Lex Fridman
(01:50:10)
It does seem like we lead with emotional turmoil and the ideas follow versus lead with the ideas and the emotion of the masses respond.
Jennifer Burns
(01:50:20)
Right, exactly. So, if we think of the evolution of conservatism, it was a whole set of ideas that was crafted, refined. The 1950s, 1960s, 1970s really found their emotional standard-bearer, translator, salesperson in Ronald Reagan, who incidentally had been following these ideas as they developed and had been honing his ability to express them and apply them politically. It’s a very opposite if we look at Trump as the political definer of the era. There’s a set of ideas, but it was more attitudes, impulses, vibes, and the ideas are coming after that, trying to figure out how they patch on. So it’s interesting to watch, to see that difference. And I hazard that a lot of it just has to do with the immediacy of the media environment we’re in, and it’s just power of the media messages to get out so fast.

Donald Trump

Lex Fridman
(01:51:17)
What do you think Milton Friedman would say about Donald Trump about him winning in 2024, and just in general, this political moment?
Jennifer Burns
(01:51:28)
I think he would love DOGE. I think that’s goes without saying.
Lex Fridman
(01:51:28)
He’s hooked on that part.
Jennifer Burns
(01:51:33)
I think he would focus on that part. I think he would really love it. He would be very alarmed by the idea of tariffs and very alarmed by the return to protectionism. I think he believed that part of what made the world peaceful in the second half of the 20th century, as opposed to during World War II, was the world was knit together more by trade. And that was the great hope that if people traded with each other, they wouldn’t fight. He was also a proponent of the free movement of capital. He would absolutely oppose this idea that Nippon Steel wasn’t allowed to invest in the United States. I think he would struggle, and he wholeheartedly embraced Reagan, and he worked to minimize the parts of the Reagan legacy he didn’t like. I think he would find it harder to embrace Trump because he’s not of that style, and he just had a different style. But I’m guessing he would’ve come around through the… I think he would just say, “Okay, we have a chance to reduce the size of government.”

(01:52:38)
At the same time, the spending plans of the Trump administration are not fiscally conservative in any way, and that was his concern, was not so much with debt but with the feeling that there’s no mechanism to stop the growth of government, that it just grows and grows and grows. And so he ended up believing even deficits aren’t so bad because they make politicians cautious, he thought, about continuing to spend, but I have to believe he would be concerned about the potential threats to the US currency’s position as the world’s reserve currency with increased levels of debt and spending.

(01:53:21)
He was concerned about low interest rates. He died, I think it’s 2004, 2006, but it was in the beginning, he didn’t see the zero low bound, but he saw low interest rates. And he said, “This isn’t necessarily good. Everyone’s talking about low interest rates as if they’re good, but there should be a price on capital. There should be a price on this. It shouldn’t be so low.” And so he had still the macro insights that I think are important.

Inflation

Lex Fridman
(01:53:52)
You wrote The Wall Street Journal essay titled How Inflation Ended Neoliberalism and Re-elected Trump, so can we weave that into this discussion in terms of inflation and Trump? What’s the main idea of the essay?
Jennifer Burns
(01:54:09)
So, the main idea is looking back and saying, “So today, we have been living in a world where people have been focused on monetary policy, steady monetary policy, free trade, reducing regulation. This is all called the neoliberal era.” And my argument was a lot of that arose was driven by inflation, so we have Milton Friedman predict inflation in 1967. It starts breaking out in the 1970s. Britain and the United States, and every institution was designed around stable prices. And once inflation broke out, prices were no longer stable. So, for example, tax rates weren’t inflation-adjusted. So if your income went up because of inflation, you might bump from a low tax rate to an extremely high tax rate, but you don’t actually have more money, on paper you have more money, but everything costs more, so you don’t actually have more money and your taxes have gone up. That kicks off the taxpayer revolt.

(01:55:14)
There’s a whole shift of American corporations towards focusing on financial investments because the tax breaks they used to get for depreciation for building new factories are not inflation-adjusted, so they no longer pay off in an inflationary environment. And then when Paul Volcker comes in early 1980s, and starts fighting inflation really pushes up interest rates to bring down inflation. And that completely reorders the banking sector because banks had statutory legal limits on the interest they could charge. And once general market interest rates exceeded that, it was proliferation of new financial forms to take advantage of that.

(01:55:58)
So my point was the era we live in was ushered in by inflation, and then everyone turned against all the formulations we had and said, “Well, these have hollowed out our industrial base. We’ve got too much immigration. We’ve got too much economic openness. We need to reshore. We need to focus. We need to turn against all these things. We need to spend more. We’ve disinvested.” And the net result of that turning away, I argued, is people forgot about inflation. They really forgot it could ever exist. And you had a whole set of theories on the left, modern monetary theory, that basically said, “We don’t really need to worry about inflation. We can spend what we want.” And lo and behold, inflation came back. And so my argument is, that has now opened the door to the presidency of Donald Trump, which is potentially a deeply transformative moment that will change the size and shape of government, that may change our foreign policy profoundly, that may change our immigration policy, that may change the demographics of our country, all of that, and my thesis is that that’s all been made possible by inflation.

(01:57:12)
And so the great mistake of the past years was to forget how fundamental inflation was to the rise of the last political order and to profoundly underestimate how much inflation would change the current political order. So I just think it’s one of these things… This is why I think you should study history because if you had studied history, you would be aware of this. And it’s so easy for people to forget, just like the banks forgot that interest rates could ever go up. They got so used to it, and it’s only a 10, 15-year thing, but to them, that seems like forever. So I really do believe what history teaches you to do is just have a much vaster scope in your vision and then take into account the possibilities of so many things happening that are different than what’s happening today. And so I just hope we don’t forget about inflation entirely, but here’s the thing, it is quite a strong chance that Trump’s policies will initiate even worse inflation, and then they will prove to be his undoing, so the ironies of inflation could be continuing.

DOGE

Lex Fridman
(01:58:21)
Like you said, Milton Friedman would be a big fan of DOGE, so if he was still here today and rolled with Elon Musk and Vivek, what advice would he give? What do you think he would focus on in terms of where to cut, how to cut, how to think about cutting?
Jennifer Burns
(01:58:39)
His signature policy move, I talk about this, is taking the price mechanism and trying to make that into the policy. And that seems obvious to us today, but in the era that he came in, so there would be rent controls, let’s take away rent controls, let’s let housing prices set themselves, or he was very against national parks, I actually think that national parks are good, so I hope that DOGE people don’t take this up, but that rather than an allocation to fund the national parks, they should be funded by the revenue that they bring in when people visit them. And so I think he was always looking to, let’s let prices make the decisions here, so I think that would be one of the key pieces.

(01:59:23)
The other thing I think he’d really be thinking about, he wrote about this a lot, about occupational licensure and barriers to entry. And he felt like one of the worst things that government does, and sometimes it’s private entities that do this, is create barriers to entry to protect industries and markets. So he talked about this in the case of the medical profession, which I think is actually not a good example because I think we all have a collective investment in having medical doctors be highly trained. But so, for instance, you could look at nail technicians or haircutting. There’s often these licensing requirements, or there’s a big kerfuffle. I think it’s DC passed a law that to run a childcare center you have to have a college degree. Well, what does that do? That disenfranchises a whole bunch of would-be entrepreneurs who don’t happen to have a college degree but probably could be really good at this particular business. So I think he would be saying, “Look out for where private interests have used the state to protect themselves and clear away those types of barriers and let competition through prices guide outcomes.”
Lex Fridman
(02:00:30)
So, open up for more competition and allow for more signals from the market to drive decisions, and so on-
Jennifer Burns
(02:00:38)
Yes.
Lex Fridman
(02:00:39)
… which would actually naturally lead to cutting a lot of the bureaucracy of government.
Jennifer Burns
(02:00:46)
I think the other thing he would probably be arguing for is, again, going back to the design of the minimum income or the negative income tax, that there’s a way he… Ultimately, he decided to run it through the tax system. The government’s already collecting this data, they already have your information, and they can just send the money out through the system rather than having a social bureaucracy where you have to come in person, you have to fill out forms, you have to document. Do you own a car? What’s your income? Who lives in the household? And his analysis of that, who that really benefited, was the bureaucracy that processed that paper, that implemented those norms, and that if you could pull that away, you could get help out where it was needed much quicker without having this drag of people doing unproductive work of administering these systems, so trying to cut administrative overhead. And what he didn’t have then, which we have now, is the technology that we have and the ability to send benefits out via smartphone or just to move so much faster and to handle information on a mass scale so much faster.
Lex Fridman
(02:01:58)
It’s painful but I think one of the big things you can do is just that, which is digitalize. I don’t know if that’s a word, but just convert everything into where the speed of signal can be instantaneous. So there’s no paperwork. It goes immediately, and then that means that the pricing signals and all these things are just immediately available to people.
Jennifer Burns
(02:02:26)
That seems to be the low-hanging fruit, government IT systems could be vastly improved among.
Lex Fridman
(02:02:33)
But that would result again with a lot of people getting fired. And I think somebody submitted a question for me saying, “What are your thoughts as a person who cares about compassion? What are your thoughts about government employees, which there’s a lot of that are going to be hurt by DOGE?” It is always a really difficult question. A lot of people get fired to make room for a new system that’s going to lead to a lot of pain.
Jennifer Burns
(02:03:05)
There is going to be a lot of pain. I don’t know what the solution is. I think that’s also part of why Friedman favored a minimum income. He talked about it being countercyclical, in other words, when things were really bad, the spending level on it would naturally go up. This is what economists today call an automatic stabilizer. And then, when it’s not needed, the cost of it goes down. Maybe there’s a way to sweeten it with honey and have people take buyouts or things like that. That would certainly be a way better way to go.

Javier Milei

Lex Fridman
(02:03:41)
I did a podcast with Javier Milei. He has consistently praised Milton Friedman and cited him as one of his inspirations. So what do you think Milton Friedman would say about what’s going on in Argentina and what Javier Milei’s trying to do in Argentina?
Jennifer Burns
(02:03:56)
Yeah, I think he would appreciate it. I think Milei is a much more of an Austrian-inspired thinker, but I think he definitely appreciates Friedman. And on the macro level, Friedman always understood it’s really painful to treat inflation, but the more you put it off, the harder it is. So I think he would be trying to get him as he’s doing to just message that short-term pain, long-term gain. I think he’d be very supportive. I think he’d be thrilled to see, also, that Milei is very good at explaining these abstract ideas and putting his policies in the framework of the bigger picture. That was really meaningful to Friedman. I don’t know how politically persuasive it is overall. Milei’s very intense. He doesn’t have the same gifts of salesmanship and setting people at ease that say someone like Ronald Reagan had, but it seems to be that’s what his country was calling for right now.
Lex Fridman
(02:04:56)
Yeah, he is more chainsaw-less, like warm blanket. Javier recollects this line from Milton Friedman. I don’t know if this is accurate, but “If you strive for equality over freedom, you often get neither. But if you strive for freedom, you often get both.” You think there’s truth to this?
Jennifer Burns
(02:05:16)
I think on the big picture, definitely. We’ve seen focusing too much on equality can be… Because equality is such an alluring word, it can lead you to downgrade all kinds of other things that are really important. But I really think it depends on how you’re defining freedom. The statement is too big and too broad, right? So, if by freedom you mean not having to pay taxes if you’re successful, that can have all kinds of knock-on effects. The idea that people are able to prosper when they’re educated. Where is education going to come from? How is that going to be paid for and supported? And again, to go back to Knight, if you’re generating too much inequality or people are feeling that you’re generating too much inequality, sometimes they value that more than they value freedom, and so I think there has to be more of a balance, and it’s hard to make such global statements. You have to break them down into what actually do you mean.

(02:06:27)
But again, Milei is coming from a very different context, a very different country that has seen so much upheaval, so much government intervention, so much inflation, so much political turmoil. He’s probably thinking about it differently than Friedman was thinking about it.
Lex Fridman
(02:06:44)
Yeah, there was, there probably still is, a real threat of hyperinflation. There seems to be a very high level of corruption or the capacity for corruption, so it’s a really messy situation. So Javier Milei likes to recollect this great line from Milton Friedman that, “If you strive for equality over freedom, you often get neither. But if you strive for freedom, you often get both.” Do you think there’s a truth to this?
Jennifer Burns
(02:07:13)
Yeah, in the macro, for sure. We’ve seen if you really put equality as your goal, it’s such a seductive ideal, and people believe in it so much that they can carry out horrible crimes in the name of equality. But then, focusing on freedom, these words are too big, they’re so hard to define. And so I think you have to ask, “What is the freedom you’re talking about?” Right? If you’re talking about the freedom of ordinary people to be entrepreneurial, to make their own way, to start new things, to continue what they’re doing, to keep what they’ve earned, for sure, that can increase the equality overall.

(02:07:52)
If you’re talking about lower taxes, if freedom is just a code for lower taxes, there has to be… Lower taxes, in general, great, but if you’re one of the top generators of wealth, there has to be some way to ensure that, say, education, right? People prosper when they’re well-educated, that’s when economies do better. Education is generally state-funded, and you need some way to support that and provide for those institutions that structure society that make competition possible. So, I think it’s just a really broad statement, but again, Milei is coming from a really different context. He’s coming from the South American context, from such upheaval, such economic devastation in pursuit of the goal of equality that I think trying to rebalance with that emphasis on freedom, I definitely see where he is coming from.

Richard Nixon

Lex Fridman
(02:08:46)
If we can pivot a little bit, we’ve talked about Reagan. What are some interesting stories about how Milton Friedman navigated the Reagan and maybe even the Nixon administrations and how he was able to gain influence?
Jennifer Burns
(02:09:00)
Well, the Nixon administration is an interesting case because… So, I’ve been talking about inflation and the different consequences it had. One consequence it had is that it began to undermine the Bretton Woods currency system that was established in the wake of World War II. Now, Bretton Woods, what it did, basically, it ended up inadvertently putting the US dollar at the center of the world economic system. But under Bretton Woods, countries of the industrialized West agreed to trade their currency in set ratios that government set, so a franc was worth so many dollars or a German mark was worth so many francs, and then also under this system, countries could come to the United States, and they could trade the dollars that they held for gold because the US was on a modified gold standard. There was a ratio of gold to paper money. And so the system was set up and very quickly, most countries, the dollar was at the heart of it in that the converting into and out of dollars was really the mechanism of trade for many of these countries.

(02:10:08)
So Friedman said, “What we should have is floating exchange rates.” This is an idea, again, of instead of having a top-down design of policy, an administered policy, we will have policies set by prices, and you should be able to trade currencies on an open market, they should trade, and they should fluctuate, and that would be fine. Totally outlandish idea, but he was pinpointing the fact that Bretton Woods had an instability, and that instability began to emerge in the time of inflation. So you have more and more dollars being printed, they’re worth less and less. If European nations keep trading their currency for dollars, they’re going to be importing inflation into their own economies. So they say, “We don’t want these dollars. We’d like some gold instead.” And they have the right to go to the treasury, send in an order, and get gold out. And so they start doing this more and more, and it becomes… It’s called the gold drain, and the United States starts running out of gold.

(02:11:16)
They’re aware this is happening through the ’60s. They’re trying various things to fix it. And when Nixon comes into office in ’68, Friedman sends him a memo, and it says, “This is going to be a real problem.” He says something like, “This is a running sore, and you have to lance it right away,” some very graphic-
Lex Fridman
(02:11:42)
Very nice.
Jennifer Burns
(02:11:44)
… metaphor, “otherwise it’s going to explode.” Nixon just files the memo away. Nixon loved people to think he was influenced by and following the wisdom of Milton Friedman, but he didn’t actually want to do that. He just wanted the political benefit that came from it. So then comes the moment where the US Treasury Department realizes we are going to run out of gold. What should we do? And everybody decamps to Camp David, and Nixon decides, we’re just going to stop redeeming currency for gold. It’s called slamming the gold window shut. Done. And he also, at that same meeting, decides to institute price controls. He does a whole bunch of stuff. It’s an emergency. He calls it the New Economic Plan, which is an unconscious echo of the Soviet New Economic Plan, so a problematic name, a problematic policy. And Friedman is livid at the price controls, but he’s like, “Actually, it’s great that you close the gold window. Let’s go all the way to floating exchange rates.”

(02:12:51)
And this idea was heresy within the Treasury Department. Everyone’s very committed to the idea of the gold standard convertibility possibility of the United States at the court, the financial system kind of hem and haw. But at this point, Friedman has a very close relationship with George Shultz, and George Shultz is a high-level appointee who will eventually, over the course of the Nixon administration, become the Treasury Secretary.

(02:13:17)
And so Friedman is feeding Shultz all his ideas about how we should move to floating exchange rates, how we shouldn’t try to reconstruct Bretton Woods and the people in Treasury… It’s funny because I read some of their accounts, and actually Paul Volcker is in the Treasury Department at this time, and he can sense that Friedman is in here somewhere, feeding his boss ideas. He doesn’t quite know. And in the oral history, Shultz talks about this quite a bit, so at any rate, Friedman exerts this behind-the-scenes influence, and what Shultz does is just lets Bretton Woods fade away. He doesn’t make grand pronouncements. It just slowly the world shifts to a regime of… For a while, it was a regime of steady prices, and then they call it a steady regime of changing prices, or whatever. The language changes, the reality changes, and they end up where they are, so that’s a real measure of Friedman’s influence.

(02:14:15)
If there had been another economist in Shultz’s ear that said, “No, catastrophe is imminent. We have to go back to Bretton Woods,” he probably would’ve worked harder. The US government would’ve worked harder. And so that becomes one of these pieces of globalization. And what people don’t realize is there used to be, in addition to these floating set capital ratios, you couldn’t bring capital in and out of different countries. You had to register. You couldn’t invest. Where all these rules and strictures and the falling of Bretton Woods really blows that all open. It’s a precursor to globalization, so Friedman is right there.

(02:14:49)
Now, he’s very ambivalent about Nixon. He sees that Nixon is not an honest person. He thinks he’s very intelligent, and Nixon’s dream is to create a new centrist majority. So, he does many things to go back on his supposed economic principles and ideals. So Friedman does not like this. He doesn’t like the price controls. He’s in communication with his old mentor, Arthur Burns, who’s now the Chair of the Federal Reserve. And Burns is basically doing everything wrong in monetary policy. And I describe this in the book in some detail, these anguished letters back and forth, and basically, as I see it, Burns doesn’t have a solid theory of inflation, and the more Friedman pushes him, it’s almost like Burns is willfully ignoring Friedman and doing the opposite of what Friedman says, so Burns is running a very loose monetary policy.

(02:15:42)
Inflation is quite considerable over the ’70s. We were all spooked by… What did it get to? 6%, something like that. Recently for a very short time, this is inflation going over 10%, hovering at 8% for basically the whole decade of the ’70s, going up and down but with extremely elevated rates. And so, the Carter presidency largely follows foreign policies. A big part of…

Ronald Reagan

Jennifer Burns
(02:16:00)
The Carter presidency largely falls. Foreign policy is a big part of it, but the failure to tame inflation is part of it. And then Reagan comes in, and now Reagan loves Friedman and Friedman loves Reagan, very mutual feeling. The Reagan administration creates an advisory economic board. Friedman’s on it. He’s retired now. He’s entering golden years, but he really has Reagan’s ear. And here what he does is he convinces Reagan of his theory of inflation, which is inflation has been caused. It’s a monetary phenomenon that has been caused by bad monetary policy. Inflation has an accelerating dynamic. The only way to end inflation is by really showing and signaling that government policy has changed. And when you do that, it’s very painful for a short amount of time, people will suffer, but then you will come out on the other side into stable prices, and this is what you need for economic prosperity.

(02:16:58)
So the man who implements this policy, Paul Volcker, he’s definitely influenced by Friedman, buys the big picture of Friedman. He even buys Friedman’s specific technique of the monetary growth rule and of the focus on monetary aggregates, which Friedman has said, “Money matters, aggregates matter, and that’s what money is.” Pretty quickly Volcker finds that because of inflation and the financial deregulation in response to it, the aggregates don’t work the way Friedman said they would. And so the specific policy Friedman recommends, Volcker tries it for a year or so, doesn’t work super well. But what does work is letting interest rates go high, go above inflation, to a point where both the general citizenry and the financial markets believe like, oh, they’re actually serious about inflation. And because we’ve had a decade of inflation with all these presidents saying, Ford, “We’re going to whip inflation now,” that monetary policy has lost credibility. This is why people focus so much on credibility today, because once it’s lost, it’s really hard to get it back. And one way Volcker gets it back is interest rates over 20%. Unemployment very high, as high as 25% in construction sectors. And as this is happening, Milton Friedman is whispering in Reagan’s ear, “This is the right thing. Stay the course. This is going to work.” Now, interestingly, he hates Volcker or Volcker hates him, and Friedman will never give Volcker credit for this policy, but he will give Reagan credit for this policy. But he owes credit himself for keeping Reagan from wobbling on this policy and just pushing it through. And he also tells Reagan, very pragmatically, “You better do this now. You’ve got a four-year term. Do this in the first two years of your term. Things will have turned around by 1984 when you run for reelection and you’ll benefit from it.” And that’s absolutely what happens.

Cryptocurrency

Lex Fridman
(02:18:56)
If we could take a small tangent, a question I have to ask about, since we mentioned Bretton Woods and maybe the gold standard, maybe just have a general discussion about this whole space of ideas, there’s a lot of people today that care about cryptocurrency. What do you think that Milton Friedman would say about cryptocurrency and what role crypto might play in the economy, whether he would be for this idea against this idea, and if we could look at it for today, and also just 10, 100 years from now?
Jennifer Burns
(02:19:34)
There’s a clip, I think it’s in 1992, where people say, “Oh, Friedman predicted cryptocurrencies,” because he’s talking about how payments will eventually be electronic. So in some ways definitely, as he was looking at the computer and money, he knew these would come together in some way. I think he probably would see a use case for crypto. He definitely would not buy the stronger forms, I think of crypto ideology in which we could be heading towards a future in which there’s many different currencies that compete or that are distributed or there’s a stateless currency. And he addresses this very, very clearly because Hayek’s Denationalization of Money, it’s a paper in the late ’70s. Hayek argues for this kind of competing currency model or regime. And so he’s responding to that. He’s responding to people writing about free banking, and he basically says, “Look, even if you developed a variety of competing currencies, eventually society would converge on one.”

(02:20:33)
And that’s because people just want one currency that they know. They don’t want a bunch of different options. Even in places where there have been options to do that, they’ve been used very minimally. And then he says, “Secondly, the state always steps in.” He says, “Technically, theoretically, it doesn’t have to. I could draw you a model. I could tell you about how it could work without the state. But in actual reality, all human societies, through time and space, the state eventually becomes involved in the provision of money because it has so many knock-on effects to so many people.” So sure, I think he would, again, find a use case for crypto, think it’s interesting, but I don’t think he would see it as this is going to displace state money and we’re going to have a variety of distributed currencies.

(02:21:21)
The other thing he really stresses is that a change in a monetary system, it only happens amid great, great crisis. So again, you see in countries where the state is not controlling the money well, that’s when people are more turning to crypto. But he says, because money is so fundamental, there is going to be so much political pressure on any country that gets the currency profoundly wrong that the government will fall and another one will replace it. So if you look at episodes of hyperinflation, they don’t go on very long because they’re so upsetting to people.
Lex Fridman
(02:21:57)
If we can go back in time, we’ve talked about it a bunch, but it’s still a fascinating time, the Great Depression. University of Chicago, there’s these folks like Jacob Viner, Frank Knight, Henry Simons, all of these influenced the thinking of Milton Friedman. There’s this Room Seven situation at the University of Chicago. Just going back there, even just speaking almost philosophically, what does it take to explore ideas together, deliberate, argue in that space, and maybe there might be interesting stories about that time. It would just be interesting to understand how somebody like Milton Friedman forms. The seed is planted and the flower blooms.
Jennifer Burns
(02:22:43)
Yeah. Yeah. So he gets to University of Chicago, he makes fast friends, and in his third and fourth year, they become what I call them, the Room Seven gang. So Room Seven is they find an old store room in the basement, they take it over, and that’s where they have their jam sessions. And what made this world come together was Frank Knight. There was a charismatic leader and there were a bunch of acolytes who clustered around him. That I think was a key piece of the ingredient. And then there was a sense that they were onto something that the rest of the economics field had forgotten or was rejecting. So there was that sense of mission. So that seems to have been, there was a formal education piece, and then there was a parallel education piece rooted in admiration for a thinker, a shared admiration. And then what that led Friedman to do, what I found, syllabi that he had from non-economics courses, lists of books.

(02:23:43)
And he’d written the prices of different ones he wanted to read. So he had John Stuart Mill, On Liberty, like 50 cents written in the margin. So he began to educate himself. He gave himself a parallel curriculum alongside this very formal economics curriculum. He started reading the traditions of political liberalism and then talking them through with friends and then developing a shared sense of mission. And the incredible thing is, of those friends in the group, they scattered for like 10 years, and then they all came back together. George Stigler, his great friend, was hired at Chicago. Aaron Director, who was his wife’s brother, was at Chicago. So many of these people continued. He became Frank Knight’s colleague. So that was the base. That was what really grew him, that really profound peer group. Now, the other piece I talk about a lot is Friedman was a collaborator, an open-minded collaborator, and he had incredible connections with economists who were women.

(02:24:43)
And he basically found first in the figure of Anna Schwartz, later in the figure of this group of women who were his wife’s friends, this untapped pool of talent. And so he immersed himself in this whole other world of consumption economics and that resulted in his more technical work on a theory of the consumption function, which is the theory of permanent income. So for Friedman, intellectual work and intellectual production was always done in this very social context, in a context that blended friendship and intellectual partnership. And he only had a handful of friends who were not also economists interested in the same questions he was. So he just lived and breathed ideas all day long.
Lex Fridman
(02:25:30)
Can you speak to the jam sessions? What do we know about the jam sessions? What are we talking about here? You’re sitting in a room, are they analyzing? Are they reading papers and discussing papers> or are they arguing more over beers kind of situation?
Jennifer Burns
(02:25:42)
Yeah, more arguing over beers. And in this case, there’s several people who say it was all about Frank Knight. What did he say? What did he mean when he said it? Is he right? And so Knight, he would say one thing and then say another. If you read him, it’s very hard to follow what he’s actually saying because he’s full of qualifications and ironies. It blends. And so he would throw out these pieces and then the students would clutch at them, and then they would come back together and try to assemble this worldview. And then Frank Knight fell into this terrible depression, and to cheer him up, they planned a big party and they went back through all of his previous writings and they assembled them into a book that was published. This is the Ethics of Competition, and you can read the introduction written in part by Milton Friedman.

(02:26:29)
So not only were they talking about Knight and what he said, but then they started pouring over his work. One of them described it as a general equilibrium system where you had to know all the parts and then all of a sudden it all fit together in a whole. So if we step back, what they were doing was getting inside the mind of a great thinker and understanding the ways it all fit together and then testing their ideas against Knight’s. And what’s fascinating is, one of the first papers that Friedman publishes in statistics is a rebuttal of Frank Knight. He publishes a rebuttal of Frank Knight’s ideas about risk and uncertainty. And Frank Knight, he took a black swan argument. He said, “Risk, you can calculate. Uncertainty, you can’t existentially philosophically. You can’t get your hands around it. It is the Black Swan.”

(02:27:22)
And Friedman publishes this statistical paper and he says, “I can put uncertainty on a graph.” And so there’s that Freudian killing of the father element when he comes back and he will in some ways turn his back on Knight’s approach and Knight’s pessimism even while it’s a foundation of his thinking.
Lex Fridman
(02:27:42)
Fascinating. Is there something you could say about the thinking process that Milton Friedman followed, how he developed his ideas? You mentioned there’s a strong collaborative component, but there’s another story I saw about that I think his son recalled about the argument number system that you mentioned, which, by the way, if you can explain that as a tangent of a tangent, that’s really awesome. I think it’s number one if the other person is right.
Jennifer Burns
(02:28:14)
Number two means you were right and I was wrong. And the number system evolved in some ways to be quick and efficient, but in other ways they also were really clear about it. So something like there’s three reasons behind it. First is, if you use a number, it reminds the listener that it’s really hard to say the words, “I was wrong.” So you’re calling on their sympathy by using the number, reminding them that you’re doing a hard thing. And then it’s also reminding them that you’re in this family with this code, and so you’re signaling your membership and your closeness and your love, really. It’s supposed to be an easy way to disagree without breaking the relationship.
Lex Fridman
(02:29:01)
So admitting you’re wrong now comes with this warm fuzzy feeling?
Jennifer Burns
(02:29:07)
Yeah, yeah.
Lex Fridman
(02:29:07)
And, really, I mean, that’s so powerful. I think so much of the friction of human interaction could be boiled down to just not being able to admit they are wrong efficiently and quickly and regularly and just often. And to be able to do that, that’s really powerful.
Jennifer Burns
(02:29:27)
I think it is a really neat aspect of their family life, for sure.
Lex Fridman
(02:29:31)
That’s a fun story, but can we just generalize to how he engaged in collaboration, how he developed his ideas? Is there a thinking process?
Jennifer Burns
(02:29:41)
So he taught at the University of Chicago, and he tended to teach for six months and then have six months off. And he spent the summers in New Hampshire or Vermont, right near that border. They had two different houses. And that to him was the deep thinking time. And so when he’s at Chicago, he’s teaching, he’s arguing. Some people love his teaching style, very much in charge, very much keeping students on their toes, confrontational. Others found it too much, overwhelming, shut them down intellectually and they couldn’t cope with it. And so I think it was go time when he was teaching. In that case, that was a lot of social time interacting, talking other professors, going out and giving papers, arguing with the people at Yale or Harvard. Then he would go and do these very deep dives over the summer.

(02:30:31)
He would also regularly do these trips to New York to see Anna Schwartz, his 12-year collaborator. Phone calls were really expensive. They did have quite an extensive correspondence, but then they would do these meetings. So he would basically come in at the beginning of the summer going to Rahway, stop in New York, see Schwartz, and then again on the way back to Chicago. So he’d have these deep check-ins at that point. The other thing that happened is people would come visit him in New Hampshire. He had his studio separate from the house, he would go and he would work, and then at night his friends would come. His friends were all economists. There’s a whole cluster of economists. They all clustered within driving distance of the Dartmouth Library so that they could get their hands on books. And so they would come over and then they would argue and talk into the night. So I think he did need that deep focus time, but he also lived a very engaged, very embedded social life.
Lex Fridman
(02:31:28)
A part of which was this marriage. Is there something you could say about love, about marriage, about relationship that made the whole thing work, it was vibrant, and they wrote a biography together?
Jennifer Burns
(02:31:40)
They did. I mean, they were very complimentary. They were the yin and the yang. She was very introverted, somewhat suspicious of others, skeptical, and he was extremely extroverted, optimistic, high energy. And they also were at a time when it was really clear. For broader society, these are the roles of a man, these are the roles of a woman. And they pretty much adopted those. Now, Rose Friedman did some very important economic work. She’s part of the early stages of the theory of the consumption function. She didn’t complete her degree because she really knew if she wanted to be married and have children in the world she lived in, there wasn’t a real pathway to also being an economist. I do think that a lot of that, although it feels very gendered, he’s the man out in the world and she’s in private, it’s interesting because her brother, Aaron Director, was the same way.

(02:32:36)
He was a very private man, very shy, very introverted, and he exerted this quiet intellectual influence on all of his friends. So I think that was just a family trait of being more quiet, preferring to be behind the scenes. It wouldn’t have worked any other way because Friedman was so out there, so extroverted. And there’s a bit of a sad thing she said, she said, “When I married Milton, I lost half of my conversations. When David came along, I lost the other half.” So this is a household that was just dominated by male voices in which she didn’t have a lot of room. What was tricky for me in my research is she didn’t leave much of a trace. She put together Milton Friedman’s archive and she took herself out of it.

(02:33:21)
I really had trouble finding her actual voice in the historical documents, and she didn’t want to leave that behind, so this absolute essential piece of his success. Because she’s the one who pushed him to do the Newsweek column to do Free To Choose. And she really wrote Capitalism and Freedom. She took all his random notes and she put them together into a book. And that became this testimony of his ideas. But she shared many of his ideas and when I think of Friedman, if you take away Anna Schwartz, if you take away Rose Friedman, if you take away the other woman who collaborated with him, you have a much thinner resume than the one he actually has.

Ayn Rand

Lex Fridman
(02:34:01)
Yeah, it’s always sad, and it always makes me wonder about the private, secret conversations between partners because they might not show up in the record, but they probably influence the person more than almost anything else, those quiet little conversations. If we can switch our path to another great mind of the 20th century, Ayn Rand, we talked about some of the similarities here, about them being fighters for freedom and fighters for capitalism. What is Ayn Rand’s philosophy? If you can give a big 10 summary of objectivism?
Jennifer Burns
(02:34:47)
Yeah, so she called it objectivism. She used to do this thing, “I can stand on one foot and say it.” It goes something like, “Epistemology, reason, ethics, selfishness, politics, capitalism.” That was how she summarized it. So what she did, there’s a couple of things she did with objectivism. First of all, she says the key defining element of humanity is rationalism, the rational faculty. That’s what defines what humanity is. Therefore, there is an objective reality that we can access a know with our reason. That’s the objective epistemology. And the one social and economic system that lets rationality flower and is based upon rationality is capitalism. And then rationality only works in her view as an individual capacity and that rationality teaches that what you should do is pursue your interests. And so she ends up calling that selfishness.

(02:35:50)
Now, it’s tricky because selfishness has so many strong and negative connotations, and she meant I think something closer to self-actualization, because she really tried to create this idea and express the idea that to be truly selfish did not mean trampling on others. It meant just being motivated by your own internal measures and metrics. And so in her fiction, she tries to show this by showing the false selfishness of some of Peter Keating, who’s an architect who steps over everybody to advance his career. And she says, “That’s not true selfishness because true selfishness would recognize it’s false to take others’ work and pass it off as your own.”

(02:36:43)
Now, the other big piece of objectivism is a very approach that’s really inspired and related to Friedrich Nietzsche’s idea of revaluing values or a genealogy of morals. And so she says, ” What’s happened here is Western culture has converged on this idea of altruism as good. Being selfless and altruistic is good. And this has led us to communism and has led us to devalue the individual in favor of the collective. So what we need is a new moral code which elevates selfishness, which elevates the individual, and which takes all the things that we have been told are bad and actually says their values.” This is what she’s trying to do with objectivism. I mean, it is about as ambitious of an intellectual project as there can be, and that’s what really draws people in.

(02:37:40)
Yet at the same time, she’s flying in the face of the way human morals and ethics and societies have evolved, and she’s not able to single-handedly recreate them the way she wants them to be.
Lex Fridman
(02:37:52)
Yeah, I mean, she’s not doing herself any favors by taking on the words and trying to rebrand them completely writing the virtue of selfishness. It’s like, can we just call it a self-actualization?
Jennifer Burns
(02:38:04)
Yeah.
Lex Fridman
(02:38:05)
There’s a negative connotation to selfishness and a positive connotation to altruism. So sometimes it seems it takes on the hardest possible form of argument.
Jennifer Burns
(02:38:18)
Yeah, I mean, she had a student who ended up being very close to her, Nathaniel Brandon, and he was the reverend advisor, and he said, “Can you please not use selfishness? Just come up with another word.” But part of her liked it. Part of her wanted to provoke and unsettle. She didn’t want to give that up.
Lex Fridman
(02:38:35)
I mean, people should listen to her public talks. Her whole aura, way of being is provocative, and she’s a real powerhouse of an intellectual. So she loves the challenge and just listening to her in itself is just inspiring. You could see the individualism radiate from her.
Jennifer Burns
(02:39:01)
Yeah, I mean, that was one of the things I found in researching and writing about her. She’s an incredibly unusual human being. That was her strength because she’s so unusual, but it was also her downfall because she looked to herself as a model or to get insight about humanity, and she never quite processed how different she was from other people.
Lex Fridman
(02:39:24)
So just because we talked about Milton Friedman so much, can we just return to, what do you, given everything we’ve said, is the interesting difference about Ayn Rand, her ideas related to Milton Friedman?
Jennifer Burns
(02:39:44)
Yeah, I mean, broadly we could put Milton Friedman and Ayn Rand in some category together, but she has this focus on ethics and rationality and this desire to be revolutionary that’s much stronger than Friedman. Friedman wanted to overthrow the economic consensus. He didn’t want to overturn the moral basis of Western society. Also, she does something. So in one of Frank Knight’s essays, he talks about the ethics of competition, and he says, “You basically cannot build an ethics out of competition because it would be monstrous to do so because it would say the winner of this competition is ethically right, and that would open the door to might makes right.” And this is what Friedman struggles with. And he says, “I can’t take capitalist outcomes as ethical unto themselves. I can’t do it. It doesn’t feel right.” And there’s this line where Frank Knight says, “No one would ever do this.” And I was like, “Oh, Frank Knight, you haven’t read Ayn Rand yet.”
Lex Fridman
(02:40:47)
Hold me beer.
Jennifer Burns
(02:40:48)
You’re a little too early. Because that’s what she does, she takes the outcomes of capitalism and of market competition and says, “These have ethical meaning, and this is where ethical meaning inheres, and it is ethical to try to succeed and to succeed in a capitalist society.” Now, what she’s able to do is create a fictional world in which people succeed in her fictional capitalist world through ethical behavior. And so she doesn’t really have to wrestle with a capitalist world in which people succeed through fraud and corruption and all the other things that might go into someone’s success. She creates the best possible take on success under capitalism, and then she holds that up as an ideal. And I think what’s important is that so few people have done that, and she comes at a time when everybody is emphasizing a downsize of capitalism, and she says, “There’s another way to look at it. Here are the good sides of capitalism.”
Lex Fridman
(02:41:42)
And like you said, she was operating, which I really loved the phrasing of that, in the Mythic Register.
Jennifer Burns
(02:41:49)
Yeah.
Lex Fridman
(02:41:49)
So she was constructing these characters, these capitalists that are the highest form, these great heroic figures, almost romanticizing them.

The Fountainhead

Jennifer Burns
(02:42:01)
Yeah, yeah.
Lex Fridman
(02:42:01)
You mentioned We The Living is one of the books that you like of hers the most, but can we stay in the Mythic Register with the Fountainhead and Atlas Rugged? What are some memorable, inspiring moments, insightful moments from those books that may be scenes or ideas that you take away from them that are important for people to understand?
Jennifer Burns
(02:42:30)
Yeah, so the Fountainhead is this story of a struggling architect, Howard Rourke, and she follows his life and his career. And the message is, really, it’s a version of to thine own self be true. And Rourke’s designs are to avant garde, nobody appreciates him and he just keeps doing what he wants to do and is just focused on his own visions, his own genius. I think that’s been really inspiring to creators of all types. I think it’s fairly unrealistic as a portrait of human achievement, but it’s an aspirational idea. I mean, one phrase that comes to mind is, there’s a character, I forget which one, who is in some adversarial relationship with Howard Rourke and says something to him like, “Well, Mr. Rourke, what do you think of me?” And Rourke says, “I don’t think of you.”

(02:43:23)
And that to Rand was the ideal. You’re not thinking of other people. You’re an island unto yourself. You’re focused on your own goals, your own capacities, and you’re not doing it to impress other people or to be better than other people or to dominate other people. You’re doing it because expressing your inner soul in a way. So that has been very captivating to so many, and the Fountainhead is one of those books we talked about that causes, people read it and they make changes in their life or they feel called to their higher self.
Lex Fridman
(02:43:55)
And I think there’s also the scene where Rourke with the Dean of Architecture at the school that’s speaking to what you’re saying, I think to me is inspiring. So this is the Dean of Architecture that expels Rourke, and then brings him into a meeting thinking Rourke will plead for a second chance. And the Dean says that, “Rourke’s work is contrary to every principle we have tried to teach you, contrary to all established precedents and traditions of art. Do you mean to tell me that you’re thinking seriously of building that way when and if you are an architect?” And then in a gangster-like way, Rourke says, “Yes.” And then Dean asks, “My dear fellow, who will let you?” And Rourke replies, “That’s not the point. The point is, who will stop me?”
Jennifer Burns
(02:44:44)
Yes. I mean, Rand’s coming from communist Russia, but it has a bit of the don’t mess with Texas flavor, I might say that really resonates with this idea of anyone who’s felt like they’re fighting the powers that be. Yeah, it’s interesting. I thought you might be going to the quote where he says something like, “I inherit no tradition. I stand at the beginning of one.” And I really think Rand’s thinking about herself when she says that, that she inherits nothing. She stands at the start. But The Fountainhead comes out in the middle of World War Two, and Rand is an unknown writer. This is a strange book. It’s a classic story. It’s turned down by 12 publishers before one takes a chance on it. And Rand really loved this story. The editor who read it said, “This book is great”.” And his boss said no. And he said, “If you don’t take this book, I’m quitting.” And so she idolized him for doing that.

(02:45:40)
So they print it and it becomes a bestseller just through word of mouth. So it’s not advertised. It gets one good book review, but people tell each other how much they like this book. And it keeps printing and selling out printings. It’s made into a movie. And so it lands in this time when Americans are engaged in this great collective endeavor of World War Two. They’re making all kinds of sacrifices for the collective and I think, paradoxically, as they do that, they’re drawn to this vision of someone who doesn’t have to compromise at all, who is leading their life exactly as they want to. Meanwhile, they might be sleeping on an ocean liner because they’ve been drafted to fight in this war, and they’re reading The Fountainhead and they’re feeling better about themselves.

(02:46:19)
And so it’s also really interesting. The Fountainhead is hugely popular in India, which is fascinating, and talk to people about this and they basically say, “This book comes like a breath of fresh air into a very traditional and conformist culture. And people just latch onto it and they love it, and it gives them that feeling of freedom and possibility that they’re hoping for.”
Lex Fridman
(02:46:41)
Yeah, I mean, it really is a book, Atlas Shrugged can be a bit of that too, but it’s more of the philosophy objectivism and the details and the nuance of that seeps into Atlas Shrugged. The Fountainhead is very much like a thing that makes you change the path of your life. I mean, that’s beautiful to see that books can have that power.
Jennifer Burns
(02:47:02)
And Rand knew that she was doing that and she knew what she was doing. This wasn’t an accident. And people say, “Oh, she’s a bad writer. Oh, her characters are so heavy-handed.” She started as a screenwriter. She started as someone who analyzed films for movie studios. She knew exactly how to manipulate plot and character and drama. And she also knew that she was writing, people say, “Oh, Rand is for adolescence, adolescent teenagers love Rand,” and that’s who she was writing for. And she said, “I’m writing for people as they start out on their life and they’re thinking about who they want to be.” So she’s not writing for the weary middle aged. She’s writing for the young who are looking for inspiration.
Lex Fridman
(02:47:46)
People say that to me sometimes about certain books like Rand, but also about The Alchemist. I know a lot of people for whom The Alchemist, and they’re adults and they’re brilliant people, The Alchemist changed their life. And the same can be said about The Fountainhead.
Jennifer Burns
(02:48:02)
Yeah.
Lex Fridman
(02:48:04)
And I sometimes get criticized for using words that are too simple. I think simple words can have power, and the cliche thing sometimes needs to be said. And sometimes it effectively needs to be said in an over the top way in the Mythic Register, because that’s the thing that resonates with us. We are heroes of our own story and we need to hear that message sometimes to take the bold step, to take the risk, to take the leap.
Jennifer Burns
(02:48:38)
Yeah, and I mean the other thing, she knew she was doing propaganda in a way. She was like, “I’m doing pro-capitalist propaganda.” She has a degree from the University of Leningrad. She was raised up in Soviet Russia. She said, “We need to present the case for the other side in the same way.” And that’s what she did.
Lex Fridman
(02:48:57)
Why do you think she’s so divisive? People either love her or hate her.
Jennifer Burns
(02:48:59)
I mean, I think it’s because of that purity that I’m willing to say, “You get what you deserve,” and that lack of charity. And part of that in her work is because she creates this fictional work where she can set everything up just so, and so you don’t have contingency or accident or bad luck. You don’t really have a lot of children. You don’t have handicapped people. You just have this idealized world. And I think it’s really infuriating for people who feel that’s so inaccurate. How can you be deriving a social theory and philosophy around this? And how can you be missing, what seems to many people, she’s missing the ethical instinct or the altruistic or charitable instinct? And so they just become enraged at that and they don’t want to see anyone go that far. And they’re outraged that someone went that far, that did the thing that Frank Knight said no one would do. It’s very unsettling.
Lex Fridman
(02:50:02)
Would you say that’s…
Jennifer Burns
(02:50:00)
Yeah, it’s very unsettling.
Lex Fridman
(02:50:02)
Would you say that’s her main blind spot, the main flaw of objectivism is just how black and white it paints the world? If not, what would you say are the flaws of objectivism?
Jennifer Burns
(02:50:19)
The big flaw is that it’s justified through a fictional world. It’s not justified through reference to the real world. It’s not empirical in a way. It’s not… Rand herself would say this. She’s not writing about things how they are, but how they should be. That idealism just really undermines it as a mechanism to understand where we’re actually living.
Lex Fridman
(02:50:47)
That is a big contrast with Milton Friedman, who would focus on how things are versus how things should be.
Jennifer Burns
(02:50:54)
Then I think it’s the problem of elevating rationality or any other mode of insight or thinking. What happens in Rand’s life, I describe this in some detail in the book, is she essentially creates a cult of reason around her and people who are drawn into this cult… It’s called The Collective. It’s a group of young people in New York City who are drawn to her work. She’s already famous, but she’s writing Atlas Shrugged. She’s sharing drafts of Atlas Shrugged as she goes along. And one of the members of The Collective, to bring all of this together, is Alan Greenspan, later be head of the Federal Reserve. And he’s incredibly taken with her. He’s one of these people who says, “I was a narrow technical thinker. I never thought about ethics or politics or anything bigger until I met Ayn Rand. She really opened my mind.”

(02:51:44)
He’s part of this tight-knit group. In this tight-knit group, they think of themselves, “We are all individualists. We’re dedicated to individualism and capitalism. We’re different than everybody else.” Over time, they all come to share Ayn Rand’s views and opinions on everything, from music to art, to clothes. She gets a dining room table and a bunch of them get the same dining room table, and it becomes incredibly conformist, because they’ve all believed they’re acting rationally. And they believe that to act rationally is to agree with Ayn Rand and they believe there’s no other way to make decisions than rationality. To disagree with her is to be irrational. They don’t want to be irrational. People get really caught up in this very damaging cult-like circle around her.
Lex Fridman
(02:52:33)
Plus, for a cult of reason, they get awfully emotional when there’s any disagreement with Ayn Rand.
Jennifer Burns
(02:52:41)
Yeah.
Lex Fridman
(02:52:43)
It’s hilarious. It’s absurd. It’s also beautiful to watch this singular figure. We’ve talked about several singular figures, like Frank Wright, that shakes up the world with her ideas.
Jennifer Burns
(02:53:00)
Yeah.
Lex Fridman
(02:53:00)
Of course, it would form a cult. And, of course, that cult would be full of contradictions and hypocrisies.
Jennifer Burns
(02:53:06)
Yeah, it’s amazing. Murray Rothbard is a famous anarchist, falls into the Ayn Rand cult. Then he disagrees and there’s some type of show trial where he’s told he’s wrong about everything. Then he has a little pseudo cult of his own, and two of his cult members switch over to Ayn Rand. Then one of them, to gesture their breaking of the relationship, mails him a dollar bill that’s been torn in half. This is high theatrics, right?

Sex and power dynamics

Lex Fridman
(02:53:41)
Okay, sticking on the drama and the theatrics, who was Nathaniel Branden?
Jennifer Burns
(02:53:46)
Oh yes.
Lex Fridman
(02:53:47)
Can you take me through the arc of Ayn Rand’s relationship with Nathaniel Branden to their dramatic falling out in 1968?
Jennifer Burns
(02:53:54)
Yes, after The Fountainhead, The Fountainhead is auctioned is sold to be a film. Ayn Rand moves to Hollywood where she’s going to help in the writing of the film. She wants a lot of creative control and she’s also still working in screenwriting and things like this. She gets a letter from… There’s a Canadian student who’s written to her several times, and then he writes again and he says, “I’m at UCLA.” And she’s like, “Young man, you’re so full of error. Why don’t you come visit me and I’ll straighten you out.” So he comes and they have this real meeting of the minds. They talk all night. He comes again, he brings his girlfriend, she loves him, and they start this very intense relationship of spending every weekend at her house, basically. Staying up all night, talking about ideas.

(02:54:39)
He becomes completely converted to the objectivist worldview. Rand begins counseling him and his girlfriend about their relationship, very intense thing. Then eventually they graduate from college and they both enroll in a graduate program in Columbia and they leave. After they’ve left, Ayn Rand is just bereft. And within a few months she packs up her home and she moves to New York. “Here I am. I like New York better.” That becomes the seedbed of The Collective. The Brandens, they get married. They change their name to Branden. They’ve never publicly spoken on this, but many people have pointed out it has the word Rand in the name. It’s some type of acknowledgement of how important she is to them. Time goes on, and romantic feelings develop between Ayn Rand and Nathaniel Branden, who’s some 20 years her junior. They discuss them and they realize that rationality has led them to the conclusion that they should be lovers.
Lex Fridman
(02:55:39)
Right.
Jennifer Burns
(02:55:39)
Right. They rationally decided this, but because they’re rational they need the consent or at least to inform their partners.
Lex Fridman
(02:55:47)
They’re both married?
Jennifer Burns
(02:55:48)
They’re both married. They call a meeting and they obtain the consent or maybe simply inform the others of the rationality of the choice. Then they say, “But this is only going to be an intellectual relationship, but we’d like a few hours alone each week and we don’t want to be deceptive, so we want you to know and approve of this.” The spouses bought into rationality, no one approve. One thing leads to another. It becomes a full romantic and sexual relationship. Although it’s open within these four people, it is not open more broadly. In all these meetings of The Collective, Alan Greenspan, all these other people coming up, drinking coffee all night, talking, talking. They all know that Nathaniel Branden is objectivist number one. They don’t know that there’s a romantic and sexual relationship happening. It’s kept a secret. Then when Atlas Shrugged comes out, it’s panned by reviewers.

(02:56:42)
People absolutely hate this book. And Rand is not Howard Roark. She falls into a deep depression because her masterpiece has been rejected. Then the romantic relationship ends, but the close personal relationship continues. And then over time Branden, who’s still married to his wife, begins an affair with another young woman. At this point he has started the Nathaniel Branden Institute to teach objectivism. He’s making good money. He’s becoming quite famous.
Lex Fridman
(02:57:15)
She supported the institute?
Jennifer Burns
(02:57:15)
She supported it. And at first it was to help her in her depression. He said, “The world needs to recognize your genius. They missed Atlas Shrugged, but I’m going to teach them. I’ll bring the message.” And it’s very successful. It becomes its own business. It has a newsletter. It’s a whole world. That small cult around Ayn Rand expands to this whole social network, and it’s very much a piece with this burgeoning conservative movement. Objectivists are involved in criticizing the draft. There’s kind of a libertarian objectivist world going on. All of this is happening. In the meantime, Nathaniel Branden has found a new partner, but he doesn’t tell Ayn Rand this because he knows she’ll be upset. It goes on for years.

(02:57:57)
Ayn Rand knows something is going on, but she can’t quite figure it out. And finally, Barbara, Branden says to Nathaniel Branden, “You know, you have to tell her. This has just gone on too long.” She finds out and the whole thing blows up, and she exiles him and she breaks off contact with him. And nobody has ever told what happens. It’s called the [inaudible 02:58:23]. Objectivism breaks in two, because some people say, “How could Ayn Rand do anything wrong?” And other people say, ” What is this letter all about? And what did Nathaniel Branden do? I’m not just going to take her word for it. I need more information.” Then a bunch of people read all the accounts of this. A bunch of people are like, “Okay, they were having an affair.” And a bunch of other people are like, “No, that couldn’t possibly be happening.”

(02:58:48)
The whole thing breaks up. But what I argue in my book is actually this is to the benefit of Rand’s ideas, because Rand herself was so controlling over her ideas. And now that she steps back from a public role, objectivism flows into the student libertarian movement. Some objectivists become conservatives. It just spreads out more generally, and you don’t have to drink the Kool-Aid. You don’t have to take the official course. Nathaniel Branden goes on to be part of the self-esteem movement, Human Potential Movement, California and Ayn Rand lives another 10 years or so, but she doesn’t do major work after that.
Lex Fridman
(02:59:27)
Since we were talking about some of the all… Although rationalized some strange sexual partnerships that their engagement in, I have to ask about The Fountainhead and the, quote, unquote, “Rape scene.” In the Fountainhead. Was she intending to add that there to be controversial? How we’re supposed to read into it? Is it a glimpse into Ayn Rand’s sexuality? Maybe broadly we can say, “Well, what was her view on sexuality, on sex, on power dynamics in relationships?”
Jennifer Burns
(03:00:04)
Yeah, there’s also an objectivist theory of sexuality in that probably the least convincing of all the parts of objectivism. And it goes something like, “Your sexual desires express your highest values.” And they are related in some ways to your rationality, right? Which is also related to your highest values. So for her that explained her attraction to Nathaniel Branden and Nathaniel Branden’s attraction to her was a function of their highest values. And, in fact, Branden imbibed this so deeply that the fact that he was later drawn sexually to a woman who was not particularly accomplished, but was beautiful, caused him deep anguish and guilt for being non-objectivist. This is the objectivist theory. Then the gender politics are just crazy. We have to back up and think, “Okay, who is Ayn Rand.” She’s born Alisa Rosenbaum in Russia. She is someone who stands out from the crowd from the beginning.

(03:01:04)
She never really fits in. She’s not conventionally beautiful by any stretch of the imagination. She struggles with her weight and she doesn’t consider herself to have a beautiful face. She’s very independent. She meets none of the metrics of traditional femininity at all. She finds love with a man who is very handsome but very passive. Yet she writes in all her fiction about strong manly heroes. There seems to be like a projection. The man she’s actually with is not a strong manly hero. The hero she writes about, she probably wouldn’t be able to be in the same room with them for more than one minute before they got in a raging argument, right? Then she develops this theory about women and men in that a woman should worship her man, and a woman finds her true expression in worshiping the man she’s with. Again, this is not at all how Ayn Rand lives her life.

(03:01:59)
This is like this. I would say compensatory theory for her lack of ability to conform to the gender norms of her day. She then articulates them in a very strong and almost distorted and exaggerated way to compensate for the fact that she doesn’t actually meet them, can’t actually enact them. The rape scene to some degree embodies that idea that to some degree that the woman should worship the man. I tend to read it more in terms of literary genre. Rand is a screenwriter, a consumer of movies, and that rape scene is paradigmatic for the romance genre. In other words, these are like pulpy romance novels. The hero rapes the heroine and then they fall in love. That’s just the trope of how it works.

(03:02:56)
It’s crazy when you read it, but if you were reading a bunch of novels in this genre, you would find this is very standard. But that is a huge part of this appeal at the time. There’s this feminist who hates Rand, Susan Brownmiller, and she wants to write an angry denunciation of the rape scene. So she goes to get The fountainhead and she’s wondering how is she ever going to find the scene in this 800 page book? It’s a library copy because she doesn’t want to buy it, and it just falls open to the rape scene because everybody’s gone and read it because it’s very racy and explicit for that time. I’m almost positive she also knew that like, “If I put in this kind of taboo-breaking sex scene, that’s also going to probably be why people tell their friends about it.” I think it’s a mess. I think all of the gender and sexuality stuff that she states is just a total mess.
Lex Fridman
(03:03:50)
I think it also reminds me of another guy related, Friedrich Nietzsche, who had very strong opinions on women and wrote about what women’s role in society should be in different power dynamics and relationships and all that kind of stuff when he himself really had trouble getting laid. You have to as to always maybe chuckle or take with a grain of salt the analysis of power dynamics in relationship from these figures which failed in many regards in their own private life. You mentioned feminists. Would you consider Ayn Rand a feminist?
Jennifer Burns
(03:04:30)
She’s almost an anti-feminist because she then goes on… Someone writes her a letter about like, “Should there be a female president?” Or something. This is the beginning of feminism. And she says, “No woman should ever be president because if she’s president, she wouldn’t be able to look up to any man because she would be so powerful, and therefore she would be corrupt and rotten in the soul and unfit to be a leader.” It just makes no sense.
Lex Fridman
(03:05:04)
But that said, she’s a woman and she’s one of the most powerful intellects in the 20th century.
Jennifer Burns
(03:05:08)
Yeah.
Lex Fridman
(03:05:11)
The contradictions… Nietzsche’s full of contradictions of this sort. The very fact that she’s one of the most powerful minds in history, to me, means that she is a feminist in the spirit she embodies, right? In what she represents.
Jennifer Burns
(03:05:31)
She lived the ideals of individualism in her life and set aside gender norms in her own life, but she did not see herself as doing this for the benefit of other women or to change society’s views about women. There was no collective essence to it. If feminism has some sort of collective aspect to it, or at least some identification, one needs to identify with a broader category of women and feel they’re acting on behalf of that, she’s definitely not doing that. She was fair to women in her life, promoted them in her life, but did not… She was very negative about feminism, because they dressed terribly. Then the other thing, it’s really interesting, there’s all these kind of homoerotic themes in her writing. For that reason, many gay men were drawn to her writing, and then she would say like, ” Homosexuals are dirty, terrible people.”

(03:06:32)
She would denounce people for being homosexual. There’s a whole actual literature of gay men wrestling with Rand and what she says about gay people. Yeah, it’s hard to make sense of. I want to be charitable. I just think of the enormous pressure she was under in the culture she was raised in, the expectations that were placed upon her, and her just utter inability to meet any of them. And it came out in this very tortured set of ideals that she tried to promote. And this lack of ability was probably too painful to introspect and to think about that. So she just tried to rationalize her way through it, and it came out in these very strange theories.
Lex Fridman
(03:07:21)
Why do you think that Ayn Rand is… Maybe you can correct me, but as far as I can see, never mentioned in the list of great thinkers in history or even the great thinkers of the 20th century or even the great female thinkers of the 20th century. You have somebody like Simone de Beauvoir, Hannah Arendt. I almost never see her in the list.
Jennifer Burns
(03:07:43)
Yeah.
Lex Fridman
(03:07:43)
If you Google those silly list whatever top thinkers of the 20th century, she’s not mentioned. Why is that?
Jennifer Burns
(03:07:52)
A lot of people just deeply dislike Rand. They deeply dislike her ideas. They don’t think they’re profound because their disconnection from other ideas and other understandings of human society. I think [inaudible 03:08:06] you could look at them and say, “These ideas are very provocative and they’re very deep because she’s not taking anything for granted, and she’s flipping everything around and forcing you to really think.” To a lot of other readers, to her critics, they just look absurd. How could you even make these contentions? And I think that because she… She’s not without precedence and she’s not without followers, but she doesn’t knit herself into an intellectual community the way that these other thinkers do very naturally. You can see who they influence. You can see who they’re in dialogue with. I think my book was one of the first to really take Rand and say, “She’s a figure in American history. Here’s who she’s connected to. Here’s who she’s influenced.”

(03:08:54)
And I got a lot of pushback for that. I think now people are more open to it, but I think the people who compile these lists really dislike her work and they think it’s shallow because they find her fiction overdrawn. They find her work, in the mythic register, simple. And she’s also a grand systematic thinker in an age that’s over systems. She’s almost creating an inverse Marxism, right? Marx was writing in 1848. He’s not a thinker of the mid 20th century. I think that’s part of it, the lack of a legacy and the dislike of what she had to say, and the feeling that she’s too detached. Her insights are not insights because they’re too idealized rather than being rooted in a theory of human nature that people find plausible.

Evolution of ideas in history

Lex Fridman
(03:09:48)
You study and write about history of ideas in the United States over the past 100 plus years. How do you think ideas evolve and gain power over the populace, over our government, over culture, just looking at evolution of ideas as they dance and challenge each other and play in public discourse? What do you think is the mechanism by which they take hold and have influence?
Jennifer Burns
(03:10:19)
Yeah, there’s a couple different ways, I think, it happens. I really am interested in the relationship between the thinker and then the reader and the interpreter of the ideas, and then the conditions on the ground that make that idea resonate or not resonate. I think a lot. As an intellectual historian, I’m studying the ideas. I’m always putting them in their historical context like, “What is happening that is making these things resonate? That is making them… People seek them out. For Rand’s case, she has this credibility because of her experience of communism. She’s one of these defining moments of the time. Then I think the idea comes out in a pure form, and then other people rework it and reshape it as they read it. And I’m really interested in how people form communities around these ideas. A bunch of people started calling themselves objectivists and getting together to read Rand’s work. That was spontaneous and ground up. It wasn’t supported by any money. Nobody planned it. It just happened.

(03:11:29)
Friedman’s a different case in that he joins an established tradition of thought that’s been institutionalized in universities. So people are signing up and paying money and getting credential to learn these ideas. To my mind, these are two different ways but really emblematic ways of how ideas spread. Rand, I think of as more bottom up. People encounter the idea in a book. They’re blown away by it, or they imbibe it without even realizing they’re imbibing it, and then they’re like, “Well, maybe I don’t like Franklin Roosevelt so much.” Or, “Maybe I’ll look at another time at Barry Goldwater.” Then whereas Friedman, you get the idea more top down. I know I’m getting the idea. I know I’m being positioned within a elite discourse of economics. I think they go top-down and bottom-up, and then they hit the events, right?

(03:12:16)
Friedman’s ideas wouldn’t have gone anywhere without that episode of stagflation that really made people think they proved out. And I think Rand’s ideas really caught fire in Cold War America, that’s looking for a statement of like, “What does it mean to be an individual? What does it mean to live in this mass society?” Because it’s also a time of great social conformity, where people are suddenly… They’re working for large corporations. They’ve been served in a large military. The United States is stepping out onto the world stage. Everything is bigger. What does it mean to be an individual in that world? That’s where Rand’s ideas catch fire. I think a lot about that. About how they trickle through different levels of society, and then how ideas collide with experience I think is critical.
Lex Fridman
(03:13:04)
What do you think about when they actually take power in government? I think about ideas like Marxism and how that evolves into the Bolshevik Revolution and how that takes hold in its implementations. Or you can think about Nazism and with Hitler where it goes from a small number of people that get real excited about a thing and then somehow just becomes viral and takes hold in power, and then that has its consequences.
Jennifer Burns
(03:13:30)
When I think about this sort of historical path of communism and the logics and dynamics of communism, in many ways it has some echoes with Rand in that the ideology in its purest form is almost… It’s a rationalist ideology of some ways. It’s an analysis of history and how things are supposed to be. I think you mentioned Hannah Arendt. I think she is one of the most penetrating analyses of communism, which she really puts in category, like it’s a logical ideology. Logic leads inexorably to its conclusions, and then experience crops up and experience is different. And what does a cult of rationality do when it hits experience? Well, it tries to bend experience to its will. That, I think, is really the story of communism writ large.

(03:14:25)
The question though is why does it catch fire? Why does it draw people into political allegiance? I think in the case of communism, it’s this dream of a more ethical world, dream of equality, dream of the powerless rising up against the powerful. That’s drawn in so many. Then you had the whole edition of Leninism, which gave a international cast to that and helped people think about what are the relations between poorer and richer countries and what can we expect out of them and what might happen. Gave us a framework for thinking about that in a time when the world was becoming more interconnected and those differences were becoming more obvious. Fascism to me is unleashing more something dark and primal within people, and it’s more a permission structure to indulge in that that is normally not there. Those impulses are normally channeled or held down, and it seems that when the fascist regimes come into power, they give people permission to let those forces out.
Lex Fridman
(03:15:32)
I think on communism, going back to that lecture that Ayn Rand gave, I think what rings true to me a little bit is that what fuels it is a kind of… Maybe not resentment, but envy towards the have-nots versus the haves. And there’s some degree to wish Nazism has the same envy towards some group, resentment towards some group. Given the environment of hard times, hard economic times, combined with a more primal, just envy of not having and seeing somebody who has it and just constructing a narrative around that. That can become a real viral idea.
Jennifer Burns
(03:16:19)
Yeah, it seems like communism is more animated by this idea of injustice. The world is unjust. It should be different. And fascism seems like the process of scapegoating, right? We’ve identified the source of the problem, and it’s this group and they need to be punished for what they’ve done to the rest of us.
Lex Fridman
(03:16:40)
There is a primal thing going back to literature in 1984, Two Minutes of Hate, where you can get everybody real excited about hating a thing. There’s something primal about us humans where once you’re in that state of hate, anyone can direct that hate towards anything. Towards any group, towards any idea, towards anything. We could get caught up in the mass hysteria of the hatred. It’s a dangerous thing. You floated the idea, I forget where, of pivoting for your next book towards maybe writing about postmodernism, which is a set of ideas almost the opposite of Ayn Rand’s philosophy. Can you maybe explain your curiosity about, first of all, spaces of ideas, but maybe postmodernism?

Postmodernism

Jennifer Burns
(03:17:40)
Yeah, I think in the broadest sense, what I’m interested in, kind of two dimensions, that guide me in doing intellectual history. One is what I talked about, like how does an idea go from a book, an elite space, out to more popular dimensions? How does that happen? What happens to the idea along the way? How is it distorted or changed? And the other is just search for meaning in a post-Christian era or a secular era. Like, “What are people coming up with?” And I think to replace that void in their religious or spiritual lives, I think both Rand and Friedman offered these alternatives, right? Objectivism, quasi-rationalist religion. People take economics as a theory of the world that you can almost believe in it, right? It can almost take that place. And in both cases, how did those ideas travel? When I think about postmodernism, it first struck me… If you read the original postmodern thinkers, it’s really tough going.

(03:18:40)
I make my students do it, and they suffer. I think they see it’s worthwhile, but it’s no fun to read Derrida. But somehow it’s trickled down into how do we go from Derrida to Tumblr? And I realized like, “Oh, this has happened with postmodernism.” It’s followed the same path, say, from Milton Friedman’s economic theory to free to choose on YouTube. We’ve had a similar path of high French theory down to Tumblr, and I sexually identify as an attack helicopter or whatever it may be. That was really interesting. Then I also thought, “Well…” At the same time, this is clearly a structure of meaning. And I actually think it’s followed the same path of objectivism, which is distilled down and then turning into its opposite.

(03:19:31)
So if objectivism was a group of people who considered themselves individualists, who ended up deeply conforming to the dictates of a charismatic leader. Postmodernism started about disrupting binaries. We’re going to be fluid. We’re going to go beyond the border. We’re going to disrupt the binary. And it’s devolved in its popular forms to the re-inscribing of many different binaries. Oppressor and oppressed has become this paradigmatic set of glasses you put on to understand the world. I think the dynamics are very, very similar. I think it’s something in the traffic of the idea from its pure form to its popular form, and then how it gets politicized or mobilized in different ways. And behind it all, I think, is this human longing for meaning and the inadequacy of the traditional ways that need was met at this point in time.
Lex Fridman
(03:20:22)
By the way, going from pure form to popular form, I remember… This might be before the internet, but when I was in college reading Derrida and Foucault, not knowing context at all, it wasn’t interesting. All right. I’m able to read pure encapsulations of an idea and just like, “Oh, all right, well that person believes that.” And you just hold it. But if you actually take the pure form of that idea and then it creates a community around it, you realize what that actually becomes and you’re like, “Oh, yeah no. I don’t…” That’s not… Although I do consider myself sexually an attack helicopter, that identify sexually. That’s beautiful. Okay, your process of researching for, let’s see, the biographies of Milton Friedman and Ayn Rand seems like an insane amount of work.
Jennifer Burns
(03:21:20)
Yeah.
Lex Fridman
(03:21:20)
You did incredible work there, going to the original sources… Can you maybe speak to that? What is required to persevere and to go for so many years, to go so deep to the sources?
Jennifer Burns
(03:21:39)
Yeah, I go to the archive. That’s where I feel like I’m communing with the dead in some ways. I’m seeing what they saw in some ways and reading what they felt. I tell my doctoral students, “It’s got to be something that gets you out of bed in the morning, because there comes a point in your doctoral career when there’s nowhere to go. There’s nowhere to be. You got to be getting up because you’re interested in what you want to study.” And so with Rand, it was this real sense of discovery. I am discovering. I want to know about this woman. I want to know where she fits. And the only way to find out is to do the research.

(03:22:13)
Yeah, I like to go deep. It’s really interesting to me. And I should say in both of these cases, I’ve done it in an institutional structure. I don’t know that I would do it independently. The first was a graduate program in history, was at UC Berkeley. I had coursework. Then I had structures, and I did have people to check in with and read, but I had a great deal of latitude. I’m very grateful for… People are like, “You wrote a dissertation on Ayn Rand at Berkeley?” I’m like, “Yeah, hell, I did.” Berkeley’s like… It’s a great place. At the time I was there, there was absolute room for free inquiry.
Lex Fridman
(03:22:49)
Oh, can you just linger on that? When you said that you’re doing that and doing a dissertation on Ayn Rand, was there… Did people get upset?
Jennifer Burns
(03:23:00)
No. I did have a friendly critic who took it upon himself to throw at me everything he thought the outside world would throw at me. I think maybe 5 or 10 years earlier, it wouldn’t have been possible. But the most important thing I had to… The person I really had to convince this was worth doing was myself, because I knew it was an unconventional choice for the field and for a dissertation. But once I convinced myself, I just said, “Well, I’m going to do this and see.” And because it was unconventional, it ended up standing out. And it really was the time… I started it during second Bush administration, George W. Bush, second term. People were interested in just conservatism in general. No matter where they stood on the political spectrum, felt like objectively we don’t know enough about this and this is a problem, so they were open to learning more. I really caught that wave in scholarship and caught that wave in American culture where people wanted to know more.
Lex Fridman
(03:23:59)
We should probably say that Ayn Rand at the very-
Lex Fridman
(03:24:00)
Probably say that. Ayn Rand is, at the very least, as you’ve mentioned, a kind of gateway to conservatism.
Jennifer Burns
(03:24:07)
Yes. I called her the gateway drug in that people start with Rand. They’re taken by her. In some ways, she takes the worldview of Milton Friedman in terms of what capitalism can accomplish economically. And then she puts it in this mythopoetic register, and she fictionalizes it. So once people have absorbed that, they want more. They go on to learning more of the ideas behind that vision, or they have become true believers. They’ve converted.

(03:24:36)
And so then they head off to work for a politician, to work for a think tank, to work for a party. t’s absolute traffic. Now, not everyone. There’s plenty of people who read Ayn Rand who don’t take the politics in. It’s a nice story. It’s interesting, just an episode in their life. But for others, it’s really foundational. It really changes them.

(03:24:52)
So those were the people I wanted to track very deliberately. I wasn’t trying to do in the round everything about Ayn Rand. I was like Ayn Rand and the American Right, Goddess of the Market. Ayn Rand and the American Right is the title. So where did they take her, those who took her in this political direction? What difference did she make?
Lex Fridman
(03:25:10)
If we return to the actual your process, so you’re showing up, you’re reading sources. And you’re like, “Is it like the process of discovery?” You’re just taking it all in and seeing what unifying ideas emerge, or maybe special moments that illustrate an idea emerge?
Jennifer Burns
(03:25:32)
Yeah. I know with the biography of a person, I am already given a start and an end date and a rough narrative of what happens. So I have a structure. And then, both with Rand and Friedman, I started by reading their major books before I really read anything about them because I wanted my own experience of the material to be fresh. And I had read some Ayn Rand, but not a lot.

(03:25:57)
Similarly, I had read some Friedman, but not a lot. So I first is like, “Let me read the major stuff, get oriented,” and then just dive into the archive and see what’s there. Who are they talking to? What’s going on? in Rand’s case, I was interested in her in the United States, not her in Russia. I didn’t have the language skills to do that. So I start her in the United States, and I start when she publishes her first book, and she starts getting letters. And who is she writing to? Who’s writing to her?

(03:26:27)
And then I start to uncover this world of nascent conservatism. And I’m putting that together. And once I have enough, I say, “Well, that’s a chapter.” I’m going to cover that chapter. And then there’s going to be the book has come out. And so now, I need to start a different chapter. What’s her life after the book has been published?

(03:26:44)
And then I look for that. Although I have this very high level structure, it’s coming out of the archive, the material I’m finding. And if I’m not finding the material there, I won’t cover it in great detail, or if I’ve decided it’s outside my ambit, I’m not going to go into great depth on it.
Lex Fridman
(03:27:02)
And you’re trying to understand the relationships. It’s so fascinating, like reconstruct in a dark room, trying to reconstruct, shine a light on relationships through reading letters. It’s interesting.
Jennifer Burns
(03:27:13)
Yeah. Yeah. Correspondence is really, really helpful, drafts, correspondence. And someone this famous, they have oral histories. Other people write about them. So you’re reading all these different things and triangulating and trying to put them together and then think about how do I present this in a compelling story, and what do I need to explain?

(03:27:34)
And then also for me, what was really helpful is that because I teach and I am explaining the broad sweep of 20th century history, so I know that Rand’s involved in a labor action at Warner Brothers, but through my teaching, I realized, “Oh, yes, this is a moment of labor strikes across the country.”

(03:27:53)
And so then that really changes the origin story of Atlas Shrugged because she’s looking at labor actions, and she originally thinking of the book as being called The Strike. So she’s really responding in real time and being inspired by what’s happening in the mid-1940s in the United States. So then I can take that and run with that and figure out where to go.

Advice to students

Lex Fridman
(03:28:16)
So you’re super passionate about teaching. You mentioned Milton Friedman had a very interesting way of teaching. So how do you think of teaching, teaching history, teaching history of ideas, teaching bright young minds about the past?
Jennifer Burns
(03:28:33)
Yeah. It’s great. It’s really inspiring the ways the old school dominating way in which Friedman taught would not fly in today’s university wouldn’t be permitted. And also, the students wouldn’t respond to it. So I try to share my enthusiasm. I think that’s almost the number one thing I bring, is my enthusiasm. Look how neat and interesting these ideas are.

(03:28:56)
I try to keep my own views out pretty much. I try to give the fairest possible rendition I can of each thinker. If I find someone really disturbing, I might side bar at the end of the lecture and say, “I find this unsettling, and this tells me something about myself.” But most of the time, I’m bringing people into the biography of a great thinker, the context of them. And then, in the lecture, we’ll literally read the work together, and we’ll talk about it.

(03:29:24)
And I’ll ask the students, “What are you finding here? What’s jumping out at you?” Kind of breaking down the language and really teaching them how to do deep reading. So I feel like that is my contribution right now. We’re having trouble reading collectively. We’re having trouble paying attention collectively, and I’m trying to cultivate their skills to doing that and showing them how I do it, and also modeling like, “This is how I would read a text. This is what jumps out to me when I look at Thomas Kuhn or something like this,” and just show them that studying a history of ideas is really fun. I feel incredibly privileged to do it.

(03:30:00)
And the other thing is, I think this is the time for students in college figuring out who they are. Their minds are developing and growing. They can really handle complicated hard ideas. They don’t always have the context behind them. So I need to give them the hard ideas and then show them this is kind of the context of what’s happening in the world. But really, I’m showing them the landscape. I don’t have time to go deep. We have a 10-week quarter giving them a flyover. And then I want them to know how to go deep and know where they want to go deep.
Lex Fridman
(03:30:33)
Do the thing that Milton Friedman did, which is in parallel, [inaudible 03:30:39] books.
Jennifer Burns
(03:30:38)
Yes. Do their own parallel curriculum. Exactly. Exactly.
Lex Fridman
(03:30:42)
What advice would you give in terms of reading about ideas you agree with and reading ideas do you disagree with?
Jennifer Burns
(03:30:48)
Even though I think the passion is important for the teaching of the ideas, like this passion is more important for the reading and understanding of them. So a lot of people have said to me, “I could never write about Ayn Rand. She makes me so angry.” I don’t get angry reading her. I am like, “Oh, there you go again,” or, “Well, that’s going to cause trouble.”

(03:31:12)
And so I guess I’m approaching it with this sort of charity, but also with, I don’t have huge expectations. I’m not expecting to have the light shine on me. I’m not expecting to agree. I’m like, “I can be very clinical about it.” So that’s worked for me. It might not work for others.

(03:31:31)
And then, I just try to find the humor in it. How funny is it, these different aspects of them? When teaching my students about Oliver Wendell Holmes, his dad wrote a poem about him. He called him the astronaut about how he came from outer space. He seemed like he came from outer space. I’m like, “This is his dad’s view of his son.” That’s how weird of a guy he was.

(03:31:55)
And so I try to find that, keep alert for those funny kind of human touches that these are ultimately just people, people with ideas that they spent enough time polishing up and developing that we still want to read about them 100 years later.
Lex Fridman
(03:32:07)
What about the dramatic formulation of that same question? Do you think there’s some ideas that are good and some of that are evil? Do you think we can draw such lines or is it more complicated like the old soul genius in line between good and evil that runs to the heart of every person?
Jennifer Burns
(03:32:24)
I philosophically agree with Solzhenitsyn, for sure. I do think some ideas pull on the good side and some ideas pull on the bad side, absolutely. And I think that’s probably why people dislike Rand so much, is they feel like she’s giving license to the bad side, and she’s saying, “It’s okay to be selfish, and it’s okay…” They feel like she’s the dark forces.

(03:32:47)
And in some cases, that may be true, but she’s also unloosing some of the light forces in terms of reflecting on yourself and trying to be true. But definitely, there are ideas that are dangerous to play with. And there are ideas that I think give license to the darker sides of human nature. But I think you can see that in the historical record. So I think that it’s possible to show that. And obviously, there’s some places like Germany. They think the ideas are so dangerous, they can’t be allowed to circulate. And in some contexts, that may absolutely be true.
Lex Fridman
(03:33:25)
And then still even that, we should take with a grain of salt because perhaps censorship of an idea is more dangerous than an idea. So all of that, that’s the beautiful thing about us humans. We are always at tension trying to figure out what ideas are the ones that are going to help humanity flourish. Pothead question, do humans have ideas or do ideas have us?

(03:33:50)
So where do ideas come from? You have Milton Friedman sitting there after Rutger’s trying to figure out what he can do about the Great Depression. Do you ever think about this? I sometimes think aliens are actually ideas. They’re just kind of travel through human brains and captivate us. And we get all real excited with the monolith in 2001 Space Odyssey, a monolith lands, and everybody gets excited. And somehow this idea just gets everybody to be on the same page, and it reverberates through the community, and then that results in an implementation of some action that results in us figuring out that that idea was actually bad, and we learn new ideas. But it feels like the idea is right in the show.
Jennifer Burns
(03:34:44)
Yeah. I think in a lot of cases, I think it’s true. Keynes has this famous quote like, “Most men are slaves of some defunct economist.”
Lex Fridman
(03:34:53)
That’s funny. That’s funny.
Jennifer Burns
(03:34:56)
So I do think it’s really hard to have an original thought. We are social creatures. We encounter the same situations again and again. And so it’s really hard. You’re born into these traditions of thinking and being and knowing. And most people are never going to question them, and most people are never going to become aware of them.

(03:35:15)
So again, that’s some of the work of what I do as an intellectual historian. It’s like, “Let’s become aware. Let’s realize that you’re carrying a map that’s orienting you to the world in a certain way.” And so I think you have to work really, really hard to have an original idea. And even then, it’s not a completely original idea. It’s a reworking and a reassembling of ideas others have had.

(03:35:38)
So I definitely think it’s possible to create autonomy in the realm of ideas and to be an autonomous consumer of ideas. But I think, on balance, most people are not. And that’s fine. They want to have experiences. They want to do other things with their life.
Lex Fridman
(03:35:55)
Well, Jennifer, thank you so much for this journey through ideas today, and thank you so much for your incredible work. It was really fun and fascinating to talk with you today. Thank you,
Jennifer Burns
(03:36:05)
Thank you.
Lex Fridman
(03:36:07)
Thank you for listening to this conversation with Jennifer Burns. And now, let me try to reflect on and articulate some things I’ve been thinking about. If you’d like to submit questions or topics that I can comment on in this way here at the end of episodes, go to lexfridman.com/ama, or contact me for whatever other reason at lexfridman. com/contact.

Lex reflects on Volodymyr Zelenskyy interview


(03:36:33)
Please allow me to say a few words about my interview with the president of Ukraine, Volodymyr Zelensky, now that a few days have passed and I’ve had the chance to think about the conversation itself, the response, future upcoming conversations, and what it all means for the war in Ukraine, for global geopolitics, and for us humans in general.

(03:36:55)
I’ve gotten a lot of heartfelt positive words from all sides, including, at least so far, literally everybody who knows me personally inside Ukraine, which includes a lot of soldiers and many high profile figures, some who are supportive of the president, and some who are critical of him.

(03:37:14)
Literally, all private communication has been positive and supportive. This is usually not the case with me. Friends usually will write to me to criticize and to disagree. That’s the whole point of friendship, to argue and have fun doing it. There was none of that here, at least so far. So thank you for your support and kind words. It means the world.

(03:37:39)
The most common message was, please keep pushing for peace. I will. But online on the interwebs, I saw a lot of attacks sometimes from swarms of online accounts which, of course, makes me suspicious about the origin of those attacks.

(03:37:57)
One of my friends in Ukraine, who by the way thinks the attacks are all propped out by Ukrainian bot farms, said, “There’s no need to say anything extra. Let the interview stand on its own. Just keep focused on the mission of pushing for peace.” Basically, he’s a Ukrainian version of my other friend, Joe Rogan, who to this day says, “Don’t read the comments.”

(03:38:20)
This is generally good advice and I try to follow it. But I’m also human being. I wear my heart on my sleeve. And this interview, this war for me is deeply personal. And the level of vitriol, misrepresentation, and lies about the conversation and about me personally was particularly intense and disingenuous. So I thought I would use this opportunity to say a few words, just speak a bit more about how I approach this conversation with President Zelensky and conversations in general.

(03:38:53)
This interview is something I poured my heart and soul into preparing a lot. I’ve described parts of the preparation process I follow in the outro to the Zelensky conversation. But in general, let me say that I’ve read a lot, listened to a lot, and had a lot of private conversations with people on the ground. I have many flaws. But being unprepared for this conversation is not one of them.

(03:39:19)
Two low effort attacks got to me a bit, if I’m being honest though I am learning to take it all in stride. First attack is that I’m unprepared, uninformed or naive. I don’t give a damn about the trolls, but I want people who listen to me, who support me, who care about my words, to know that this is not the case. It never will be the case for future conversations, especially ones of this importance. I work extremely hard to prepare.

(03:39:52)
Second low effort attack that got to me a bit is that I’m a shill for Zelensky or a shill for Putin. Both accusations were hurled readily and freely by the online mob of all persuasions by the left and the right in the United States and Europe by the pro and the anti-Zelensky people in Ukraine or of Ukrainian origins and by the pro and anti-Putin people in Russia or of Russian origins.

(03:40:20)
As I’ve said over and over, this is not the case. I will never be the case. I’m a shill for no one. More than that, I just simply refuse to be caught in any one single echo chamber. It’s an ongoing battle of course, because social media algorithms and the various dogmatic groups and tribes out there want to pull you in to their warm embrace of belonging. And humans want to belong.

(03:40:48)
But the cost of the path I have chosen is that I will never belong to any one group. In the end, like many of us must, I walk alone. And I try to do my best to do what is right to my independent heart and mind, not what is popular with any one group.

(03:41:08)
My goals for this conversation were twofold. First, give a large platform to President Zelensky to explain his perspective on the war and to do so in a way that brings out the best in who he is as a leader and human being. Second goal was to push for peace, and to give him every opportunity possible to signal that he’s ready to make peace, and to provide his vision for what that might look like.

(03:41:35)
And just to be clear, by peace, I mean long-lasting peace that minimizes suffering of people in the region and maximizes the flourishing of humanity in the coming decades. The war in Ukraine has led to over one million casualties and growing every single day. For some people torn apart by loss, tormented, and forced into a state of anger and hate, peace is a dirty word. To them, nothing less than justice must be accepted.

(03:42:10)
I hear this pain. I’ve seen the bodies and the suffering. It’s true. Peace will not bring back your loved ones. But it’ll prevent further slaughter of more people, each of whom are someone else’s loved ones. So again, the second goal of this conversation was to push for this kind of peace.

(03:42:35)
So how did I approach it? Every conversation is its own puzzle. So let me try to explain my approach for this one. As I’ve said, I read and listened to a lot of material since February 24th, 2022. There would be many weeks over the past three years where I would spend every day over eight hours a day of focused reading and research.

(03:42:58)
There were several rabbit holes that I consistently returned to and researched. But the most important line of inquiry was always peace talks, not just in this war, but in other wars in modern history. For this specific war, as part of the background prep, I would take notes on every single perspective I could find on every single major diplomatic meeting and negotiation that happened in Ukraine-Russia relations since 1991.

(03:43:27)
There is a lot of material to go through. And there are a lot of perspectives, even on the very 2019 meeting that President Zelensky spoke about in this podcast. Just as a small but important example, Andrii Bogdan was interviewed twice by Dmytro Gordon and gave a deep inside look of the administration of President Zelensky, including that very 2019 meeting.

(03:43:52)
The two interviews are seven and a half hours by the way. And from my interviewer perspective, are a masterclass of interviewing. Andrii Bogdan worked directly with President Zelensky as the head of the Office of the President of Ukraine. He was there for the 2019 face-to-face meeting between Volodymyr Zelensky and Vladimir Putin at the Paris Summit, along with French President Emmanuel Macron and German Chancellor Angela Merkel. This was part of the Normandy Format Peace Talks.

(03:44:26)
In those two interviews, Andrii Bogdan gave a very different perspective on that 2019 meeting than did President Zelensky to me in our conversation, the perspective being that the failure to negotiate a ceasefire and peace was not a simple one-sided story.

(03:44:46)
I don’t think this is the right time for me to dive into that data point and be critical. I’m not interested in being critical for the sake of criticism. I am interested once again in productive conversations, critical or otherwise, that push towards peace, the kind I described earlier. This is merely an example of a data point I was collecting in my brain.

(03:45:09)
There are many, many others. But all of it taken together made it clear to me, and I still believe this, that it is indeed very difficult but possible to negotiate long-lasting peace with Vladimir Putin. It’s certainly true that Ukraine is best positioned to negotiate from a place of strength. After the invasion of February 24th, 2022, I believe there were three chances where peace was most achievable. First chance was March and April of 2022 with a successful defense of the north. Second chance was the fall of 2022 with the successful counteroffensive in Kherson and Kharkiv.

(03:45:54)
The third chance is now. As he has stated multiple times publicly, Donald Trump is very interested in making peace. It is likely that the US financial support for this war will continue to dwindle. So the leverage and the timing for peace negotiation is now. There is unlikely to be another chance like this for a long time.

(03:46:18)
Just to zoom out on the conversation piece of this, I interviewed Donald Trump and may do so again. I interviewed Volodymyr Zelensky and may do so again. And it seems likely that I will interview Vladimir Putin in Russia in the Kremlin. I understand the risks and accept them. The risks for me are not important. I’m not important. I merely want to do my small part in pushing for peace in a moment in history when there’s a real chance for that piece to actually be achieved.

(03:46:52)
I may be speaking too long, I’m sorry, but I can probably speak for many more hours. So this is in fact me trying to be brief. So again, my two goals were to bring out the best in the President Zelensky as a leader and a human being, and to give him every opportunity possible to signal that he’s ready to make peace and to lay out his vision for what that piece might look like.

(03:47:16)
Like I said, step one through 10 is prepare well. I did. But step 11 is the actual conversation. They’re the specific psychological and personality quirks and qualities of the guests matter a lot. My job is to try to cut through the bullshit walls we put up as human beings and reveal directly or indirectly who the person truly is and how they think.

(03:47:40)
With Zelensky, he is a deeply empathic and emotional human who personally feels the suffering of the people of Ukraine in this war. This is a strength and perhaps also a weakness, but it is an important part of the reason why I said many times that he is a truly historic figure.

(03:48:02)
Very few leaders in recent history would be able to pull off what he did, to stay in Kiev, to unite the country, to convince the west to join the war effort to the degree they did. He is also a showman. To borrow the title of the biography, I recommended a man with many layers of humor and wit, but also ego and temper, sometimes fully self-aware and sometimes losing himself in the emotional rollercoaster of a painful memory or a turn of phrase that he can use as a springboard for a angry soliloquy.

(03:48:40)
Add to this, the fact that we didn’t agree to anything, what we will talk about or how long we will talk about it. The interview could have easily been five minutes or three hours. So I had to quickly gain his trust enough to open up and stay for a long-form conversation, but push him enough to reveal the complexities of his thought process and his situation.

(03:49:03)
This is where humor and camaraderie was essential. And I would return to it often though it was very difficult given the stakes, the heaviness, the seriousness of the topic of the war. So in this case, the approach I followed for this conversation is constant nudges and questions about peace, often using almost childlike statements or questions.

(03:49:25)
I generally like these kinds of questions. On the surface, they may seem naive, but they’re not. They are often profound in their simplicity, like a lot of questions that children ask. Remember, it was a child who pointed out that the emperor was not wearing any clothes. I like the simplicity, the purity, the boldness of such questions to cut through the bullshit to the truth. And that truth is that hundreds of thousands of people died in this war and are dying every day, and all the other problems from corruption, to suspended elections, to censorship cannot be solved until peace is made.

(03:50:06)
I give the president every single chance to signal willingness to negotiate, knowing that both Trump and Putin will listen to this conversation. I don’t think he took it and, instead, chose to speak very crude words towards Vladimir Putin. This is fully understandable, but not directly productive to negotiation.

(03:50:28)
To clarify, I have hosted many conversations that were intensely critical of Vladimir Putin from Serhii Plokhy to Stephen Kotkin. But this conversation is with a world leader speaking about another world leader during a opportunity for peace. Crude words of disrespect while powerful may harm negotiations. Peacemaking in this situation requires compromise in order to avoid further death and suffering. And I believe it requires treating the other leader with the seriousness you expect him to treat you with.

(03:51:06)
This is what I was pushing for, all that while also putting my ego aside and letting the president shine, which is necessary to accomplish both goals one and two that I mentioned previously. This is also why I wanted the president to speak about Elon and Trump to extend the olive branch for further avenues of peacemaking.

(03:51:27)
This is not about politics. It is once again simply about peace. Now, all of this, my words, my attempts were taken out of context and used to attack me by some online mobs. As an example, President Zelensky said in a mocking tone that he thinks that Vladimir Putin is simply irritated by people who are alive in Ukraine. And I answered, quote, “If you believe this, it will be very difficult to negotiate.” If you think that the president of a country is completely crazy, it is really hard to come to an agreement with him. You have to look at him as a serious person who loves his country and loves the people in this country,. And he conducts, yes, destructive military actions. The president interrupted me at this point and said, “Who are you talking about now? Who loves this country?”

(03:52:22)
And I said, “Putin. Do you think he doesn’t love this country?”And the President answered, “No.” Again, this is not a podcast conversation with a historian or activist. And I somehow out of nowhere, just for fun, [inaudible 03:52:38] poetic about Putin’s or Zelensky’s or Trump’s love of nation. It is a conversation with a world leader discussing the opportunity to negotiate peace when a large number of people are dying every single day. Even if the hard ball’s over with hate, leadership now requires sitting at the negotiation table and compromising.

(03:53:03)
This may be painful, but it is necessary. There are a few other places in the conversation where some online mobs took my words out of context and used them to call me naive and to call for more war, saying peace is impossible with a man who they claim is the second coming of Hitler.

(03:53:23)
My friends, if you make such attacks on this conversation, it is in fact you who are naive and ignorant of the facts of history and geopolitics. Peace must be made now in order for death and suffering to stop, in order for Ukraine to have a chance to flourish, and in order for the drums of a global war to stop beating, a global war that would cripple humanity. This was my goal once again to push for peace. And I will continue this effort to the best of my ability. Thank you. I love you all.

Transcript for Volodymyr Zelenskyy: Ukraine, War, Peace, Putin, Trump, NATO, and Freedom | Lex Fridman Podcast #456

This is a transcript of Lex Fridman Podcast #456 with Volodymyr Zelenskyy.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Introduction

Lex Fridman
(00:00:00)
I hope the Kyiv Airport will open soon then it will be easier to fly in.
Volodymyr Zelenskyy
(00:00:05)
Yes. I think that the war will end and President Trump may be the first leader to travel here by airplane. I think it would be symbolic by airplane.
Lex Fridman
(00:00:16)
Again, January 25th around that date, right. Flying in, meeting the Air Force One.
Volodymyr Zelenskyy
(00:00:21)
That would be cool.
Lex Fridman
(00:00:23)
There is a perception of corruption. People like Donald Trump and Elon Musk really care about fighting corruption. What can you say to them to gain their trust that the money is going towards this fight for freedom, towards the war effort?
Volodymyr Zelenskyy
(00:00:41)
In most of cases, we did not receive money, we received weapons. And where we saw risks that something could be happening with weapons, we cracked down hard on everyone. And believe me, this is not only about Ukraine. Everywhere along the supply chain, there are some or other people and companies who want to make money, they try to make money on the war. We did not profit from the war. If we caught someone, believe me, we cracked down hard on them, and we did that, and we will continue to do so because to this day when someone says that, “Ukraine was selling weapons,” and by the way, Russia was the one pushing this narrative, we always responded, “Our soldiers would kill such people with their own hands without any trial.”

(00:01:34)
Do you honestly think anyone could steal weapons by the truckload when we ourselves don’t have enough on the front lines and yet we have to provide proof to defend ourselves? Because when there’s an abundance of such misinformation, distrust starts to grow. And you’re right, people listen to various media outlets, see this and lose faith in you. In the end, you lose trust and with it you lose support. Therefore, believe me, we are fighting more against disinformation than against particular cases, although I still emphasize once again at the everyday level, such things are still important. We catch these people and we fight them.

(00:02:22)
… as if Putin wants to sit down and talk, but Ukraine does not. This is not true.
Lex Fridman
(00:02:31)
I think that yes, he is in fact ready to talk.
Volodymyr Zelenskyy
(00:02:35)
Did you talk to him?
Lex Fridman
(00:02:36)
On the phone or what?
Volodymyr Zelenskyy
(00:02:37)
How do you normally talk to him?
Lex Fridman
(00:02:39)
I don’t know. Normally by the sea, the same as with you. He invites you to the sea with me, just the three of us.
Volodymyr Zelenskyy
(00:02:45)
No, no. One of us may drown.
Lex Fridman
(00:02:48)
Who? Are you good at swimming?
Volodymyr Zelenskyy
(00:02:49)
Yes, I’m a good swimmer.
Lex Fridman
(00:02:50)
You’re a good swimmer. Well, if you think that the President of a country is completely crazy, it is really hard to come to an agreement with him. You have to look at him as a serious person who loves his country and loves the people in his country, and he conducts, yes, destructive military actions-
Volodymyr Zelenskyy
(00:03:10)
Who are you talking about now? Who loves his country?
Lex Fridman
(00:03:12)
Putin. Do you think he doesn’t love his country?
Volodymyr Zelenskyy
(00:03:16)
No. What is his country? He happened to consider Ukraine, his country. What is his country?
Lex Fridman
(00:03:22)
When do you think there will be presidential elections in Ukraine?

Introductory words from Lex

Lex Fridman
(00:03:29)
The following is a conversation with Volodymyr Zelenskyy, the President of Ukraine. It was an intense, raw and heartfelt conversation, my goal for which, was to understand and to do all I can to push for peace. Please allow me to say a few words first about language, then about the President, and finally about history. Please skip ahead straight to our conversation if you like. We spoke in a mix of languages, continuously switching from Ukrainian to Russian to English, so the interpreter was barely hanging on. It was indeed in many ways a wild ride of a conversation. As the President said, “The first of many. Language, like many other things in a time of war is a big deal.” We had a choice, speaking Russian, Ukrainian or English. The President does speak some English, but he’s far from fluent in it and I sadly don’t speak Ukrainian yet, so Russian is the only common language we’re both fluent.

(00:04:39)
In case you don’t know, the Russian language is one that the President speaks fluently and was his primary language for most of his life. It’s the language I also speak fluently to the degree I speak any language fluently, as does a large fraction of the Ukrainian population. So the most dynamic and powerful conversation between us would be in Russian without an interpreter, who in this case added about two to three second delay and frankly translated partially and poorly for me at least, taking away my ability to feel the humor, the wit, the brilliance, the pain, the anger, the humanity of the person sitting before me, that I could clearly feel when he was speaking fluently in the language I understand, Russian. But all that said, war changes everything. The Ukrainian language has become a symbol of the Ukrainian people’s fight for freedom and independence, so we had a difficult choice of three languages and faced with that choice, we said yes to all three.

(00:05:49)
To the consternation and dismay of the translators. We make captions and voice over audio tracks available in English, Ukrainian, and Russian, so you can listen either to a version that is all one language or to the original mixed language version with subtitles in your preferred language. The default is English overdub. On YouTube you can switch between language audio tracks by clicking the settings gear icon, then clicking audio track and then selecting the language you prefer English, Ukrainian, Russian. To listen to the original mixed-language version, please select the English (UK) audio track, big thank you to ElevenLabs for their help with overdubbing using a mix of AI and humans. We will continue to explore how to break down the barriers that language creates with AI and otherwise. This is a difficult but important endeavor. Language, after all, is much more than a cold sequence of facts and logic statements.

(00:06:58)
There are words when spoken in the right sequence and at the right time they can shake the world and turn the tides of history. They can start and end wars. Great leaders can find those words and great translators can help these words reverberate to the outskirts of a divided civilization. On another note, let me say that President Zelenskyy is a truly remarkable person and a historic figure. I say this as somebody who deeply understands the geopolitical complexity and history of the region. I am from this region. My parents were both born in Ukraine, Kyiv and Kharkiv, both my grandfathers too. I was born in Tajikistan and lived for a time there, then in Kyiv, then Moscow, then United States, and while I am now for almost 30 years and to the day I die, I’m a proud American. My family roots grow deep in the soil of nations that comprised the Soviet Union, including Ukraine, Russia, Belarus, and Tajikistan.

(00:08:13)
I’ve gotten to know and have spoken for hours with members of the President’s team and people close to him. I spoke to hundreds of Ukrainians since 2022, including soldiers, civilians, politicians, artists, religious leaders, journalists, economists, historians and technologists. I listened to hundreds of hours of programs that both support and criticize the President in Ukraine, in Russia, and the United States. I’ve read countless books about this war and the long arc of history that led up to it. It forced to recommend two at this moment, I would say The Russo-Ukrainian War by Serhii Plokhy and The Showman by Simon Shuster, which is a good personal behind-the-scenes biography of the President focused on 2022, but there are many, many more. This is why I can comfortably say that he is a truly singular and remarkable human being. It was an honor and pleasure to talk with him on and off the mic.

(00:09:18)
Now, it is true that I plan to travel to Moscow and to speak with President Vladimir Putin and I hope to be back in Kyiv as well as President Zelenskyy said, this was our first of many more meetings. In all of these cases, I seek to do my small part in pushing for peace. And in doing all this, I’m deeply grateful for the trust people have given me on all sides. For the people attacking me, sometimes lying about me, for the critics in the stands chanting the latest slogans of the mass hysteria machine like the sheep in Animal Farm, I love you too. And I assure you that drawing lines between good and evil on a world map is much easier than seeing that line between good and evil in every human being, including you and me. This is what I try to do. I’m simply a human being who seeks to find and surface the humanity in others, and as I’ve said, no amount of money, fame, power, access can buy my opinion or my integrity. Now, finally, please allow me to briefly overview some history to give background for several topics that President Zelenskyy references in this conversation. I recommend my conversation with Serhii Plokhy and many others about the history of the region. But here, let me start with 1991. When Ukraine declared its independence and the Soviet Union collapsed. From this point on Russia-Ukraine relations were defined in large part by whether Ukraine aligned more with Russia or with the West, meaning Europe, United States, NATO, and so on. In 2004, with the Orange Revolution, a pro-Western candidate, Viktor Yushchenko became President. In 2010, it went the other way. A pro-Russia candidate, Viktor Yanukovych became President. The internal tensions grew and in 2013 Euromaidan protests broke out over Yanukovych’s decision to suspend talks with the European Union in favor of closer ties with Russia. This set forward a chain of important events in 2014. On the politics front, Yanukovych was ousted and fled to Russia leading to the election of a pro-Western President.

(00:11:45)
Also, in 2014, on the war front, Russia annexed Crimea and war broke out in the Donbas region of Eastern Ukraine, which eventually killed over 14,000 people and continued all the way to 2022. When on February 24th 2022, Russian forces initiated a full scale invasion of Ukraine. This is when the world started to really pay attention. Now some history of peace talks. Volodymyr Zelenskyy won the presidency in 2019 and he discusses in this conversation the ceasefire agreements he made with Vladimir Putin in 2019, which was one of many attempts at peace from the two Minsk agreements in 2014 and ’15 to a series of ceasefire agreements in 2018, ’19, and ’20, all of which failed in part or in whole. All this shows just how difficult ceasefire and peace negotiations are, but they are not impossible. It is always worth trying over and over again to find the path to peace.

(00:12:55)
I believe that Presidents Zelenskyy, Putin and Trump should meet soon after January 20th this year and give everything they got to negotiate a ceasefire and security guarantees that pave the way for long-lasting peace. We discussed several ideas for this in this conversation. As I said, this was one of my main goals here, to push for peace. This trip to Kyiv and this conversation was a truly special moment for me in my life. It is one I will never forget, so to reflect I say a few more words and answer some questions at the very end if you like to listen, but here let me say thank you to everyone for your support over the years. It means the world. This is a Lex Fridman Podcast and now dear friends, here’s the President of Ukraine, Volodymyr Zelenskyy.

Language

Lex Fridman
(00:13:55)
If we can explain why the Ukrainian language is very important, our conversation will be most effective and impactful if we speak in Russian.
Volodymyr Zelenskyy
(00:14:01)
I speak Russian perfectly of course, and I understand everything you are talking about. However, I can’t respond in Russian the entire interview. It’s because this is how it is today. I’m not making anything up. You can see it all for yourself. You can feel and hear it. Today, there were 73 missile attacks against us and people were killed. There were over 100 drones today, and this is a daily occurrence. The people who attack us, they speak Russian. They attack people who were only recently told that this was actually in defense of Russian-speaking people, and this is why I respect neither the leader or director of today’s Russia, nor the people. That’s it. And I don’t think that you can just pretend that nothing’s happening and give Putin a pass once again for saying that, “We are one people, that we speak one language,” et cetera. They speak the language of weapons. That is a fact, and we are peaceful people. Peaceful people who want to protect themselves and defend their freedom and their human choice. At the beginning of the war, I addressed Russians in Russian, zero effect. They’re mute.

(00:15:33)
They do not listen. They did not listen. Some are afraid, some have other issues. They have different reasons. It’s like when a person is drowning and people walk by because they can’t hear them. And someone walks on by crying, afraid to save them. It doesn’t change anything for the one drowning. They need someone to help them. This is why I honestly despise these people as they are deaf. They began the occupation in the supposed defense of the Russian language and that’s why with all due respect, I would like to give an interview in Ukrainian. This is very important to me. If there are some points that you want me to explain in Russian, I can certainly do that. I can certainly occasionally speak Russian, but in general, no. I’m not sure that you will understand me completely, despite your Ukrainian roots, you are a citizen of the United States, right?
Lex Fridman
(00:16:39)
Yes.
Volodymyr Zelenskyy
(00:16:40)
That’s why I’m surprised that you don’t understand. Well, it was a long time ago. I understand that it was a long time ago. Moreover, a lot has changed. A lot has changed.
Lex Fridman
(00:16:58)
If I may, please allow me to say this in Russian. Yes, many things have changed, but I have hope. I hope that today many Russians will hear this, that Vladimir Putin will hear this, that the American President, Donald Trump, and the American people will hear this, that everyone will hear this. And yes, Ukrainian language is important symbolically, but what is also important is that we understand each other well.
Volodymyr Zelenskyy
(00:17:24)
For Donald Trump? Is it important for Donald Trump whether I speak Russian or not?
Lex Fridman
(00:17:28)
Yes, yes, yes. Because unfortunately, and it hurts to admit, but I cannot speak or understand Ukrainian yet, so your wit, dynamism and your humanity will not come through as well and as quickly. Remember, I need to wait for two to three seconds to hear it. You have a great sense of humor, great stories. With an interpreter translating, I simply won’t see this, but I understand that it’s painful. Another reason is that I hoped we could show that even though it is sometimes said that Russian is banned in Ukraine-
Volodymyr Zelenskyy
(00:18:03)
This is not true. I’m speaking Russian now, right. We have people who speak Russian. This is not true. Really, it’s not. It’s really not true. We disrespect Russian now because of Russians. That’s all. When they were saving Russian speakers, they killed Russian speakers. Many people who actually… Many of whom are in the East, right. They lived in the East. They destroyed their houses, destroyed their lives. It’s not a rhetorical thing. It’s not all talk and blah, blah, blah. I don’t have time for blah, blah, blah.

(00:18:38)
Yes, so it’s a very, very, very important and sensitive moment. The message is that we are not one nation. We are not the same country. We’re different countries. Yes, different countries, and I think what is most important is what we’re talking about, not how we’re speaking about it. This is what I think. You are a smart guy, so you have a lot of experience in dialogue of this kind. That’s why I think you will understand me. Yeah. Anyway, I think it is far better for Donald Trump to hear my English, not my Russian.
Lex Fridman
(00:19:21)
Your English is much better than my Ukrainian. You’re getting better and better every day.
Volodymyr Zelenskyy
(00:19:25)
That’s true. I’m a very honest guy. That’s why I will be very honest with you.
Lex Fridman
(00:19:30)
Okay.
Volodymyr Zelenskyy
(00:19:31)
Your Ukrainian is not very good, but we will work on it.
Lex Fridman
(00:19:36)
Yes. I have many flaws, that’s one of them.
Volodymyr Zelenskyy
(00:19:38)
Sometimes I can speak English. Sometimes as I understand, we can be very flexible, right?
Lex Fridman
(00:19:44)
Very flexible. Spanish, Swahili.
Volodymyr Zelenskyy
(00:19:47)
Yeah. You see?
Lex Fridman
(00:19:48)
Yeah.
Volodymyr Zelenskyy
(00:19:49)
You’re very flexible-
Lex Fridman
(00:19:50)
Javier Milei needs to understand this, so…
Volodymyr Zelenskyy
(00:19:51)
By the way, Javier understood me without any words.
Lex Fridman
(00:19:55)
The language of love, maybe.
Volodymyr Zelenskyy
(00:19:57)
Of respect.
Lex Fridman
(00:19:58)
Respect.
Volodymyr Zelenskyy
(00:19:58)
I respect him. I had a very good conversation with him. Really brilliant.
Lex Fridman
(00:20:03)
May I sometimes speak Russian and sometimes English?
Volodymyr Zelenskyy
(00:20:05)
Yeah, yes. You can use any language you like, and I think that’s a very good rule for this first meeting between us. As you said, maybe we will meet in the future for the second time.
Lex Fridman
(00:20:15)
Second and third and fourth?
Volodymyr Zelenskyy
(00:20:17)
Yeah, this is good. You can ask questions in the language you’d like and I will answer in the language I can.
Lex Fridman
(00:20:23)
Well, you said you wanted to meet by the sea at some point, so for our next meeting, let’s meet by the sea.
Volodymyr Zelenskyy
(00:20:30)
With pleasure next time. It would be much better to meet by our Ukrainian Black or our Azov Sea.
Lex Fridman
(00:20:39)
I have traveled to many cities in Ukraine, but I have never been to Odessa and everyone tells me that, and I don’t know why.
Volodymyr Zelenskyy
(00:20:46)
You have to.
Lex Fridman
(00:20:47)
Can you explain to me why everyone loves Odessa so much? What’s there?
Volodymyr Zelenskyy
(00:20:54)
What’s in Odessa? That’s how they say it. What’s there? In Odessa, we’ve got it all.
Lex Fridman
(00:20:59)
Okay.
Volodymyr Zelenskyy
(00:21:00)
Odessa. I love Odessa because of its particular temperament. People have their own accent, and there are many nationalities. There are a lot of stories, authentic, Odessa cuisine. By the way, the cuisine is very different from others. The dishes are not like any other dishes and everything is very tasty. Also, there are beautiful people, and today, you understand people very well, especially after the attacks on Odessa. You understand what the people are like, just how Odessites are, very Ukrainian, and that’s very cool. I love Odessa. I go there several times a year. I go there several times a year now because… Well, now because of strengthening of air defense systems, because of this grain corridor, et cetera. I go there more often. They have the sun there. They have the sea. It’s Ukraine, and it’s very cool there.
Lex Fridman
(00:22:10)
Well, when you come and visit me in Texas as a guest for the third time-
Volodymyr Zelenskyy
(00:22:16)
With pleasure.
Lex Fridman
(00:22:16)
Let’s do this. How about you, my friend Joe Rogan and I, we’ll go get some Texas barbecue together.
Volodymyr Zelenskyy
(00:22:25)
Who will pay?
Lex Fridman
(00:22:27)
That’s a good question.
Volodymyr Zelenskyy
(00:22:27)
Putin. Putin for everything. He has to pay.
Lex Fridman
(00:22:31)
Well, yes. We’ll invite him too.
Volodymyr Zelenskyy
(00:22:33)
No, no, no, no.
Lex Fridman
(00:22:34)
Okay.
Volodymyr Zelenskyy
(00:22:34)
Without him.
Lex Fridman
(00:22:35)
Okay, I get it. Understood.
Volodymyr Zelenskyy
(00:22:37)
But if the Rome Statute will be accepted by your government before this moment.
Lex Fridman
(00:22:46)
Okay. By the way, I don’t know if you know this, but Joe has a great comedy club in Austin.
Volodymyr Zelenskyy
(00:22:52)
Joe Rogan?
Lex Fridman
(00:22:53)
Joe Rogan, yes. And I think that as a person who respects comedy and stand-up comedy, it would be interesting for you to have a look at it.
Volodymyr Zelenskyy
(00:23:00)
No, no. He is, I know him and I saw a lot of different videos. He’s a very talented person, so it would be a pleasure if you invite me and I’m able to do it. I am a little bit busy, but if I’ll be in the United States, I hope that I will have a conversation and a meeting with President Trump, and of course during my visit, if I’ll have the time, it would be a pleasure if you’ll invite me. With pleasure.
Lex Fridman
(00:23:30)
You know what? I will pay.
Volodymyr Zelenskyy
(00:23:33)
Good.
Lex Fridman
(00:23:33)
Yeah. I had to think about it, but you are the President.
Volodymyr Zelenskyy
(00:23:37)
Yes, with you, with pleasure.
Lex Fridman
(00:23:39)
When the war is over, please come.
Volodymyr Zelenskyy
(00:23:42)
Thanks so much.
Lex Fridman
(00:23:43)
And when you’re less busy.
Volodymyr Zelenskyy
(00:23:43)
Thanks so much.

World War II

Lex Fridman
(00:23:44)
If we can go back many years, World War II, tell me the story of your grandfather who fought in World War II.
Volodymyr Zelenskyy
(00:23:52)
My grandfather, he graduated from the military academy, and from the very beginning of the war, he went to fight. He was in the infantry and he fought through the entire war. He had many wounds, as they used to say back then, “His chest is covered in medals,” and it’s true. He had more than 30? Yes, more than 30. He was the kind of man… He was such a serious man. I loved him very much, and we had a very close relationship. He didn’t like to tell details about the war. He never boasted, although I asked him as a boy would, “How many fascists did you kill?” He never talked about it. He believed that the war was a great tragedy. A tragedy for everyone. And Ukraine was occupied, and it was a tragedy for Ukraine, a tragedy for Europe, and a tragedy for the Jewish people.

(00:25:10)
His own brothers, his entire family were executed. They were tortured by fascists who had occupied Ukraine and their village. His father was the head of the village and he was killed. They were shot. It was a mass… A mass grave, right? Yes. It was a communal burial. Some of them were killed outright and others, they were buried alive. His four brothers, they all went to war. As soon as the war began, they were all there. He was the only one who had a military education, and they all died in the war. He was the only one who came back. He had nobody. He came back and he found my grandmother, his future wife, and she managed… What was it called then? I don’t know. They don’t have them anymore. It was a child care facility, an orphanage, so to speak, a place where orphans lived, children who don’t have parents, children of war.

(00:26:34)
And she managed this child care facility with difficult children, as they used to call them, difficult children who went through the war, who saw their parents killed, and this is how they met, because these difficult children, they, well, sometimes behave differently. They could steal something, do something bad. There were many, many children in the orphanage. Yes, that’s how she met my grandfather, and I loved him very much, and I think that my grandfather, frankly would never have believed that this war is possible. He would never have believed it because he worked in the police after the war. He was a colonel. He worked in a criminal investigation all his life, so he fought with bandits all his life, after the Second World War, but also I believe he fought for justice all his life. And we all lived in one apartment, and even after his death, I lived with both of my grandmothers and my parents, two grandmothers who both lost their husbands. Both of them died.

(00:28:05)
Well, it was an ordinary family, an ordinary family that lived like everyone lived back then in the Soviet Union and even after the Soviet in the nineties, we lived in one apartment all together. What else is there to say? But I think the most important thing was values, respect. They gave me an education. My parents gave me an education, no one left me money or apartments, so I didn’t inherit anything material. But I believe that our real inheritance is here, in our minds and in our hearts. I believe that. Understood.
Lex Fridman
(00:28:47)
There’s just a one-second delay, so if… I’m sorry, if you-
Volodymyr Zelenskyy
(00:28:59)
It’s fine.
Lex Fridman
(00:28:59)
… tell a joke, I will laugh about one, two or three seconds later. There’s a delay. So ordinary family, but not an ordinary time. World War II. Speaking of mass graves, I was at Babyn Yar yesterday. A large part of my family died there. In moments like this, such a place serves as a stark reminder of the profound historical gravity of the Second World War. I remember this song from my youth, “On June 22nd at four o’clock, Kyiv was bombed and the war began.” I always wondered how it would feel to live in a moment when everything changed. The path of humanity completely shifts in a single moment just like that. What do you think about that moment in 1941 now, after the 2022 invasion, how do you perceive the Second World War after you have witnessed all of it?
Volodymyr Zelenskyy
(00:30:05)
Well, firstly, the war actually started earlier. It started here in Ukraine. Kyiv was bombed as you quoted, but the war had already begun before that. And I think I perceived it as a start of the full-scale invasion. Well, I think it’s hard to understand why nobody wants to listen, look at and analyze history. War, the rise of fascism and Nazism, the emergence of Hitler, Goebbels, and their entire team at the time, this wasn’t just about one party or even one country. It was essentially a wave. A wave of hatred, a wave of one race, one race above the rest.

(00:31:17)
They were in fact constructing and ultimately implemented a theory around this idea later seizing Europe. They created a theory of one nation, one race, one world, their world. Of course, this idea is absolutely senseless, but it has become radicalized over the years and even gained support. A vision of one world and in principle, the so-called Russian World, the ideology Putin promotes and imposes, it wasn’t originally like that. He was a different person back then, or maybe he was always like this, but his rhetoric was different. At the beginning, remember, he talked about the EU, and-
Volodymyr Zelenskyy
(00:32:00)
At the beginning, remember he talked about the EU and even about Russia’s future being tied to NATO. There were even talks of joining the European Union, NATO. He spoke about shared values with the West. That’s how it all sounded back then. We must also look at Hitler, who was seriously… Before the radical idea of taking over the whole world, he actually made certain steps and everyone believed he was helping the economy. And to be fair, he did take some steps in that direction but he was a terrifying person. None of those actions justify him, nor do they excuse his actions and that’s why we cannot look at the Second World War as if it started in 1939. It didn’t begin in 1941 either. We need to draw conclusions. When did it start? With the weaknesses of the world? The division of European states, the Molotov-Ribbentrop Pact. All of this happened before 1941. People who were more informed, those who dug deeper, whether they were politicians or not, whether they were from different walks of life including business, which was different back then, we’re speaking about all of this.

(00:33:26)
Hitler won’t stop. There’ll be a world war. Hitler will destroy nations, nations, and that’s what happened. Someone looked the other way. What I told you about. Europe was sinking then, I gave you an example of it but the whole world looked the other way and didn’t pay attention and said, no, we can negotiate with him. I’m telling you he’s okay. We can negotiate with him. He’s just more right leaning, or it does not matter what they said. He’s just very pro nationalist. This is all nonsense and this is not the first time and Hitler isn’t the first such case in history. We’re dealing with a person who is allowed to stick to this desire to destroy. He was consumed by it and enjoying it. And what happened to Hitler?

(00:34:31)
Now, what about Putin? This invasion was also at around four in the morning. There were missile strikes on Ukraine. This is the same. I believe that intentions are also the same but more on that later. By the way, you tell me if this is too long, you can stop me.
Lex Fridman
(00:34:53)
Never long enough. It’s beautiful.
Volodymyr Zelenskyy
(00:34:55)
Okay, so it happened here around four in the morning. Before this, I must honestly say everyone said something, predicted something, et cetera but I asked only for one thing primarily from the United States, if you are sure, if you have the evidence, if you talk to him and he tells you that there’ll be an invasion, if all this scares you, I only asked for two things. Send us weapons or better yet, strengthen us with preventive measures so there would be no war. It wasn’t the weapons that I was asking for. I asked for sanctions. Intimidate him. Please don’t say that if he comes, if he crosses borders, if he kills, we are imposing sanctions. Well, this is complete bullshit. Sorry but really.
Lex Fridman
(00:35:53)
Oh, I understand this.
Volodymyr Zelenskyy
(00:35:54)
Oh, wonderful. Yes.
Lex Fridman
(00:35:55)
I understood one word.
Volodymyr Zelenskyy
(00:35:57)
Yeah.
Lex Fridman
(00:36:00)
So, they did not help.
Volodymyr Zelenskyy
(00:36:02)
I believe that no and this is a fact. We didn’t receive help. If we assume that words are help, well then yes, we received a lot of it because there were plenty of words. Even more than plenty, yes. At four in the morning there were strikes. Morally, is it possible to prepare for war? No. It doesn’t happen like you read in books, see in movies and so on. What happens to you? I was just looking at my wife and children. My children were asleep but my wife was awake. There were strikes, missile strikes. We heard them. To you as a living person, how can this be? You just can’t fully believe this. You just don’t understand. Why now, given everything that happened in World War II when millions of people died? None of it mattered still at four in the morning around 4:00, 3:40, 3:45. Remember, around this time? Yes, there were missile strikes and later.

(00:37:27)
By the way, a few days after the first days of the war, I spoke with Lukashenko on the phone and he apologized and he said that it was not me. Missiles were launched from my territory and Putin was the one launching them. These are his words. “I have witnesses and I apologize,” he said. But believe me, that’s what he told me. “Volodymyr, this is not me. I’m not in charge,” he told me, “I’m not in charge. These are just missiles. This is Putin.” I told him, “Don’t do that.” This was done without me. That’s it. On the phone, I remember this conversation. I told him, “You are a murderer too, I’m just saying.” He told me, “You must understand, you can’t fight the Russians.” I told him that we never fought them. I said, “It’s war. The missiles came from your land, from Belarus. How did you allow this?” Then he replied, “All right, retaliate then.” I still remember him telling me, “Hit the refinery. You know how much I care about it.” Mozyr Oil Refinery, is that it? Can’t recall. Mozyr Oil Refinery. I told him, “What are you on about? What retaliation?”
Lex Fridman
(00:39:00)
Forgive me, Volodymyr.
Volodymyr Zelenskyy
(00:39:02)
Yes.
Lex Fridman
(00:39:03)
This was at five in the morning?
Volodymyr Zelenskyy
(00:39:06)
No, no, no. This was during the first or maybe the second day. Second or third day of the war.
Lex Fridman
(00:39:11)
Ah, I see.
Volodymyr Zelenskyy
(00:39:12)
After that I went back home. I was home with my children, with my wife. I just went to my wife very quickly that night at four o’clock and just told her, “Get the children, get ready. You’ll probably need to go to my office very soon.” And I left. That’s it. At this moment, you are no longer a father. What happened to me, unfortunately, because I believe that this is, and not only do I believe I understand, especially now, that all of this is the most important thing because your country is your family. The strength is in your family and this is the most important thing. I’m the president and therefore, I had to stop being a father in my own family. And my wife had to do everything… She had to do everything regarding children, regarding safety and I had to deal with the state because I’m the president and this is my duty. And I, by the way, am taking this very seriously. I went to the office and here we are now. You’re very welcome.

Invasion on Feb 24, 2022

Lex Fridman
(00:40:31)
Well, at that moment on February 24th, 2022, everything changed again. Just like in June 1941, everything changed and history took a turn. The history of humanity took a turn and for you too, you were the president. You were talking about fighting corruption, about the country’s freedom, about interesting and innovative reforms but that morning of February 22nd, everything changed. Could you tell me about that morning, the details of your actions when you had to quickly make difficult decisions? What was the process for you? How did you make these decisions? Did you discuss them with people you trust to understand how to respond to this invasion in every technical, political and military aspect? What was the process for you? How did you make the decision?
Volodymyr Zelenskyy
(00:41:34)
According to our legislation, in principle, I am the supreme commander of the Armed forces of Ukraine, so I had to give corresponding orders. Yes, I have a military office and then later there was a military headquarters where all key people gathered. This is not only about the military, it’s about energy, et cetera, all key things. But at that moment, I made the decisions quickly and without a doubt and I cannot say that I am just that kind of person. I’m just a living person who believed that if help is needed right now to help, to help evacuate people, help with children. Several cities were blocked. I was only thinking about how to deliver food there within a day. We did a lot of things, although we understood that they in fact occupy part of our state.

(00:42:43)
We distributed weapons to people, that’s how it was. Trucks came and simply distributed weapons to people so that they could defend the capital, to ordinary people just on the street, to ordinary people who understood that if the Russians entered a city, then we would have the same thing that’s happening in other cities per the information we received. Thanks to digitalization, by the way, we had very good digitalization before this. We preserved a lot and even when they were surrounding certain cities, a lot of things still worked. The banking system, the internet, we had television, and thanks to this, I made several decisions to ensure that people are united and have all the information. Russia is very good at spreading large scale disinformation. Fortunately, I have two decades of experience managing a production studio, TV channels and large media resources. I understood that we needed to build an information network very quickly.

(00:44:08)
Thanks to this, I began to address the people constantly. This happened several times, three to five times a day. In fact, I became an information source for people who were in cities that were cut off from other information and it was very important for me to keep all things digital, to keep the internet, to stay in touch with everyone with all the people. Initially, that’s the contact we had and then we also built a media platform where we had all the news agencies of Ukraine and this network was called Marathon. It was also very important for the people to trust us and people had to receive information. Why? There were waves of Russian on the first day who said he ran away. I had to go out into the street, I left the office and went outside. I had to do this because I was showing that this was no green screen to show that it was the street, not some digital manipulation.

(00:45:25)
I did these things. Then I touched various objects. Now, people might think that these are small things, but I was actually showing that I was in a real place. All of this had an impact. I was absolutely sure of my actions. And these contacts, several contacts and then I spoke to the Russians, I addressed Russians, I really did, and then only after that I gathered… It was the first day when I invited all of the journalists here, wasn’t it? That was on the first day, I think. Well, not here, to the press center in this building. I talked to journalists. I asked them not to leave because we needed weapons. At that moment, they were handing out rifles to people and for me, journalists and media platforms were essential voices.

(00:46:22)
There were various journalists from different countries here and they were essentially stuck and I asked them for contacts, those who had access to Russians, Belarusians, Kazakhs who understood everything, the same information, and I spoke to them and I spoke to them and spoke in Russian. I told them, “You must stop Putin. This is terrible. This is horror. This is war. You must stop him. And if you stand up now, if you speak out and if you go out into the streets…” This was very important. I spoke to them in Russian to show them that there was no problem and that all of these pretexts were made up. This is why it’s so painful to talk about the Russian language too because look, if a person does not want to listen, they will not listen no matter what language we speak.

Negotiating Peace

Lex Fridman
(00:47:19)
I disagree with you here. I think and hope that many people in Russia will hear us today.
Volodymyr Zelenskyy
(00:47:27)
They blocked YouTube recently. Are you aware of this? In their country.
Lex Fridman
(00:47:32)
I know. I simply guarantee that this conversation will travel fast on the internet. Everyone will hear you. They will hear you, including the president of Russia will hear you. This is why I have hope.
Volodymyr Zelenskyy
(00:47:46)
He is actually deaf. Even if he speaks to you, he is deaf by his very nature. Do you understand the difference? For instance, when you talk to Musk, you are talking to an innovator, a scientist about rockets. You talk about how to save on costs and how they land. On the other hand, Putin doesn’t launch rockets to save money but to kill people. Do you think you can talk to Putin about technology? Your guys were interviewing him and he told them about tribal history. Do you understand? Imagine a Russian man in his country listening to him. You know what Musk is about, technology, Mars, artificial intelligence. And this guy Putin is standing there bare-assed pontificating about tribes. You’ve got to understand, you think that when you do interviews like Mr. Tucker who did an interview there that you’re about to make them friends. What does this have to do with friends? He is different. He’s simply different.
Lex Fridman
(00:49:09)
But it’s still necessary.
Volodymyr Zelenskyy
(00:49:10)
A mammoth stands before you.
Lex Fridman
(00:49:13)
By the way, I must say that when you said bare-assed, it was not translated. Could the interpreter please translate?
Volodymyr Zelenskyy
(00:49:19)
This is so that you can understand.
Lex Fridman
(00:49:21)
Now he explained everything to me, I understand.
Volodymyr Zelenskyy
(00:49:23)
Yeah, that’s great.
Lex Fridman
(00:49:24)
Now I fully understand.
Volodymyr Zelenskyy
(00:49:25)
That’s great.
Lex Fridman
(00:49:25)
Anyway, but we still need to talk.
Volodymyr Zelenskyy
(00:49:28)
One should always speak with someone who listens and you must speak when you know that this will benefit you bring peace and calm to the world, not the other way around. I love President Trump’s message. When he speaks, I think that we share a position on peace through strength. That is very important. It means that if you are strong, you can speak. And we need to be strong and Ukraine has to be strong enough, otherwise what for? Like Voldemort who must not be named. Yes, he’s like Voldemort. He thrives subsists and lives on being subjectivized instead of isolation. He has offered to step out into the light. He is darkness personified and you offer him, as it were, to be subjectivized.

(00:50:40)
Why? There’s only one reason, fear. And you say we need to talk. Listen, we need to be in a strong position and not talk but end the war. Yes, it is possible through dialogue. We’re not opposed to it. You just need to be in a strong position to make the other person want it. Do you think he wants to end the war? That’s what you suggested. I think this is naive. I’m sorry. With all due respect, it’s naive to think he wants to finish the war.
Lex Fridman
(00:51:20)
Tell you what-
Volodymyr Zelenskyy
(00:51:22)
The circumstances… Sorry for interrupting. There’s something we need. I think that President Trump not only has will, he has all these possibilities and it’s not just talk. I really count on him and I think that our people really count on him. He has enough power to pressure Putin not into wanting to stop it. No, he will not want to. To pressure him to actually stop it, that is the difference. Don’t rely on Putin’s will to stop. You won’t see it. That’s what I think. Sorry.
Lex Fridman
(00:52:04)
No, sorry. I interrupted you first. I do have what some might call a naive dream of you sitting down with Putin and Trump and negotiating a deal about a ceasefire and together, finding a path to long-term peace. I think this requires strength, requires negotiations. There are a lot of carrots and sticks here that can be used to make a real deal. Trump is very keen on making a deal and ready to negotiate.
Volodymyr Zelenskyy
(00:52:41)
Can I ask you a question?
Lex Fridman
(00:52:42)
Yeah.
Volodymyr Zelenskyy
(00:52:44)
I just really want you and I to be on the same page. It’s very important to be in the same information space. Extremely important. Let’s talk a bit about the ceasefire. Let me describe the situation to you. In December 2019 in Normandy, in Paris at the Elysee Palace, Macron, Merkel, Putin and I agreed on the ceasefire. The U.S. wasn’t there and this, by the way, was a weak point of the meeting. If you’d like, we can later discuss why they weren’t there. It’s a security guarantee thing in general. It’s Germany’s position, et cetera. We agreed on an exchange of hostages, an all for all exchange. We made a deal to exchange everyone for everyone, I think you know that. There was also a meeting that lasted many hours, a meeting where we made a deal with him. Everyone was tired. It was just the two of us in the end and I proposed a ceasefire. By the way, no one in Ukraine believed. Few believed in the ceasefire and he wanted troop withdrawal.

(00:54:05)
I calculated that if there were a withdrawal of troops from the line of contact the way Russians proposed, it would take 20 years. I proved it to him just in terms of time, square kilometers, namely the length of the line of contact or delimitation line, and we agreed on what I told him that it will not work out but I had many points because I was deeply involved in the issue. I was involved very deeply. It’s my thing in general. If I start doing something, I can’t stand there like that guy I spoke about with my ass out. I must be dressed. I must be prepared. I must be prepared better, better than anyone in front of me. You do sports, right?
Lex Fridman
(00:54:59)
Mm-hmm.
Volodymyr Zelenskyy
(00:55:00)
I practiced for many years and we know what fights are like what boxing is, what Thai boxing is. This is what I did and I loved it very much. When you step into the ring, you understand everything pretty much, and so I stepped into it and I was definitely well-prepared but he wasn’t, no. He was not deeply involved in the process. What border? Where is it? How long will it take to disengage troops? Why wasn’t he involved you want to know? Because he wasn’t going to do any of this. This is what confused me. If you are not deeply involved in the issue, it’s as if you don’t really need the result. That’s what I think. So, what happened? We agreed that there will be gas continuation, gas transit in 2019. We agreed with him. This was the security for Europe. Merkel asked me for it and this was extremely important for Germany. We agreed with him. For him it was just money.

(00:56:21)
Secondly, we agreed on an exchange. For me, this was the most important thing. For them, gas was. For me, was the people and this is a fact because I wanted to have a humanitarian advantage so that there would be further meetings that would lead to sustained peace. And third, ceasefire. Ceasefire, you spoke about. What happened? The gas contract was signed because he needed it. And by the way, he knew everything about it. As for exchange, we took the first step and exchanged the people. Regarding the ceasefire, well, they started killing us in about a month, so I called him and I told him, “We agreed on a ceasefire, didn’t we?” Well, it wasn’t a piece of toilet paper, was it? This is serious business or so it seemed. It really was serious. “Merkel, Macron, you and I, we all agreed on this together. A ceasefire is important, isn’t it?”

(00:57:43)
Not for New Year’s because everyone was celebrating New Year’s. And now, they’re offering us a Christmas ceasefire. It’s all the same. A ceasefire for two, three days just to get some praise but this isn’t a performance. This isn’t some kind of theater. No, this is about people’s lives and that’s what happened. After that, I called him a few more times. I think I only had two, three calls with him in total. I asked him for a ceasefire. He told me it couldn’t be. We will figure it out now. People from the occupied territory, Russians and separatists, they were all there together. They continued to shoot and kill our people. Yes, the front lines were quiet but they killed people. They were killing people and I kept calling him.

(00:58:35)
I called again and again, but there was nothing until after a few months the Russians stopped answering the phone. We did not have any contact since. I wanted another meeting like we had in Normandy. I wanted the next meeting. I wanted to find a solution but the Russians refused. We tried to make it happen through various European countries and not only European, but the Russians refused. They passed along some kind of bullshit, made excuses. They didn’t want it. Meanwhile, they were sending their snipers. We had evidence, living proof, even video evidence because some of them were captured back then. Those were the snipers in training. They were training them. They were training them. And later, those snipers operated in Syria and Africa. These snipers were training in our country in the east. Ukrainians were living targets.

(00:59:35)
They were shooting from the other side, killing people, women, people, children. They were shooting. It was a hunt. By the way, it was in the Russian-speaking region in the east where according to him, everyone is speaking Russian. That’s where they were shooting. Where the situation currently is the most tense. They killed people. We sent this information, sent pictures, we sent them to the UN, sent them everywhere. We worked very hard, very persistently. I met with everyone but who thought of Ukraine back then. They didn’t notice it much. They didn’t pay much attention to Crimea being illegally occupied either. To be honest, the United States of America too, everyone was somewhat silent about this issue. That’s how it was. It was like that before a full-scale war.

(01:00:30)
I want to ask you a question about the ceasefire. For example, in Mariupol today, there are American and Ukrainian journalists, and everyone will tell you who has contact now with Mariupol, who fled from there in the last minutes, just before the occupation or who was able to leave to escape after the occupation? Chernov, who won an Oscar was among them, and the journalists that left Mariupol, they are here by the way. We had a conversation. They will tell you that 20,000, 30,000 civilians were tortured and buried there. We do not know the number of victims. People who didn’t want to work with them, who refused to cooperate with them, people who went on strikes to protest, people who did not want to work with the Russians who occupied Mariupol. And this is one example just with this city. And I have a question for you. What about the millions of children?

(01:01:40)
I will ask you in Russian so that you hear this without delay, what about the millions of children over there? What if we just arranged a ceasefire without understanding what would happen next without understanding what will happen to Ukraine’s security guarantees? What about the millions of children in the occupied territories? What should I tell them? What am I to tell them? What is it I should tell them? What? Whatever? Hey, all of you over there, see ya. And those tens of thousands of people buried there, they were… Is that what we want? Are we ready to forgive them for this? We must at least take the first step. If this is a ceasefire, we must know that there is a security guarantee for the part of Ukraine under our control. We need it so that he will not come back. This is very important.

(01:02:38)
What do we say to the people who live in those territories? These are millions of people. Did you know that since 2014 in Donetsk, in the Crimea, this is happening in Melitopol as well as in Berdiansk now, they are making all these kids of drafting age go and fight? And if they don’t go, they will be killed. So, you understand what’s happening? That is why a ceasefire, everything I said, what I wish for, and I believe in President Trump’s power to use all of this information to come up with a way to make Ukraine strong. Why am I saying that? I will give you an example. President Trump will be in the same situation as I was in 2019. Precisely the same situation. I want to end the war. We want a lasting piece for Ukraine. We must do this, the ceasefire, exchange people, and then diplomatically return all territories. And we will do this through diplomacy. What will happen next with-
Volodymyr Zelenskyy
(01:04:00)
… to diplomacy. What will happen next with President Trump? If the ceasefire happens without security guarantees, at least for the territory we control, what does he get? If he manages to make a ceasefire deal, and three months later, Putin launches a new wave of attacks, what will Trump look like? What will Ukraine look like? What will everyone look like? Putin will just do it, and why would Putin do it? Because today he’s afraid of Trump. But once Trump manages, for example, to do a ceasefire deal without serious security guarantees for Ukraine, he will give a pass to Putin. Not that he wants to, no, he does not want that. I believe in what he says. But he will give Putin an opportunity, because in Putin’s head, he wants me to fight with Trump.

(01:05:04)
Putin’s plan is to end the occupation of our territory. This is in his sick head and I’m absolutely sure of this. That is why I told you don’t wait for Putin to want to stop the war. Pressure him so that he is forced to stop the war, that’s important.
Lex Fridman
(01:05:30)
It’s important to say that what you said about the children is a tragedy. War is hell, but let me say again, we must find a path to peace.
Volodymyr Zelenskyy
(01:05:38)
There is one.
Lex Fridman
(01:05:40)
What is it?
Volodymyr Zelenskyy
(01:05:41)
There is one. Before ceasefire, strong Ukraine. Strong Ukraine’s position, yes, we can speak about it with Trump. For me, we can speak about security guarantees, but a quick step is NATO. A partial membership, NATO, yes, I understand. I understand Trump’s feelings about NATO. I heard him, he’s thinking through all of it, of course.

(01:06:10)
But anyway, yes, NATO is a strong security guarantee for all the people, for us, part of security guarantee. The second part is the arms aid package, which we will not use. If a ceasefire works, nobody will use the weapons. For what? But it has to stay. But with all due respect to the United States and to the administration, not like before, I don’t want the same situation like we had with Biden.

(01:06:39)
I ask for sanctions now, please, and weapons now and then we will see. If they start it again, of course, we’ll be happy if you’ll give us more and you will stand with us shoulder to shoulder. Of course, that is right, but, but it’s different when you have weapons. Putin wouldn’t have been able to occupy so much territory.

(01:07:03)
It was very difficult for us to push him out, but we didn’t have weapons before and that is the same situation. It can be the same situation. I’m just sharing this with you, like I said at the very beginning, I want to be very honest with you and with your audience. Yes, it’s true. If we do not have security guarantees, Putin will come again.

NATO and security guarantees

Lex Fridman
(01:07:24)
To make it clear, let’s describe the idea that you are speaking about. I would like to offer you other ideas too.

(01:07:32)
But right now, your idea is that NATO accepts Ukraine, minus the five regions of Luhansk, Donetsk, Zaporizhzhia, Kherson and Crimea.
Volodymyr Zelenskyy
(01:07:43)
Just so you understand the situation, the invitation to NATO is legislatively issued to Ukraine. So to us, all those territories are still Ukraine, but NATO so far can only act in the part that is under Ukrainian control. This can be negotiated, I’m sure about that. Yes, this would not be a great success for us, but if we see a diplomatic way to end the war, this is one of the ways.

(01:08:16)
So it is, sorry, that is a start. Secondly, weapons, arms aid package, I’m not ready to discuss this publicly right now. It’s all written down and President Trump might have seen it or not, but we’ve got no secrets from him, yes. But mostly it depends on the willingness of the United States, because some of it will come from the EU, some from the United States, of course, together.

(01:08:43)
So not just from the United States, no, no, no, we need unity with this package, so the package and sanctions. Yes, sanctions, but I think it’s in the interest of all the smart people to not have Russian energy on the market in general, so he has to stop it. That’s all, it’s fine. American oil, American gas is okay. Why not? And it’s cheaper, so it will be cheaper for the whole world.

(01:09:11)
The money will go to the United States and I think he will be happy, and the president and your people will be happy, but it’s your decision. I’m just sharing. Yes, and cheaper oil. So Putin won’t have so much money for the war, and that’s it.
Lex Fridman
(01:09:27)
But this is difficult because it’s a lot. You’re saying to continue the sanctions on Russia, to accept Ukraine into NATO, I need to ask you some difficult questions about this.
Volodymyr Zelenskyy
(01:09:37)
Yes, go on.
Lex Fridman
(01:09:38)
I trust and respect your words today. Many people respect and love you in America. Trump respects you.
Volodymyr Zelenskyy
(01:09:46)
Loves me.
Lex Fridman
(01:09:47)
Oh, come on now. Remember, last time you corrected me when I said that you love Javier Millet, you said, “No, no, no, I respect him.”

(01:09:54)
So let’s not talk about love today, but could we talk seriously about guaranteeing Russia’s security?
Volodymyr Zelenskyy
(01:10:04)
Okay. Can I interview you a little? Question is what land is the war happening on and where did it start? On our soil, on our territory, international law was violated. The sovereignty of our country was violated, civilians were killed. Tens of thousands of our people were taken hostage, and everyone will tell you this happened.

(01:10:28)
This is what happened when I speak with the Global South, which is trying to balance the two sides because of the history, because of their roots and because of their shared economic interests with Russia in the past. And now, of course, when you talk to them, they are speaking a little bit like you. They’re balancing a little bit. Yeah, a little bit in between, but we will work on it.
Lex Fridman
(01:10:56)
Yeah.
Volodymyr Zelenskyy
(01:10:57)
It’s our first meeting. During the second one, you will be more on our side, but it’s just-
Lex Fridman
(01:11:03)
You’re very convincing, very charismatic.
Volodymyr Zelenskyy
(01:11:05)
Yeah, thank you. But when I speak with them, when I speak, it’s very important. Even with their balancing attitude towards the war, they all recognize that this is a war. This is not just internal conflict, this is a full-scale war that began, that Putin began. And all of them, all of them, if you talk to them, they say…

(01:11:41)
But then they all recognize that, that it’s his own big mistake, Putin’s mistake, and that he’s not right. That’s why I said, “No, no. He’s not right, and you have to begin from this.” If you begin at the middle, between Ukraine and Russia, of course, we can speak like this. You are in the middle and say, “Okay, what’s going on? There is a fight. Where is the fight?”

(01:12:10)
It’s not the fight like in Europe when Napoleon is fighting against somebody in the middle of Europe. No, this is not in the middle of somewhere of the planet. Not the planet, it’s concretely on our land. So one country with one army, one person came to another. That’s it, it’s very clear.
Lex Fridman
(01:12:38)
Again, I would like us to find a path to peace, so let us nevertheless try to start in the middle. What other ideas do you think might? You are a very intelligent person and-
Volodymyr Zelenskyy
(01:12:50)
Your Russian isn’t that good either, and I told you that this is only our first meeting.
Lex Fridman
(01:12:56)
My English is not very good either.
Volodymyr Zelenskyy
(01:12:58)
Your English is very good.
Lex Fridman
(01:13:00)
Thank you. To be honest, I’m terrible at speaking in every language. Well, there are other ideas. For instance, sorry to say this, it sounds crazy.
Volodymyr Zelenskyy
(01:13:08)
Please.
Lex Fridman
(01:13:08)
But what if both Ukraine and Russia are accepted into NATO?
Volodymyr Zelenskyy
(01:13:12)
Putin himself spoke about Russia, maybe about NATO. What you just said is very correct. What are the guarantees for Russia? It’s not like I’m even interested what happens to them. To be honest, I don’t care what will happen to them in the future after the war ends, but these are our borders and we must understand what is going on there. Well, the NATO guarantees for Ukraine.

(01:13:43)
Actually, this is also a security guarantee for the Russians. Frankly, I talked about this many times before. Sorry, I’m speaking figuratively, but as an example, if you were a father who lost his children, a grown man. A grown man, a man, an adult, and the war has ended and he never got justice for real. For example, somebody decides to freeze support. We won’t give you anything.

(01:14:16)
You can’t fight. You can’t continue, so we stop when we stop without any guarantees, without any support, without financing, without okay. And nobody is held accountable, but the man lost his children. He will not get anything. None of the killers will be in prison. All the sanctions will be removed, and he lost his children, and we have thousands of such people.

(01:14:47)
Why do you think they will not go to Russia? We’ll find a way and will not kill the Russian soldiers there or somebody there. Why wouldn’t they? It’s human nature. It’s not about us, it’s everyone. Read American writers. Always after any war, if there is no justice for people, there must be punishment for the crime, it is only justice. How come my child was taken away? The war took him.

(01:15:21)
This is very scary. And even whether it was my son, who was fulfilling his constitutional duty, or simply a missile that struck a civilian child, and if there is no justice and the killers are not punished, why wouldn’t these people come back with hate? They will definitely come back. So when we talk about NATO, NATO is not only stopping Russia.

(01:15:51)
Do not forget, NATO is stopping us too, because there will not be justice for everyone. We know that NATO does not have the right to solve certain issues with war. NATO is a security alliance, it is protection, not brainwashing. What Putin claims that this is offensive is not true. NATO is a defensive alliance, a security alliance, and it is security for Russia.
Lex Fridman
(01:16:22)
But unfortunately, there are many options for peace that don’t involve NATO inviting Ukraine as a member. Can you imagine security guarantees without NATO membership?

(01:16:35)
For example, if America simply leaves NATO, I believe there is a high likelihood that Donald Trump would do such a thing.
Volodymyr Zelenskyy
(01:16:45)
I think it’s very bad for NATO. That’s the end, that is, that’s the death of NATO. It is a pity, because I think that it’s a very good alliance. Maybe not everything’s good there from the bureaucracy or money, et cetera, but totally countries who are in NATO, they don’t fight.

(01:17:07)
There is no war on the land of any of these NATO countries. I think that is the answer. It works or not. It works politically or militarily, I don’t know, but it works. So without Trump, without the United States of America, there will not be NATO. That is the first, so and you say, “Can we imagine that?” That what?
Lex Fridman
(01:17:32)
That there could be security guarantee without.
Volodymyr Zelenskyy
(01:17:34)
No, we don’t need guarantees without the United States. That’s it, because the United States is a very strong, powerful country. The United States puts the point. Of course, Putin said that it’s just the Soviet Union where, by the way, Ukraine was the second strong republic militarily, yes, by the way. But he, of course, always forgets about it.

(01:18:01)
But during the World War II, without help of the United States, support of your troops, support of your industry. Industrially, militarily, without your money, without your people, Hitler could win, so the United States helped a lot. Of course, Europe, USSR, and of course, everybody fought. Everybody did a lot.

(01:18:29)
But without the United States, it couldn’t be such. I don’t use the word success, because I think that there is no war which ends successfully because this is a war. Seven figure losses, heavy losses in World War II, millions of people, and that’s why without the United States, security guarantees are not possible.

(01:18:53)
I mean these security guarantees, which can prevent Russian aggression. Of course, we have security guarantees bilaterally with some countries financing support of our internal military and defending, and humanitarian issues and demining, which is very important, in helping our children in the school networks. By the way, this is a very sensitive point.

(01:19:18)
How many bomb shelters? How many bomb shelters we built with the partners for the children? And it’s a pity that they are underground, but can you imagine their eyes when they came after COVID? You understand what does it mean COVID, but they had COVID and the war, and together they didn’t see each other for so many years.

(01:19:39)
And when they saw each other even underground, they were very happy and smiling. So we have such security guarantees, but it’s not enough to prevent. Yes, preventive measures also work to prevent the aggression of Putin.
Lex Fridman
(01:20:00)
Your English is better than my Russian. This is wonderful.
Volodymyr Zelenskyy
(01:20:05)
I’m not sure.
Lex Fridman
(01:20:06)
I’m just giving you a compliment.
Volodymyr Zelenskyy
(01:20:07)
Thank you. No, no.
Lex Fridman
(01:20:08)
You’re supposed to do that kind of thing to a president.
Volodymyr Zelenskyy
(01:20:10)
Thank you so much.

Sitting down with Putin and Trump

Lex Fridman
(01:20:11)
Okay. Once again, without NATO guarantees, I have a dream that, let’s say, on January 25 or sometime at the end of January this year, you will sit down with Donald Trump, with Vladimir Putin.

(01:20:28)
And together negotiate a ceasefire with strict security guarantees and an agreement will be signed. What will this look like without NATO?
Volodymyr Zelenskyy
(01:20:41)
I will make it clear. So first of all, I think January 25th or some other day, well, you just call it January 25th and I don’t mind, it’s my birthday. And we sit down, first of all, with Trump. We agree with him on how we can stop the war, stop Putin. It is important for us to sit down with him.

(01:21:16)
Secondly, it is very important for us that Europe, which is very important for us because we are part of Europe. And not only geographically, geopolitically, but also in the European Union where we will be. For us, it is very important that Europe also has a voice. It’s the second thing. It won’t be long because Europe will be looking at us and we’ll be looking at Trump.

(01:21:44)
And by the way, I now see that when I talk about something with Donald Trump, whether we meet in person or we just have a call, all the European leaders always ask, “How was it?” This shows the influence of Donald Trump, and this has never happened before with an American president. I tell you from my experience, this also gives you confidence that he can stop this war.

(01:22:15)
That is why we and Trump come first and Europe will support Ukraine’s position, because they understand that Ukraine has every right to have its voice heard in this because we are at war. Trump and I will come to an agreement. And then if, and I am sure that he can offer strong security guarantees together with Europe, and then we can talk to the Russians. That’s right.

(01:22:46)
Not just three of us sitting down at once, and you still talk to me like that? Do you know how? As if Putin wants to sit down and talk, but Ukraine does not. This is not true.
Lex Fridman
(01:23:02)
I think that yes, he is, in fact, ready to talk.
Volodymyr Zelenskyy
(01:23:06)
Did you talk to him?
Lex Fridman
(01:23:07)
On the phone or what?
Volodymyr Zelenskyy
(01:23:09)
How do you normally talk to him?
Lex Fridman
(01:23:10)
I don’t know. Normally, by the sea, the same as with you. He invites you to the sea with me, just the three of us.
Volodymyr Zelenskyy
(01:23:17)
No, no. One of us may drown.
Lex Fridman
(01:23:19)
Who? Are you good at swimming?
Volodymyr Zelenskyy
(01:23:20)
Yes, I’m a good swimmer.
Lex Fridman
(01:23:21)
You’re a good swimmer. Well…
Volodymyr Zelenskyy
(01:23:25)
And I would like to add that if you have any contact with him, I just want to hear what happens then.
Lex Fridman
(01:23:32)
I have never talked to Vladimir Putin, but I have a feeling that he is ready, because Donald Trump is ready. I hope you’re ready, and this is not just a feeling, but a dream.

(01:23:46)
I have a dream here that the three of you will get together in a room and make peace, and I want to understand what it looks like. What security guarantees look like that will satisfy Ukraine, that will satisfy Russia.
Volodymyr Zelenskyy
(01:24:04)
Ukraine needs security guarantees. First and foremost, we are in danger, that is why they are called so. This is no joke to me. Let’s take a few steps back. Interesting, why are security guarantees a strong position of Ukraine, strong weapons and so on, so important? I will give you a little history lesson, although I think you have prepared yourself and know everything perfectly well.

(01:24:38)
You can correct me on that. Yes, Ukraine had security guarantees. The Budapest Memorandum, nuclear weapons are the security guarantees that Ukraine had. Ukraine had nuclear weapons. I do not want to characterize it as good or bad. Today, the fact that we do not have them is bad. Why? Because this is war.

(01:25:01)
Today we are at war because you have unleashed the hands of a nuclear power. A nuclear power is fighting against us, against Ukraine, and doing what it wants. By the way, even you are now talking about ceasefire, just a ceasefire. Maybe give flowers to Putin, maybe to say, “Thank you so much for these years. That was a great part of my life.” No, we are not just ready for this.

(01:25:33)
Why? The Budapest Memorandum, nuclear weapons, this is what we had. Ukraine used them for protection. This does not mean that someone attacked us. That doesn’t mean that we would have used it. We had that opportunity. These were our security guarantees. Why am I talking about this in detail? Because if you take the Budapest Memorandum, by the way, I discussed this with President Trump.

(01:25:58)
We have not finished this conversation yet. We will continue it. Regarding the Budapest Memorandum, the Budapest Memorandum included security guarantees for Ukraine at first, three, three. The most important security guarantors for Ukraine, three strategic friends and partners of Ukraine. This was in agreement. United States of America, Russia, Britain, France, and China joined.

(01:26:28)
There were five states that these are not even security guarantees. We now understand that this is not a guarantee of security, because on the one hand, these are security guarantees, but there was an English word, as far as I understand, assurance. It is translated as assurance, assurance, right? And in Russian, it’ll be and what assurance?

(01:26:59)
That is give up nuclear weapons because you were under pressure of the US and Russia for Ukraine to give them up. These two powers were exerting pressure. These two states negotiated to ensure that Ukraine does not have nuclear weapons. Ukraine agreed. These are the largest states. This is the nuclear five that does not even provide security guarantees.

(01:27:27)
Now we just need to find these people, and we just need to put in jail all of those who frankly invented all this. So confidence, assurance, assurance that Ukraine will be territorially integral with its sovereignty. It was a piece of paper. If you are curious, by the way, that after occupying part of our Donbas and Crimea, Ukraine sent diplomats three times.

(01:28:04)
I don’t think I remember, three times within a few years, we sent letters to all security guarantors, to all members of the Budapest Memorandum. What did they send? That what was written on the piece of paper, consultations, Ukraine holds consultations. If it’s territorial, integrity is violated and everyone should be in consultation. Everyone must come.

(01:28:32)
Everyone must meet urgently, USA, Britain, Russia, France, China. Did anyone come, you ask? No. Did anyone reply to these letters, official letters? They are all recorded by diplomats. Did anyone conduct consultations? No, and why not? They didn’t give a fuck. This is understandable in Russian, that as Russia didn’t give a damn, neither did all the other security guarantors of the Budapest Memorandum.

(01:29:08)
None of them gave a damn about this country, these people, these security guarantees, et cetera. We take a break, this will be a Budapest Memorandum. The last time with me, imagine how many years it was with me in February 2022? In February 2022, the war began, a full-scale war. Letters for consultations have been sent. No one answers. Next, we are taking a break from the Budapest Memorandum.

(01:29:46)
The question is simple about Budapest. Can we trust this? No. Whichever country out of these five sat at the negotiating table, just a piece of paper. Believe me, we will save you. No. Another, this is a train. This is a train with wastepaper, with security guarantees, which Ukraine has been riding for many years. The second car on this train is the Minsk agreements.

(01:30:21)
The Normandy Format and the Minsk agreements where it was written, where the signatories were, the United States of America was no longer there. I understand that Obama was here at the time, and as far as I know, I think they were simply not interested in what happened to Ukraine, and where it was in general, where it was located. Well, somewhere there, part of something.

(01:30:43)
People, well, people and let it be, let it be with these people. The United States simply did not participate. In the Minsk agreements, there are no claims to the US because they were not guarantors. Where is the claim? A step back. 2008, Bucharest, everyone has already learned from the Budapest Memorandum. Bucharest, 2008.

(01:31:14)
Bucharest, Mr. Bush, President of the United States, Republican says that Ukraine should be in NATO. This is the voice of Republicans. Check it out, Ukraine should be in NATO. Everybody’s looking at the US always, all in favor. Who is against? Merkel, so she opposes and she forced everyone not to give Ukraine an invitation to join NATO because that would be a step.

(01:31:46)
Seriously, Republicans were in favor, the US was in favor, because Republicans and Bush were not afraid of anyone. They were not afraid of anyone, and they knew that Ukraine rightly wanted to join NATO. She chooses so. And what is the question? Well, people made their choice. Well, and the Russians will not look that way. That was not the case then. Why? Because the Russians were different.

(01:32:16)
Next, Minsk, we didn’t succeed. After the Minsk agreements, as I told you, hundreds of meetings were held. I have had hundreds of meetings since 2019. We could not think about a ceasefire. A ceasefire is our offer, this is not somebody’s suggestion. This is mine, I would like. I wanted to. In Ukraine, society was divided. Not everyone wanted to, half did not want to.

(01:32:52)
Half were against, half were in favor. Some of them shouted, “Do not believe it.” Some of them shouted, “Believe it.” I am the president of Ukraine. I was given a mandate of trust by 70% of the population to take appropriate steps and I made them. This is not a joke, we’ll just sit the three of us. I am simply telling you what is. This is how can I tell you?

(01:33:19)
These meetings must be serious and prepared, and prepared with those who want peace. Ukraine wants peace, US wants peace. We have to sit down with Trump, and that is 100%, first and foremost, number one. Moreover, he told me on the phone that he is waiting for us to meet, and there will be an official visit. And my visit would be the first or one of the first to him.

(01:33:48)
And for him, this topic is very important. I know that he has his own matters, American issues, I understand. I heard his election program, but regarding international affairs, I think our issue is one of the most pressing issues for President Trump. Therefore, I believe very much I trust his words, and I hope we will meet again. We need to prepare.

(01:34:11)
We have many plans to build on, and they exist and they are supported by many countries, but we need his vision. He needs to look at all these details, but his vision, please, because he can stop Putin because Putin is afraid of him. That’s a fact, but Trump is a president of a democratic country, and he does not come for life. He is not Putin. He will not come for 25 years.

(01:34:43)
He will come for his term. Please, tell me. Well, for example, he came for four years and for the fifth year, Putin came with a war. Will it make Trump feel better that there was no war during his time, and that Ukraine was destroyed after him? Why destroyed? Putin is whoever, a killer, whoever, but not a fool. He will be prepared. He knows our mistakes.

(01:35:17)
He understands how we defeated his army after the invasion began. He realized that this was not a Soviet war, and that this would not happen with us. He will prepare. He will let everything into arms production. He will have lots of weapons, and there will be a very large army. And you think that after such humiliation, four years without a war, he did not finish us.

(01:35:42)
He will return and fight only against Ukraine. He will destroy everything around. And if you say there is a risk that President Trump will withdraw from NATO, for example, this is a decision of the United States. I’m simply saying that if it does, Putin will destroy Europe. Calculate the size of army in Europe.
Volodymyr Zelenskyy
(01:36:01)
Calculate the size of army in Europe. It’s just that I say it for a reason. Do the calculation. Why did Hitler conquer all of Europe then? Almost. Just count, remember his armies of millions. Calculate what Europe has. What are the largest armies? We have the largest army. The Ukrainian army is the largest in Europe. The second place after us is four times smaller than us.
Lex Fridman
(01:36:33)
France?
Volodymyr Zelenskyy
(01:36:33)
Yes. 200,000. I think the French have about 200,000. We have 980.
Lex Fridman
(01:36:41)
So this powerful coalition of European nations.
Volodymyr Zelenskyy
(01:36:44)
That will not be enough.
Lex Fridman
(01:36:45)
Yes, it’s not going to be enough. But you’re a smart man. There’s a lot of ideas. Partnerships with global South India, Middle East, Saudi Arabia, [foreign language 01:36:57] partnerships, political partnerships. It all protects you.
Volodymyr Zelenskyy
(01:37:01)
First of all, look at one example. North Korea. Just look at this example. 12,000 has arrived. Today, 3,800 killed or wounded. They can bring more. 30-40,000 or maybe 500. They can bring many people. Why? Because they have order, autocracy and everything. Can Europe bring people together? No. Will Europe be able to build an army consisting of two to 3 million people? No, Europe will not want to do this. And for what? We definitely don’t want a world war with you. There is no such purpose. There is no such purpose as gathering everyone.

(01:38:07)
We do not want any war. We want to stop the Russians and they invite North Korean soldiers. Invited. Their faces are burned. They themselves burn their faces. Those who cannot escape, injured or killed. There’s a video. Everything I’m telling you, there is evidence of this so that they are not recognizable. Right? What does it mean? It’s out of values which share Europe. Europe counts. It means that those guys, they don’t count. [foreign language 01:38:54] It’s count. Yes. They don’t count the number of people. That is the answer.

(01:38:58)
Can they move more? Yes. Can they move dozens of thousands? Yes, because we see what they have. Last year, for example, Europe gave us 1 million artillery rounds. We produced a lot ourselves, but they gave us initiative. It was initiative. 1 million artillery rounds and of 155 and et cetera. We produce more but North Korea gave Putin 3.7. Just gave him. So he also has a deficit for today. It means he needs what? He needs time.

Compromise and leverage

Lex Fridman
(01:39:46)
But the number of soldiers…
Volodymyr Zelenskyy
(01:39:48)
Yes.
Lex Fridman
(01:39:49)
And the number of artillery rounds is not everything. As you have said. Let’s say Donald Trump guarantees security for four years. You can form partnerships with India, with Saudi Arabia that enforce punishment to stick on oil prices, for example, if any aggressive action is taken. You can actually even build the… I’ve met a lot of incredible Ukrainian tech people, IT people. You could build great companies that form partnerships with the United States, that form partnerships with China, and that is a big leverage against aggression of however many million artillery rounds. And that, a sheet of paper, you don’t need a sheet of paper of protection.
Volodymyr Zelenskyy
(01:40:41)
Ah, that’s you. Well, when you speak in English. You don’t even need answers because when you now are talking, you already answered on all the questions. The first one is that during this time you need just cooperation. A lot of money for this military industry. In Ukraine or in Europe, with India, Saudi Arabia, and the United States, you need a lot of money. So the question is where you will get it. My answer was to Trump. I said this is one of the security guarantees. Take 300 billions of frozen Russian assets. We will take it. Take money, what we need for our interior production, and we will buy all the weapons from the United States.

(01:41:35)
We don’t need gifts from the United States. It will be very good for your industry, for the United States. We will put money there. Russian money, not Ukrainian, not European. Russian money, Russian assets. They have to pay for this. We will put it and we will make it. This is one of security guarantees. Yes, of course. Because this is a military guarantee. Yes. But then the second you said that energy price and a lot of sanctions on products and the Russian shadow fleet and etc. That is the second answer we spoke about before. Yes put more sanctions on them. More sanctions. It’s okay, but not to take off sanctions.
Lex Fridman
(01:42:20)
It’s okay with you, but it’s not going to be okay with the president of Russia.
Volodymyr Zelenskyy
(01:42:24)
Yes, but I’m not thinking how it’ll be very good for him. He’s still a killer.
Lex Fridman
(01:42:29)
I understand, but unfortunately the reality is that a compromise is needed in order to reach an agreement.
Volodymyr Zelenskyy
(01:42:35)
So in your understanding, the fact that he is not in jail after all the murders, he’s not in jail assuming all the murders and no one in the world is able to put him in his place, send him to prison. Do you think this is a small compromise?
Lex Fridman
(01:42:50)
This is not a small compromise. And to forgive him will not be a small compromise.
Volodymyr Zelenskyy
(01:42:55)
To forgive. No one will forgive. It is absolutely impossible to forgive him. We cannot get into the head and soul of a person who lost their family. Nobody never will accept this. Absolutely impossible. I don’t know. Do you have children?
Lex Fridman
(01:43:09)
No, not yet. But I would like to.
Volodymyr Zelenskyy
(01:43:11)
Yes. God bless. And this is the most important thing in life, and they simply took away the most precious thing from you. Will you ask who ruined your life before going to rip their head off? I’m just curious. They took your child away. Are you going to ask who did this? And they will answer that that dude did this. You will say, “Oh, well then there are no questions.” No, no, no. You will go fucking hell and bite their head off and it will be fair. Can murderers be forgiven? That’s why you make security guarantees. What I told you, for those who are here and what we control and what will not happen.

(01:43:52)
And that those who lost, we will never forget in a matter of time. But when you gave us NATO, I just said this means that after a while, everything I said about NATO. After a while, Ukraine will not go against Russia and Russia will not go against Ukraine because you are in NATO. I am just saying is not that a compromise? NATO is a compromise. This is not just a security guarantee, in my opinion. Look, when rockets were attacking Israel and Israel is not in NATO. NATO countries, aircrafts were deployed. The air defense worked, operated by different Middle Eastern countries.

(01:44:44)
These are also security guarantees. And by the way, Israel has nuclear weapons. So why do they need NATO when in fact they have more than NATO has? The American, British and French aviation stepped in. There was ADA. I don’t remember from Jordan. Listen, thousands of missiles were shot down that way. What is this? So it’s a guarantee of safety. It’s just that it’s not called NATO. Is some Uncle Vova irritated by the word NATO? There’s a problem with the word and I think he’s simply irritated by people who are alive and living here.

Putin and Russia

Lex Fridman
(01:45:33)
If you believe this, it will be very difficult to negotiate. If you think that the president of a country is completely crazy, it is really hard to come to an agreement with him. You have to look at him as a serious person who loves his country and loves the people in his country. And he conducts, yes, destructive military actions.
Volodymyr Zelenskyy
(01:45:55)
Who are you talking about now? Who loves his country?
Lex Fridman
(01:45:56)
Putin. Do you think he doesn’t love his country?
Volodymyr Zelenskyy
(01:46:00)
No. What is his country? He happened to consider Ukraine, his country. What is his country? Explain it. Tomorrow he will say that it’s America. No pity for the Chechens. Do they look like Russians? Do they speak Russian? Of course, they learn in schools like anywhere there’s been Russification. Who are the Chechens? A different people, another faith, other people, another language. A million eliminated. And eliminated how? How did he kill them? With love? I know fuck by hugging. In Ukrainian, as we say, “Strangling by hugging. I love you so, so much. I love you so much that I want to kill you.” That’s his love.

(01:46:55)
And that’s not love. You are mistaken. He does not love his people. He loves his inner circle. It’s only a small part of the people. He doesn’t love them. Why? I’ll explain. You cannot send your people to another land knowing that they will die. Children, my daughter, she is 20 years old. For me, This is a child. She’s already an adult of course, but she’s a child. The boys he sends are 18 years old. They are children. He sends them. It’s not that fascists came to his land and he needs to defend it. He came to ours and he sent them Chechnya. He sent them Syria, he sent them Africa. He sent them Georgia.

(01:48:06)
He sent them Moldova. Transnistria, that was before him. Fine, we can leave that aside. He has enough sins of his own. And then there’s Ukraine, the largest part. 788,000 killed or wounded Russians. He calls them all Russians. Even those who don’t know how to speak Russian on his territory of Russia, everything they’ve enslaved. Yes. Proud [foreign language 01:48:48]. So I wonder, is that love? What love is this? And for what? Does he love his people? No. Does he love his land? His country is bigger than America. How much land do you need? America is huge. America is simply an outstanding country. Outstanding country.

(01:49:09)
Russia is bigger. Well, just bigger. So ask yourself, does he love them? What is he doing and what does he love? Do you think he’s been everywhere in his Russia? It’s impossible to get around it. He hasn’t been everywhere. He just hasn’t.
Lex Fridman
(01:49:32)
Well, I believe that Donald Trump loves America and I don’t think he has been to every single American city.
Volodymyr Zelenskyy
(01:49:39)
No, no, no. I saw his rallies. So many rallies. No, no. Let’s be honest. Let’s be honest. He had it and I saw it and it’s very difficult. I mean he’s not 18. Yes, but he’s strong and this is his will. Everywhere where the war is, I’m sure, I pray to God it never will be on your land. Yes. And I’m sure that it will not be, but I’m sure that if you have in some region the problems, how to say, earthquake, hurricane you have it all. Well, I’m sure that President Trump would be there after one day, two or three days. I don’t know the security of all these things, but he will be. Otherwise, how will people look at him?

(01:50:29)
Yes, of course he will. Of course the same about me. I’m not comparing myself with him. I’m just where it is difficult for people, I have to come. The next question is very simple. Region, Kursk region. The operation there. Was Putin in Kursk during four months? No.
Lex Fridman
(01:51:00)
Listen, I have tremendous respect for you, admiration for many reasons. One of which is you stayed in Kiev and another one is that you visit the front and you talk to the soldiers in the front and you talk to people all across Ukraine. Absolutely tremendous respect for that. And not enough people say that. I had a conversation with Tucker Carlson for example, and I said that, “You’re hero for staying in Kiev.” And he said, “Well, he just did a thing that every leader should do.” But I think not enough leaders do the thing that every leader should do. So tremendous respect. I agree with you totally.

(01:51:44)
Yes a leader should go to the front of a war. That said, America has waged wars all across the world. The war in Afghanistan and Iraq costs $9 trillion and killed over a million people. War is hell. And just because war is waged in terrible ways that it is does not mean the leader does not love their country. But I take your point. I once again have a dream that even if there’s hate, that you sit down with Donald Trump and Vladimir Putin and you find a way to peace. Let me ask you a question. What do you think?

(01:52:33)
Will there ever be a day when the Ukrainian people forgive the Russian people and both peoples will travel back and forth again and marry each other, rekindle and form friendships? Will there be such a time in the future?
Volodymyr Zelenskyy
(01:52:47)
I think history has long answered this question. I don’t know how be for us. It’ll be in the future without a doubt. History has shown this time. And again after every devastating war, one generation, one country recognizes that it was an aggressor. And it comes to realize this is impossible to forgive. This is precisely the kind of education they’ve had in Germany for many years. Even though these children had nothing to do with it. It was their grandfathers who participated and not all of them were participants of Nazi Germany’s war essentially against the world. Yes. And against life. And therefore they’re still apologizing.

(01:53:54)
Apologizing is not easy. They know that they were the aggressors, they were guilty. They do not look for compromise in history. Compromise in itself buys time. And they understand this. There are convicted murderers condemned both historically and by their own people. Reparations have been paid and security guarantees have been established by the way. And all this is done. And when all this is done and recognized, in any case, people develop relations with each other. That’s clear. But it can only happen the way it always has, always has in history. Russia will have to apologize. It will.

(01:54:45)
This will happen because they are guilty. They’re guilty. And as I told you, the guilty are different. Both those who participated and those who remain silent because silence is also about participating, in my opinion.

Donald Trump

Lex Fridman
(01:55:07)
Can I ask about Donald Trump? We’ve already mentioned him a lot, but let’s focus there. What do you admire? What do you respect about Donald Trump? And also maybe why do you think he won overwhelmingly the election in 2024, that American people chose him?
Volodymyr Zelenskyy
(01:55:28)
He was stronger. He was much more stronger than Kamala Harris. Biden first and then Kamala Harris. Yes. He showed that he can intellectually and physically. It wasn’t important point to show that if you want to have a strong country, you have to be strong. And he was strong. And this number of rallies, what I said is not a simple thing. He showed that he can. He’s strong. So he doesn’t have any questions with his, I mean this age and et cetera. Nothing. He is young. He is young here and his brain works. So I think it’s important, very important. And of course a lot of interior questions.

(01:56:15)
I understand the prices and et cetera. Economic questions and you have the questions with other things.
Lex Fridman
(01:56:24)
Immigration.
Volodymyr Zelenskyy
(01:56:26)
A lot of things. I understand. So maybe he answered on those questions, which people had.
Lex Fridman
(01:56:34)
One of the questions-
Volodymyr Zelenskyy
(01:56:35)
That he will finish the war.
Lex Fridman
(01:56:37)
That he will finish the war.
Volodymyr Zelenskyy
(01:56:38)
For me, this is the main question, but I said that for him, he’s the President of the United States. For him, his priority is his questions in the United States. And I understand and I respect it, but the second he was speaking about the world, yes, he said that he will finish the war. And I hope very much because I think that our people really support his idea. And that’s why I said it is for me. It’s very, very important to have enough people around him who will have connections with him, with the right things.

(01:57:21)
For me, the truth is very right things. What’s going on really in the battlefield, what’s going on really with Putin and Russia, what he really wants and that is just to have it. Before any decision, you have to be at the same level of information. Really we need him to know everything from us, from you, from people in Ukraine, from people around who are really afraid. Afraid that Putin doesn’t want to stop the war, afraid that he will come back with his aggression.
Lex Fridman
(01:58:07)
So first of all, I should mention that our conversation today will be translated and dubbed into Ukrainian, English, Russian, other languages, Spanish. So you’re in your voice. So there are great guys originally from Poland. It’s a company called ElevenLabs. They’ve trained an AI. Artificial intelligence sounds truly remarkable in your voice. You have the freedom to speak in any language you choose, but no matter what, you will always find yourself returning to speaking in Ukrainian. That is, when you talk about Donald Trump, you can do it in Ukrainian or Russian.
Volodymyr Zelenskyy
(01:58:49)
Everybody understands.
Lex Fridman
(01:58:50)
Everybody understands. But you said that there’s some things about the war that maybe Americans don’t understand. So we talked about Putin, we talked about the security guarantees, the reality of war, what’s happening on the ground? What do you think that people should understand?
Volodymyr Zelenskyy
(01:59:13)
First of all, they have to understand the idea of Putin’s war. It is very important for him. I consider this process. I think it is very important for him not to give Ukraine independence. To prevent Ukraine from developing as an independent country for him, influence. Influence on Ukraine cannot be lost. And four, for him, it is like… I think for him, this is such a goal in this last mile and certainly for him, the last mile and of his political life. And I think that this is the goal for him. The second story, I do not want to talk about these banalities that he wants to return all the territories of the Soviet Union influence over them. He does this little by little.

(02:00:22)
I just don’t want to… People need to know details. For example, Georgia, which was headed towards the EU and NATO completely turns towards Russia regardless of the fact that they have frozen conflicts. They have in Abkhazia what we have with Donbas, which is controlled by militant rebels. Abkhazia is not developing. It’s just a very beautiful part of Georgia that has died. And if you have the opportunity, then go there someday. You will understand. It simply died because Putin wanted to. He wanted not to allow them to develop because a frozen conflict means that you will not be accepted in the EU and certainly will not be accepted into NATO.

(02:01:02)
Because right now, yes, they do not take you because of a frozen conflict. And this is what Putin did. It’s very important for him not to lose this influence. That is he turned back Georgia, young people, students, everyone leaves. And this is a fact. Georgia is quite small and they will leave. They want to live in Europe. They want to develop. Somebody in the United States, somebody in Europe, somebody in the EU, somebody in Britain. He will now fight for the Moldovan parliament. This is his second step. You will see in April what happens. You will see he will start turning Moldova away from Europe.

(02:01:42)
Although they want to go there, he does not care. They will be a pro-Russian party and they will do something with the current president because she has won the elections. She is pro-European but he will turn this back. The next steps are completely clear. He will do everything wherever he has lost influence, where there was influence of the Soviet Union. He’ll turn it back as much as possible. And we understand at what price. You have seen Syria, you saw these tortures. What we saw in [inaudible 02:02:19], what we saw everywhere we came and where our territories were occupied.

(02:02:24)
In Syria, the same happened. There were a thousand people there. And you have seen it. Scientists were found. Doctors were found. It’s clear that any people are capable of generating their own opinion. So their skills developed society, everyone who can express an opinion, everyone who can shape the independence and maturity of society. Such people are not needed. And he wants this in Ukraine. And therefore everyone should understand that Ukraine is like a large wall. From that Europe, and if God willing, President Trump does not withdraw from NATO. Because again, I believe that this is the biggest risk.

(02:03:15)
I think two steps. Two steps that Putin would like to see is a weak NATO and this without Trump. And a weak Ukraine, which cannot survive on the battlefield, simply cannot survive and prevent me from building a strong relationship with Trump. I think these two steps, leaving NATO and Ukraine’s weakness will lead to a large-scale war, which Putin will wage on all the territories of that post-Soviet Europe. I mean Soviet Europe, not post-Soviet, but post-World War II period. That is Soviet era Europe, in order to completely control everything there. This is what he will do. And besides this, this will happen in any case.

(02:04:19)
Even if the US is thinking about leaving NATO, this war will affect the United States because North Korea is the first sign. North Korean skills, North Korean knowledge, which they are now gaining from this war. These include mastering new technologies, large-scale drones, missiles, how it works, the kind of technological war we have today, cyber war, etc. All these skills, Korea will bring home and scale up in that region. And this will be a risk for the Pacific region. Security, first and foremost. For Japan and for South Korea, they will face these risks a hundred percent and it will be clear that Taiwan will also have to face them.

(02:05:18)
Without this, it is impossible. This is already happening. This is already happening. Therefore, I think that President Trump has all power to stop Putin and give Ukraine strong security guarantees.

Martial Law and Elections

Lex Fridman
(02:05:40)
We’ve been talking for two hours. Have to pause. You want to take the break?
Volodymyr Zelenskyy
(02:05:45)
We will make a pause. We can have coffee, right? Coffee.
Lex Fridman
(02:05:50)
Let’s do it. And give the interpreter, he’s struggling.
Volodymyr Zelenskyy
(02:05:57)
Some water.
Lex Fridman
(02:06:00)
We keep switching languages
Volodymyr Zelenskyy
(02:06:01)
Like a dragon. Three heads, three translators.
Lex Fridman
(02:06:05)
So one of the difficult decisions you had to make when the war began is to enact martial law. So when you won the presidency, you were the warrior for freedom. In fact, this war is for freedom. For freedom of the individual, freedom of speech, freedom of religion, freedom. But a lot of freedoms had to be curtailed, sacrificed in this fight because there’s so much focus on the war. Do you see the tension? Do you feel the tension of that, the sacrifice that had to be made in democracy, in freedom, in fighting this war?
Volodymyr Zelenskyy
(02:06:52)
In any case, this war is for our freedom. Generally speaking. To be honest, when you understand, over time when the war passes, you understand that your main values are at home. This is your home, your children, your love, God willing, parents are alive and if not alive, then their memory, visiting their grave, choosing how to work, how much, preferably choosing where to work. All this is freedom. Freedoms are not just a desire, they are an opportunity. In any case, you are right because war is a limitation of opportunities. In any case, you fight for these opportunities. Your parents and God gave you life, right?

(02:07:57)
You fight for your life. Your life. But we need to understand that first there is a war. And then-
Volodymyr Zelenskyy
(02:08:00)
But we need to understand that first there is a war and then martial law is introduced. Martial law is not introduced because someone wanted to. You say, this is not Pinochet, this is not Pinochet, and so on. This is a completely different story. An aggressor came and according to your legislation, if the border is violated, if there is armed aggression, you have all this written down long ago, written out in legislation, you introduce martial law and the introduction of martial law everywhere at all times means, in any case, a restriction of opportunities. If opportunities are limited, rights and freedoms are restricted. Therefore, the war itself restricts rights and freedoms. Yes, and you can’t do anything about it. We try, honestly, to balance as much as possible. I believe that the business sector works despite the difficulties of the war, and we do everything somewhere, there somewhere to reduce some load. Unfortunately, we cannot reduce taxes.

(02:09:12)
On the contrary, military tax is used for war. You need to take money somewhere. This, by the way, is about the fact that the US gave us a lot and Europe too, but compared to how much we needed for the war, this is not all. As for military salaries, you know that we could not pay the salaries of a million strong army. We could not pay it using the money from our partners. These are all expenses. This is all the money that the country and people have accumulated. You can’t do anything. I really want to reduce taxes. I will tell you frankly, I really want to.

(02:09:55)
Well, I think that the whole new tax system, new deregulation, new steps, new reforms, all this will be after the war. Although there is something to brag about, this is proof. And this is a document because if you want to get a candidacy for European Union, you must implement the appropriate number of reforms. We do everything. During the war, we voted for many reforms, including anti-corruption, banking reforms, land reforms, major reforms. We started a large privatization and the war did not stop us. Yes, it slowed down, but we went through a lot.
Lex Fridman
(02:10:42)
When do you think you’ll hold elections? Because for people who don’t know, part of the martial law elections were suspended and they were delayed and delayed and delayed and I think the next sort of plan is in February of 2025, but when do you think there will be presidential elections in Ukraine?
Volodymyr Zelenskyy
(02:11:02)
Elections were postponed once. They were not delayed, to be clear. Elections did not take place in 2024 that year. First of all, we need to understand the constitution. They were scheduled to be held in the spring of 2024. Due to martial law under the constitution, you cannot do this. These are the presidential elections. The parliamentary elections did not take place in the fall of 2024 according to the constitution. Yes, there are security things, there is the constitution, but there are security things. That is, everyone in Ukraine understands that this cannot be done until the war is over or legislation needs to be changed.

(02:11:52)
I believe that elections will take place immediately after the end of martial law. This is according to the law or members of the parliament need to get together and change legislation, which will be very difficult to do because society is against it. Why society against it? It is understandable why. Because we want elections that we want to trust. 8.5 million people went abroad. The infrastructure needs to be created for these millions of people to vote. Millions of people in the occupied territories. I’m not even talking about the occupation of 2014. I’m talking about the occupation right now. What to do with these people? This is a difficult question. And one of the most unfair ones is how to vote without having a million soldiers. It is impossible.

(02:12:57)
We need to think about how to change the system if the elections are held in times of war, change the legislation, which should include changes to the voting system, to think about online voting. Everyone is afraid because of certain attacks like cyber attacks and so on, but we need to think about it. I really think that it’s possible that we can end the war in 2025.
Lex Fridman
(02:13:27)
In January?
Volodymyr Zelenskyy
(02:13:29)
We’ve already agreed on it. I would very much like to. I would very much like to-
Lex Fridman
(02:13:34)
After the war?
Volodymyr Zelenskyy
(02:13:35)
And immediately. Yes, immediately. In the year of the end of the war. It’s a fact. Why? Because when martial law ends, you can immediately vote in parliament to hold elections and then everyone will vote because there are no restrictive measures. And after they vote, I think elections can be held in 90 days, something like that. Yes. And this means that immediately after the end of the war, elections may take place in 90 days.
Lex Fridman
(02:14:11)
Are you running for reelection?
Volodymyr Zelenskyy
(02:14:14)
Even I don’t know, really. I don’t know. I don’t know. It is a very difficult question. It depends on how this war will finish. It depends on what people will want. Mostly, it depends on people, first of all, and of course my family. We had no time to speak about it with my family and of course didn’t have a chance because we don’t think about it now. I mean, it’s something… There are a lot of, some, not a lot of, but enough voices in Ukraine from politicians, opposition and et cetera, about this, I guess. But we don’t think really seriously, didn’t think seriously with my family about it. So this is war. I mean, how to think about what will be after. It’s very difficult, really very difficult.
Lex Fridman
(02:15:18)
If we look at the field of candidates, maybe you can give your opinion about the set of ideas you see out there, including your own about the future of Ukraine. As I understand the candidates include Poroshenko, Zaluzhnyi, Arestovych, Budanov, Klitschko, there are many others. This is the internet speaking to me. What do you think are the space of ideas that these candidates represent?
Volodymyr Zelenskyy
(02:15:44)
I think there can be even a bigger number of candidates. Yeah, I don’t really know what will be. They have rights to participate if they want to. Yes, if they really want to, and can, they can go and do what they want, honestly. Most important is what are they doing now? I think that all these people are famous Ukrainian people and it’s important for them to do everything they can today, not begin any election campaign. I think this what can divide our people to have the elections during the war. I mean this make steps, speak about elections a lot, make a big mess about it. I think this is not right. That’s why I’m not agreeing with some of these people. But they can and I think that they can and maybe some of them will. And it’s okay. It’s normal. It’s very normal. Our system differs from the system in the United States. You have two parties and the parties decide who will be the leader. And in Ukraine, everybody can participate. Let them.
Lex Fridman
(02:16:57)
You think you’re going to win the debate? You versus Zaluzhnyi, Poroshenko or Arestovych and you decide to run, do you think you’re going to win the debate or you’re again focused on the war and everybody should be focused-
Volodymyr Zelenskyy
(02:17:11)
Oh, I’m really focusing on the war and-
Lex Fridman
(02:17:13)
I understand.
Volodymyr Zelenskyy
(02:17:14)
… I think the most difficult debate is what will be brought to the table and we spoke about it. It’ll be during the war, how to finish the war. I think that is my goal because it will be one of my most complicated debates and for any president who is in a war, of course, but I think this is my goal to win those debates and the other things are not for today.

Corruption

Lex Fridman
(02:17:44)
As I said, the dream I have is it’s a historic opportunity to make peace, to make lasting peace soon. So I’m glad you’re focused on that. Let me ask a question that a lot of people in the United States think about, and I care a lot about the future of Ukraine is corruption. This is something you have cared a lot about for a long time. You won the presidency in 2019, in big part your message of fighting corruption. But there’s a lot of accusations that during war, I mentioned $9 trillion in the United States, war breeds corruption. So can you speak to that, how you have been fighting corruption and can you respond to the accusations there has been corruption in Ukraine?
Volodymyr Zelenskyy
(02:18:42)
It’s very simple. First of all, we really have a very sophisticated anti-corruption system. Sophisticated not in the sense that it’s difficult to understand, but in that it really consists of many elements. It’s the most sophisticated in all of Europe. This is another requirement of the European Union. It was a requirement for Ukraine and for many years, Ukraine was not trusted. I want to tell you that under me, we all voted for bills, all the anti-corruption reforms, well, almost all reforms and all anti-corruption bodies today are independent. They work as requested. I still believe that they are not perfect yet. There are many issues. There is a judicial system, but also a judicial reform that our partners, the United States plus the EU, demanded from us. This is all written out. This is written out in specific laws, in specific decrees, in specific decisions. We did this, we’ve done 99% of this.

(02:19:50)
If something has not been done, it means that it is on the way. But in principle, all this exists and there is no such system as we have in Europe. To say that we do not have corruption would be lying. We just talk about it openly. We are genuinely fighting against it. Look, we have sitting in our prison, Ihor Kolomoyskyi, who is the most influential Ukrainian oligarch since independence and no one could do anything about him. The United States of America wanted to have Kolomoyskyi and they went to great lengths because of money laundering, etc. There are criminal cases in the United States, I think in Delaware, something like that. Neither Europe could do anything about it. That is, we did a lot with oligarchs. Russian oligarchs, sanctions were imposed, they were thrown out. Some of them fled the state, but they are all under sanctions. We exchanged some of them for our soldiers such as Medvedchuk to whose daughter Putin is godfather.

(02:20:57)
That is, we fought against the strongest influential oligarchs, which are, and were in Ukraine and we eliminated a lot of corruption. Of course corruption exists in everyday life. It exists. But institutionally, I am sure that Ukraine will overcome all this. This takes a little time. I would say honestly, that listen, what we call corruption and in some state of the world it’s called lobbyism, but this does not mean that there is no corruption there.

(02:21:36)
Let’s take the aid you mentioned during the war. First of all, we have no money. We have no money except for the war. We received weapons from the United States of America, from Europe. If we take for example money from the United States of America during all this time of the war, around 177 billion have been voted for or decided upon, 177 billion. Let’s be honest, we have not received half of this money.

(02:22:20)
The second point, which is very important, just as an example, is it corruption? The first question, whose corruption? This is the second. Here is just one small example for you. When the United States began to transfer US weapons, it was American money, but American weapons, money for these weapons. As a president, I had cargo jets, not in Ukraine. Because of the war, we moved them very quickly to Europe. We had cargo. We have good cargo fleet, very good because of Antonov. So I asked American side to grant me the opportunity because our jets are at another airfield and I asked America to give me the opportunity to use our jets for transfer, not to pay a lot. To whom? To your companies, to American companies. No, I didn’t get this opportunity. My jets stayed put and the United States jets, cargo jets moved these weapons. But everywhere you have to spend money so we could get more weapons, but we have to pay for this very expensive fleet. My question, is this corruption or not? Or lobbyism? What is it?
Lex Fridman
(02:24:05)
You mean corruption on the part of the US companies?
Volodymyr Zelenskyy
(02:24:08)
Yes. Making such decisions.
Lex Fridman
(02:24:10)
Yes, I got it.
Volodymyr Zelenskyy
(02:24:11)
The lobbying for such decisions involves some companies that make these decisions, but I can’t be open about it and I couldn’t speak loudly about it. I didn’t want, nor did I intend to cause any scandals to arise because otherwise you can freeze the support and that’s it. And that’s why when we talk about corruption, we must ask who is involved? If we had 177, and if we get the half, where’s the half? If you will find the second half, you will find corruption.
Lex Fridman
(02:24:45)
There is a perception of corruption. People like Donald Trump and Elon Musk really care about fighting corruption. What can you say to them to gain their trust that the money is going towards this fight for freedom, towards the war effort?
Volodymyr Zelenskyy
(02:25:03)
In most of cases, we did not receive money, we received weapons. And where we saw risks that something could be happening with weapons, we cracked down hard on everyone. And believe me, this is not only about Ukraine. Everywhere along the supply chain, there are some or other people and companies who want to make money, they try to make money on the war. We did not profit from the war. If we caught someone, believe me, we cracked down hard on them, and we did that, and we will continue to do so because to this day when someone says that, “Ukraine was selling weapons,” and by the way, Russia was the one pushing this narrative, we always responded, “Our soldiers would kill such people with their own hands without any trial.”

(02:25:56)
Do you honestly think anyone could steal weapons by the truckload when we ourselves don’t have enough on the front lines? And yet we have to provide proof to defend ourselves because when there’s an abundance of such misinformation, distrust starts to grow. And you’re right, people listen to various media outlets, see this and lose faith in you. In the end, you lose trust and with it you lose support. Therefore, believe me, we are fighting more against disinformation than against particular cases. Although I still emphasize once again at the everyday level, such things are still important. We catch these people and we fight them.

Elon Musk

Lex Fridman
(02:26:45)
I mentioned Elon Musk. I would be interested to hear what you think of him, why you respect him as a person, as an engineer, as an innovator, as a businessman. I would just like to hear from you, what do you think about Elon Musk?
Volodymyr Zelenskyy
(02:27:00)
First of all, I had a conversation with him at the beginning of the war. I talked with him. I respect him, first and foremost. I respect the self-made man, right?
Lex Fridman
(02:27:14)
Yes.
Volodymyr Zelenskyy
(02:27:14)
In English, I love such people. No one and nothing fell into their lap. But the man did something, did it all himself. I worked myself, created a big production company and I know what it means to make money, to select talented people, to impart knowledge to them, to invest money and to create something, something important for certain people. And I’m not comparing myself to Musk, he just, well, the man is a great leader of innovations in the world. And I believe that such people move the world forward. Therefore, I respect the result of his work. And we see this result. And for me, it has always been important that your result can be used. That these are not words but facts.

(02:28:16)
Let’s take the war. We are very grateful for Starlink. It has helped. We used it after Russian missile attacks on the energy infrastructure. There were problems with the internet, et cetera, with connection. We used Starlink both at the front and in kindergartens. It was used in schools, it helped children. We used it in various infrastructure and it helped us very much. And I would very much like Elon to be on our side as much as possible to support us. And yes, I’m grateful to him for Starlink. Truly, I am. First of all, so that our guys have a connection, and children too. And I am really grateful to him for that. I think I would like him to come to Ukraine, to talk to people here and to look around and so on.
Lex Fridman
(02:29:21)
Has Elon visited Kyiv or Ukraine yet?
Volodymyr Zelenskyy
(02:29:23)
No.
Lex Fridman
(02:29:25)
I hope the Kyiv airport will open soon, then it will be easier to fly in.
Volodymyr Zelenskyy
(02:29:30)
Yes, I am looking forward to it. Maybe we will open it, but only, and you must understand if the war is over, there must be sustainable peace and air defense systems to be honest. And we must ensure that they are long-lasting and effective. Let’s take the airport for example, and let’s focus on the airport in Dresden, which very well as it is handling important cargo for Ukraine in Poland. And there are patriot systems there because everyone understands what the risk is. Well, Russia is a risk and therefore we need air defense systems. And today, today, take for example, the air defense system of one city or another that is being shelled and move it, move it to the airport. Well, that would be dishonest. People are more important than planes. But there will be a moment, and Trump, by the way, I think that the war will end and President Trump may be the first leader to travel here by airplane. I think it would be symbolic by airplane.
Lex Fridman
(02:30:36)
Again, January 25th around that date. Right? Flying in, meeting the Air Force One.
Volodymyr Zelenskyy
(02:30:41)
That would be cool.
Lex Fridman
(02:30:42)
Elon Musk. I will meet you there for the second time too on the plane.
Volodymyr Zelenskyy
(02:30:46)
With pleasure.

Trump Inauguration on Jan 20

Lex Fridman
(02:30:47)
And you, by the way, before I forget, let me ask, are you coming on January 20th for President Trump’s inauguration?
Volodymyr Zelenskyy
(02:30:58)
I would like to, of course. I will be considering what is happening then in the war because there are moments of difficulties, escalation, many missiles, etc. But honestly, well, I can’t. I can’t come especially during the war, unless President Trump invites me personally. I’m not sure it’s proper to come because I know that in general, leaders are for some reason not usually invited to the inauguration of presidents of the United States of America. Well, and I know that there are leaders who can simply come, want to come and will come. Yeah, I know. And I know the temperament of some of these people. They can come at their discretion. This is very, very difficult for me. I am the kind of person that cannot come without an invitation. This is Putin. We did not invite him. He came to us, so to say. And me, I can’t do that.
Lex Fridman
(02:32:09)
No, but he publicly say that it would be great if you came to the inauguration or you mean did he invite it officially?
Volodymyr Zelenskyy
(02:32:15)
No, wait. Look, look, look. Listen, I am against any bureaucracy. I get rid of it as much as I can. But well, there are some complexities involving security. I decide and I fly, and the United States of America officially provides security. Not that I need this, mind you. I do not ask for helicopters to fly around and protect me, but they will simply do it themselves. The security service itself. They had to do it. I don’t want it, and sometimes I don’t need it. And I’m asking them.

(02:32:51)
It was for example, before the war, I think, yes, it was before the war, I had a meeting, yes, with President Trump. It was in 2019. I just wanted to go for a run early in the morning because I really wanted to exercise. And they, those tall bodyguards, a lot of them, they decided to join me, but I couldn’t really do it because they were in suits and I was in sportswear. I said, no, I can’t. It’s always funny. I don’t want to disturb anybody and cause anyone problems with me. And that’s why if he will invite me, I will come.
Lex Fridman
(02:33:34)
I thought he invited you.
Volodymyr Zelenskyy
(02:33:36)
Yeah?
Lex Fridman
(02:33:37)
Yeah. I thought he publicly invited you. But okay, I hope to see you there.
Volodymyr Zelenskyy
(02:33:40)
I think they had to to do some of their steps. I don’t know, but…
Lex Fridman
(02:33:46)
Step, yeah. The stamp was missing.
Volodymyr Zelenskyy
(02:33:49)
But with pleasure with my wife of course. And I think it’s important. It’s important.

Power dynamics in Ukraine

Lex Fridman
(02:33:55)
All right, let’s get back to a serious question. Sometimes they say it in America, this question of who is really in power? So let me ask, is someone controlling you? For example, oligarchs, American politicians, Yermak? I wanted to bring this up because I have been here in Ukraine twice since the invasion of 2022. And one of the things I’ve learned well is that actually nobody controls you. And this is one of your strengths as a president, as a person that oligarchs and other rich and powerful people like that cannot control you. Can you explain why that is? How you see it?
Volodymyr Zelenskyy
(02:34:44)
I think, and it is indeed true that I’m generally difficult to deal with. I am an ambitious person. I can’t submit to anyone. I can live by rules, by laws. I believe that this is the only thing that can control any person today. These are the rules and laws of the society or state where you live. And I believe that this is the most important thing. There’s no person who could control me as I once told President Trump when we had a meeting. By the way, journalists asked if Trump influenced me during the phone call. I told the journalist the truth then, who can influence me? Only my boy, my son. This is a fact. When he calls asking for something, well, then I lift up my arms, yes, and I cannot do anything about it because children are children. I have so little time with them. And therefore when there are these moments, they are precious and important to me. I am ready to do anything.

(02:35:58)
Also, probably my parents, they are an authority for me. Beyond that, I view it more as a system. No one can control the president. Therefore, we have oligarchs who either fled or are in prison because oligarchs usually control cash flows and people and influence politics. And we have concrete examples with sentences. They are not just under house arrest. Not just that there are some judgments under which their assets were frozen or sanctions were imposed. There are specific people who are behind bars. I think this is the answer regarding the influence. Would they like to influence me in the same way as any president of Ukraine because finance and cash flows always influence politics? Well, at least they want to do this. This is regarding the influence and other people on the vertical, they perform tasks as my managers. Andrii, you mentioned is one of those managers. Well, I am glad that I have such people. Well, probably there is nothing else to add here.
Lex Fridman
(02:37:19)
I will just say that your team that I spoke with is an excellent team. Excellent people.
Volodymyr Zelenskyy
(02:37:25)
Thank you.

Future of Ukraine

Lex Fridman
(02:37:26)
Okay, one last question. The future of Ukraine. If you look 5, 10, 20 years into the future, what can help Ukraine flourish economically, culturally, politically in the future?
Volodymyr Zelenskyy
(02:37:37)
Digital, it’s very important. Digitalization of all the process. We began this work. We have special ministry of digital transformation.
Lex Fridman
(02:37:38)
Yeah?
Volodymyr Zelenskyy
(02:37:47)
Yeah. So this is very good. And we also have our Diia. This is the name for all of these services. So I think that is the most important. This is, again, this is not only convenient, that will cancel any possibilities for future corruption because you don’t have any personal connections with people in the government or elsewhere. So you are just on your phone or any other device. That’s it. And I think we are doing very well. We are the best in Europe. All of Europe recognizes it. Some countries of the African Union asked us to provide this, the same service and we will do it after the war immediately. And I think that we can bring money to Ukraine from this. And I think what we also need, we need a tax reform. I think it will be very important for the businesses to return.

(02:38:43)
A lot of support will come, I think from USA business investment, not as direct aid to us, just to the private sector and resources. And I mentioned this to President Trump and to some European leaders who are our key strategic partners that will be happy, especially with the Americans, will be happy to sign these contracts and engage in joint investments in many areas. And I think we can develop oil, gas, green energy, including solar power. And we already have the resources. We can invest money into this. We have oil reserves in the Black Sea that we can exploit and we need your expertise and the investment of your companies. We have gold and uranium reserves, the largest in Europe by the way, which is also very important. For example, Russia has pushed France out of Africa. They urgently need uranium, which we have. So we are ready to open up for investments and this will give us of course, opportunities, jobs for people, revenue. I don’t want cheap labor, honestly. What I truly want, especially after the war, to open up for those people…
Volodymyr Zelenskyy
(02:40:00)
I truly want, especially after the war, to open up for those people who can really contribute and earn.
Lex Fridman
(02:40:08)
Yes, and give a reason to the eight million people to come back.
Volodymyr Zelenskyy
(02:40:11)
Yes, it’s so important. And they will come and we will recover and rebuild Ukraine. We will be very open to companies, and of course we will welcome our people back. It’s so important culturally.

(02:40:28)
I think the most important thing is to remain open and not change our direction because culturally aligning with Russia, it’s one idea, while aligning with Europe is another. Our people have chosen Europe. It’s their choice, it’s our choice, the choice of our nation, and I think it’s very important.
Lex Fridman
(02:40:45)
But first, you have to end the war.
Volodymyr Zelenskyy
(02:40:47)
Yes, you’re right. And we will. We want peace. Just to make it clear, we want peace. Just what I always say, you have to come to Ukraine and see for yourself. And people will tell you, “No, we can’t forgive those murderers who took our lives, but we still want to make peace.”

(02:41:12)
And honestly, I think that the highest approval rating of the president of the United States, of Trump now is in Ukraine. People really believe that he can truly help bring peace. Now they have faith, faith that he can make it happen, that he can support Ukraine and he can stop Putin and that he will make sure Putin doesn’t get everything he wants. This is very important, and it’s why we believe that we must not lose this opportunity.
Lex Fridman
(02:41:54)
I hope you find the path to peace. Thank you.
Volodymyr Zelenskyy
(02:41:57)
Thank you so much.
Lex Fridman
(02:41:58)
Thank you for talking today.
Volodymyr Zelenskyy
(02:41:59)
Thank you for coming.
Lex Fridman
(02:42:01)
[foreign language 02:42:01].
Volodymyr Zelenskyy
(02:42:02)
[foreign language 02:42:02] Yeah. You started. Thank you very much.

Choice of language

Lex Fridman
(02:42:09)
Thank you for listening to this conversation with the President of Ukraine, Volodymyr Zelenskyy. And now let me answer some questions and try to reflect on and articulate some things I’ve been thinking about. If you would like to submit questions, including in audio and video form, go to lexfridman.com/ama. Or to contact me for whatever other reason, go to lexfridman.com/contact.

(02:42:36)
First, I got a bunch of questions about this, so let me chat about the topic of language and let’s say the mechanics of multilingual conversation. Perhaps the details are interesting to some people. It also allows me to reflect back on the puzzle of it in this episode and what I can do better next time. I already explained in the intro the symbolic, historic, and geopolitical complexity of the choice of language in the conversation with President Zelenskyy. As I said, the Russian language is one that the president speaks fluently and was his primary language for most of his life. I speak Russian fluently as well. It’s the only common language we are both fluent in, so any other combination of languages required an interpreter, including when I spoke English. He did need an interpreter when I spoke English, and just like I was, was visibly encumbered and annoyed by the process of interpretation. This is why I tried to speak in Russian to the president instead of English, so that he can directly understand me without an interpreter.

(02:43:47)
I’m willing to take the hit for that, as I am for everything else. I’m not trying to protect myself, I’m trying to do whatever is best for the conversation for understanding, though it has been getting harder and harder to stay open, vulnerable and raw in public while the swarms of chanting internet mobs stop by with their torches and their color-coded hats, flags, frogs, pronouns, and hashtags.

(02:44:17)
Anyway, there is a lot of nuanced aspects of the conversational language that I would like to explain here. I’ll try to be brief. I can recommend a lot of books on this topic of language and communication that reveal just how amazing this technology of language is. For example, for a good overview, I recommend John McWhorter’s books and especially his lecture series for the Great Courses on language. There are several. In the Story of Human Language series, he gives a great discussion on spoken language versus written language, and that spoken language often relaxes the rules of communication. It uses shorter packets of words, loads in a bunch of subtle cues and meanings, all of which, like I’m trying to describe, are lost when there’s an interpreter in the loop.

(02:45:09)
Let me also describe some relevant characteristics of my peculiar language “abilities” in quotes. I was never good at speaking. I listen to think and understand better than I speak. For me, this is true for both English and Russian, but it is especially true for Russian. The Russian language allows for much more room for wit, nonstandard terms of phrase, metaphors, humor, rhyme, musicality, and let’s say deforming of words that create a lot of room for creativity and how meaning and emotion are conveyed. You could do the same in English, but it’s harder. I actually find that Brits are sometimes very good at this. One of my favorite humans to talk to is Douglas Murray. Setting the content of the conversation aside, the sheer linguistic brilliance and wit of dialogue with Douglas is a journey in itself. I think Christopher Hitchens had the same, and many others, like I said, especially Brits. Anyway, I’m able to detect and understand a lot of dynamism and humor in the Russian language, but I’m slow to generate it in part because I just don’t practice. I have very few Russian-speaking friends. Funny enough, most of them are Ukrainian, but they speak with me and each other in Russian. But of course, as I mentioned, this is slowly changing due to the war. But I tried to speak to the president in Russian so he would avoid needing an interpreter as much as possible.

(02:46:47)
One of the things I want to improve for next time is to make sure I get very good equipment for interpretation and arrange for an interpreter I trust to be exceptionally good for the dynamism and the endurance of a three-hour conversation in the style that I tried to do. Just to give you some behind-the-scenes details of the experience, equipment-wise, funny enough, it’s not actually so trivial to set up wireless connections from us, the two people talking, to the interpreter, and then back to us in a way that’s super robust and has clean audio. The audio I had in my ear from the interpreter had a loud background noise, so the whole time I’m hearing a shh sound with the voice of the interpreter coming in very quietly. What a wonderful experience this whole life is, frankly. Plus, his translation was often incomplete, at least for me, so I had to put together those puzzle pieces continuously. But again, it worked out. And hopefully our constant switching of languages and having a meta-discussion about language provided good insights as to the complexity of this fight for a nation’s identity and sovereignty that Ukraine has gone through.

(02:48:07)
Behind the scenes, off-mic on a personal level, President Zelenskyy was funny, thoughtful, and just a kind-hearted person. And really, the whole team were just great people. It was an experience I’ll never forget.

(02:48:24)
After the conversation was recorded, the next challenge was to translate all of this and overdub it and do it super quickly. These words I’m speaking now have to be translated and dubbed into Ukrainian and Russian. ElevenLabs were really helpful here, especially in bringing the president’s voice to life in different languages. But even more than that, they’re just an amazing team who inspired me and everyone involved. Please go support ElevenLabs. They are a great company and great people. The translation is separate from the text to speech and was done in part by AI and a lot by human. This is where the fact that we had constant switching between three languages was a real challenge. There are six transition mappings that have to be done: English to Ukrainian and Russian, Ukrainian to English and Russian, and then Russian to English and Ukrainian continuously, sentence by sentence, sometimes word by word. And each combination of language to language translation is best done by a person who specializes in that kind of mapping. It was all a beautiful mess, all of it.

(02:49:41)
And on top of all that, great translation is super hard. For example, I’ve read and listened to a lot of Dostoevsky both English and Russian and studied the process of how these books are translated by various translators. You can spend a week discussing how to translate a single important sentence well. Obviously, in this situation, we don’t have weeks, we have hours for the whole thing.

(02:50:06)
One of the things I regret is not putting enough time into the hiring and selecting great translators, from Russian and Ukrainian to English especially. I think translation is an art, so getting a good translator that works well with us is a process that needs more time and effort. I’ll be doing that more this month.

(02:50:27)
By the way, we have a small but amazing team. If you want to join us, go to lexfridman.com/hiring. If you’re passionate, work hard, and everyone on the team loves working with you, then we’ll do some epic stuff together. Would love to work with you. Like I said about ElevenLabs. There are a few things is awesome in life as being able to work hard with an amazing team towards a mission all of us are passionate about.

(02:50:54)
Anyway, I’ll probably be doing a few more interviews in the Russian language. I do have a lingering goal of interviewing the Mathematician Grigori Perelman, but there’s also others. I will also work on improving my whole pipeline both equipment-wise and interpreter-wise in doing these conversations in other languages because there are many that I would like to do in languages I don’t speak at all like Chinese, Mandarin or Spanish, Arabic, Hindi, Portuguese, French, German. I see language as both a barrier for communication and a portal into understanding the spirit of a people connected by that language. It’s all a weird and beautiful puzzle, and I’m just excited to get the chance to explore it.

Podcast prep and research process


(02:51:39)
All right, I got a question on how I prepare for podcasts. This has evolved and expanded more and more over time. There are some podcasts that I prepare hundreds of hours for. In AI terms, let’s say, first, I’m training a solid background model by consuming as much variety on the topic as possible. A lot of this comes down to picking high-signal sources, whether it’s blogs, books, podcasts, YouTube videos, X accounts, and so on.

(02:52:09)
For this conversation with President Zelenskyy, for example, since February 2022 I’ve spoken with hundreds of people on the ground, I’ve read Kindle or audiobook, about 10 books fully, and then I skimmed about 20 more. And I don’t mean books about Zelenskyy, although he does appear in some of them, I mean books where this conversation was fully in the back of my mind as I’m reading the book. For example, I read Red Famine by Anne Applebaum. It’s about Holodomor. Does it directly relate to Zelenskyy? Not on the surface, no. But it continues to weave the fabric of my understanding of a people, of a history of the region.

(02:52:56)
But it’s really important for me to read books from various perspectives. And I’m always trying to calculate the bias under which the author operates and adjusting for that in my brain as I integrate the information. For example, Anne Applebaum’s book, Gulag, is very different from Aleksandr Solzhenitsyn’s Gulag Archipelago. The former is a rigorous, comprehensive, historical account, the latter is a literary, psychological, and personal portrait of Soviet society. Both I think are extremely valuable. On the bias front, for example, The Rise and Fall of the Third Reich by William Shirer is a good example. It is full of bias, but he was there, and to me, he has written probably one of the greatest if not the greatest book on the Third Reich ever. But like I said, it has a lot of inaccuracies and biases. You can read about them online if you like. But my job in this case and in all cases is to adjust based on my understanding of the authors’ biases and take the wisdom from the text where it could be found and putting the inaccuracies aside into the proverbial dustbins of history.

(02:54:12)
As I’m reading, I’m writing down my thoughts as they come up, always digging for some deeper insight about human nature. If I’m at my computer, I’ll write it down in Google Doc, sometimes use Notion or Obsidian. If I’m not on my computer, I’ll use Google Keep. For example, if I’m listening to an audiobook and I’m running along the river, if a good idea comes to mind, I’ll stop, think for a few seconds, and then do a speech to text note in Google Keep. By the way, listening to audiobook at 1X speed. Old school. And eventually I get a gigantic pile of thoughts and notes that I look over to refresh my memory. But for the most part, I just throw them out. It’s a background model building process. By the way, LLMs are increasingly becoming useful here for organization purposes, but have not yet been useful, at least for me, and I do try a lot for insight extraction or insight generation purposes.

(02:55:14)
I should mention that my memory for specific facts, names, days, quotes is terrible. What I remember well is high-level ideas. That’s just how my brain works, for better or for worse. I realize that sometimes forgetting all of the details and the words needed to express them makes me sound simplistic and even unprepared. I’m not, but that’s life. We have to accept our flaws and roll with them.

(02:55:44)
Aside from books, I also listen to a lot of podcasts and YouTube videos where people are talking about the topic. For the President Zelenskyy episode, I listened probably to hundreds of hours of content from his supporters and from his critics; from all sides. Again, I choose who to listen to based not on their perspective, but based on SNR, signal-to-noise ratio. If I’m regularly getting insights from a person, I will continue listening to them whether I agree or disagree.

(02:56:14)
In the end, this turns out to be a lot of hours of prep, but to say that it’s X hours per episode is not accurate because a lot of this preparation transfers from one guest to another even when there’s an insane level of variety in the guests. We’re all humans, after all. There is a thread that connects all of it together somehow if you look closely enough.

(02:56:35)
For more technical guests in STEM fields, I’ll read papers, a lot of papers, and also technical blog posts and technical tweet threads. This is a very different process. For AI or CS-related topics, I will run other people’s code, I will write my own, implement stuff from scratch. If it’s a software company, I’ll use their tools and software relevant. But in the actual conversation, I constantly am searching for simple but profound insights at various levels of abstraction. Sometimes this means asking a trivial question in hopes of uncovering the non-trivial, counterintuitive but fundamental idea that opens the door to a whole new way of looking at the field.

(02:57:18)
And actually, every guest is their own puzzle. Like preparing for Rick Rubin was me listening to hundreds of songs he produced and even learning some on guitar like Hurt by Johnny Cash. Preparing for the Cursor Team episode meant obviously I had to use Cursor fully for several weeks; all of its features. I switched completely from VS Code to Cursor. For Paul Rosolie, round two especially, I literally went deep into the jungle with Paul and almost died fully taking the leap toward adventure with him.

(02:57:56)
When it gets close to the conversation, I’ll start working on the actual interview questions and notes. And there I’m asking myself what am I personally curious about? I love podcasts. I’m a big fan of many, many podcasts, and so I ask myself, “What would I want this person to explain on a podcast? And maybe what aspect of their thought process or their humanity would I want to be surfaced or have the chance to be surfaced?”

(02:58:26)
In the actual conversation, I always try to put my ego aside completely and do whatever it takes to have a good conversation and serve the listener. This means asking questions, simply trying to define terms and give context if needed, being open-minded, vulnerable, curious, and challenging the guests when needed. Despite the claims on the internet, I do ask a lot of challenging questions, including follow-ups, but always with empathy. I don’t need to be right. I don’t need to signal my moral or intellectual superiority to anyone. I try to do the opposite actually, because I want the guest to open up. And I trust the intelligence of the listener to see for themselves if the guest is full of shit or not, to detect the flaws and the strengths of how the guest thinks or who they are deep down. A lot of times when interviewers grill the guest, it doesn’t reveal much except give a dopamine hit to the echo chambers who hate the guest.

(02:59:29)
As I said in the intro, I believe the line between good and evil does run through the heart of every man. The resulting conversations are sometimes a failure, sometimes because they are too short, sometimes because the chemistry was just not working, sometimes because I fucked it up. I try to take risks, give it everything I got and enjoy the rollercoaster of it all no matter what. And as I said, I trust the listener to put it all together and I trust the critic to tear it apart. And I love you all for it.

Travel and setup


(03:00:04)
All right, I got a bit of a fun question. It’s a long one. Delian, cool name, wrote in saying he spotted me out in the wild and had a question about it. He wrote, “I saw Lex working at the Detroit Airport between flights. I hesitated and ultimately decided not to interrupt since he was in focus mode.” True. “Lex had his headphones, earbuds on,” listening to brown noise, “Microsoft’s surface propped up at eye level, Kinesis Advantage keyboard on the table. The use of Microsoft Windows is surprising, but it has been discussed in the past.” True. “The ergonomics of the setup surface at eye level means that Lex cares about his health. But the anomalously large Kinesis advantage keyboard seems like such a burden to lug around airports. I cannot help but ask why is it that Lex is going through the hassle to bring this absolutely large keyboard with him as carry-on? It barely fits in a backpack. Carrying it around must be necessary for Lex for some reason.” I love the puzzle of this, that you’re trying to think through this. “The pain of lugging this tool around must be much smaller than the problem it solves for it? What problem does this keyboard solve? What makes it necessary at the airport? Productivity? Health? RSI?”

(03:01:24)
Good questions. Thank you, Delia. Great question. It made me smile, so I thought I’d answer. I remember that day. There was something else about that day, aside from the keyboard that I miss, so I am filled with a melancholy feeling that is appropriate for the holiday season. Let me try to set the melancholy feeling aside, answer a question about my computer setup when I’m traveling. Whether I’m going to SF, Boston, Austin, London, or the front in Ukraine, I am always bringing the Kinesis keyboard. I don’t have RSI or any other health issues of that kind that I’m aware of, even though I’ve been programming, playing guitar, doing all kinds of combat sports my whole life, all of which put my hands and fingers in a lot of precarious positions and situations. For that reason, and in general, ergonomics have never been a big concern for me. I can work on a crappy chair and a table, sleep on the floor. It’s all great. I’m happy with all of it.

(03:02:36)
Why Kinesis? Which, by the way, is right here. I had to think about it. Your question actually made me reflect. And I was hoping as I’m answering it the truth will come off on many levels. It is true that I’m more productive with it. I can type and correct mistakes very fast compared to a regular keyboard, both in natural language typing and in programming. Fast enough, I think, where it feels like I can think freely without the physical bottlenecks and constraints of fingers moving. The bit rate in Neuralink parlance is high enough for me to not feel like there is cognitive friction of any kind.

(03:03:26)
But the real answer may be the deeper, more honest answer or something else. I’ve used the Kinesis keyboard for over 20 years, so maybe it’s like one of those love stories where a guy and a girl love each other and you try to quit because it doesn’t quite work, but every time you leave, you ask yourself, “Why?” And then you realize that when you’re together, your life is just full of simple joys, so what’s the point of leaving? What’s the point of life if not to keep close to you the things that bring you joy, Delia? Like this keyboard, it brings me joy. It’s a bad metaphor, over anthropomorphized perhaps, but I never promised a good one. I’m like a cheap motel on a road trip; low quality is part of the charm. I do have some good motel stories for another time. This does not feel like the appropriate time. All that said, to disagree with myself, I did use Emacs also for over 20 years, and in a single week recently switched to VS Code and then Cursor and never looked back. Take my romantic nature with a grain of salt.

(03:04:38)
Yes, eventually I’ll have to leave, but for now, you’ll keep finding me on occasion in a random airport somewhere listening to brown noise, writing away the hours on this Kinesis keyboard. Now, if you see me without it, maybe it’ll give you the same change of melancholy feeling I feel now in looking back to that airport in Detroit.

(03:05:03)
Anyway, more about my travel setup, if anyone’s curious. I usually do travel with a Windows laptop, but I am mostly using Linux on it through WSL, Windows Subsystem for Linux. And in some cases, I’m dual booting Linux and Windows. I also need to be able to video edit, so on a longer trips, I usually have a bigger laptop with a bigger screen, lots of memory, good CPU, good GPU. All of that helps with video editing on Adobe Premiere. In general, I’m extremely minimalist except for the few, let’s call them the sentimental things, like all my podcast recording equipment fits into a small suitcase. I try to keep it as simple as possible. Thank you for the question, and see you at the next airport.

Conclusion


(03:05:51)
All right, I think it’s time to bring things to close. I’d like to give a big thanks to you for giving me your time and your support over the years. It means the world. If you want to get in touch with me, go to lexfridman.com/contact. There you can give feedback, ask questions, request guests for the podcast, or submit the Coffee with Lex form if you just want to chat with me over a cup of coffee. I’ll be traveling across the world a bunch this year, from Europe to South America and more, so it would be cool to do some small meetups and meet some interesting people. This has been a journey of a lifetime. Thank you for everything. Onto the next adventure. I love you all.

Transcript for Adam Frank: Alien Civilizations and the Search for Extraterrestrial Life | Lex Fridman Podcast #455

This is a transcript of Lex Fridman Podcast #455 with Adam Frank.
The timestamps in the transcript are clickable links that take you directly to that point in
the main video. Please note that the transcript is human generated, and may have errors.
Here are some useful links:

Table of Contents

Here are the loose “chapters” in the conversation.
Click link to jump approximately to that part in the transcript:

Introduction

Adam Frank
(00:00:00)
If we don’t ask how long they last, but instead ask what’s the probability that there have been any civilizations at all, now matter how long they lasted. I’m not asking whether they exist now or not, I’m just asking in general about probabilities to make a technological civilization anywhere and at any time in the history of the university. That, we’re able to constrain. What we found was basically that there have been 10 billion trillion habitable zone planets in the universe. What that means is those are 10 billion trillion experiments that have been run. The only way that we’re the only time that this whole process from abiogenesis to a civilization has occurred is if everyone one of those experiments failed.

(00:00:51)
Therefore, you could put a probability, we called it the Pessimism Line. We don’t really know what nature sets for the probability of making intelligent civilizations, but we could set a limit using this. We could say, look, if the probability per habitable zone planet is less than 10 to the minus-22, one in 10 billion trillion, then yeah, we’re alone. If it’s anywhere larger than that, then we’re not the first. It’s happened somewhere else. To me, that was mind-blowing. It doesn’t tell me there’s anybody nearby, the galaxy could be sterile. It just told me that unless nature’s really has some bias against civilizations, we’re not the first time this has happened. This has happened elsewhere over the course of cosmic history.
Lex Fridman
(00:01:36)
The following is a conversation with Adam Frank, an astrophysicist interested in the evolution of star systems and the search for alien civilizations in our universe. This is The Lex Fridman Podcast. To support it, please check out our sponsors in the description. And now, dear friends, here’s Adam Frank.

Planet formation

Lex Fridman
(00:01:58)
You wrote a book about aliens. The big question, how many alien civilizations are out there?
Adam Frank
(00:02:04)
Yeah, that’s the question. The amazing thing is that, after two-and-a-half millennia of people yelling at each other, or setting each other on fire occasionally over the answer, we now actually have the capacity to answer that question. In the next 10, 20, 30 years, we’re going to have data relevant to the answer to that question. We’re going to have hard data finally that will, one way or the other … Even if we don’t find anything immediately, we will have gone through a number of planets. We’ll be able to start putting limits on how common life is.

(00:02:38)
The one answer I can tell you, which was an important part of the problem, is how many planets are there? Just like people have been arguing about the existence of life elsewhere for 2500 years, people have been arguing about planets for the exact same amount of time. You can see Aristotle yelling at Democritus about this. You can see they had very wildly different opinions about how common planets were going to be, and how unique Earth was. And that question got answered. Which is pretty remarkable, that in a lifetime, you can have a 2500-year-old question. The answer is they’re everywhere. There are planets everywhere.

(00:03:14)
It was possible that planets were really rare. We didn’t really understand how planets formed. If you go back to, say the turn of the 20th Century, there was a theory that said planets formed when two stars passed by each other closely, and then material was gravitationally squeezed out. In which case, those kinds of collisions are so rare that you would expect one in a trillion stars to have planets. Instead, every star in the night sky has planets.
Lex Fridman
(00:03:42)
One of the things you’ve done is simulated the formation of stars. How difficult do you think it is to simulate the formation of planet? Like simulate a solar system through the entire of the evolution of the solar system. This is a numerical simulation sneaking up to the question of how many planets are there.
Adam Frank
(00:04:01)
That, actually, we’re able to do now. You can run simulations of the formation of planetary system. If you run the simulation, really where you want to start is a cloud of gas, these giant interstellar clouds of gas that may have a million times the mass of the Sun in them. You run a simulation of that, it’s turbulent. Gas is roiling and tumbling. Every now and then, you get a place where the gas is dense enough that gravity gets hold of it and it can pull it downward, so you’ll start to form a proto-star.

(00:04:32)
A proto-star is basically the young star, this ball of gas where nuclear reactions are getting started. But it’s also a disc. As material falls inward because everything’s rotating, as it falls inward, it’ll spin up and then it’ll form a disc. The material will collect in what’s called an accretion disc or a proto-planetary disc. You can simulate all of that.

(00:04:56)
Once you get into the disc itself and you want to do planets, things get a little bit more complicated because the physics gets more complicated. Now you got to start worrying about dust, because actually dust … Dust is the wrong word. It’s smoke, really. These are the tiniest bits of solids. They will coagulate in the disc to form pebbles, and then the pebbles will collide to form rocks. And then the rocks will form boulders, et cetera, et cetera. That process is super complicated. But we’ve been able to simulate enough of it to begin to get a handle on how planets form. How you accrete enough material to get the first proto-planets, or planetary embryos as we call them.

(00:05:37)
The next step is those things start slamming into each other to form planetary-sized bodies. Then the planetary bodies slam into each other. Earth, the Moon came about because there was a Mars-sized body that slammed into the Earth and basically blew off all the material. Then eventually formed the Moon.
Lex Fridman
(00:05:54)
And all of them have different chemical compositions, different temperatures?
Adam Frank
(00:06:00)
Yeah. The temperature of the material in the disc depends on how far away you are from the star.
Lex Fridman
(00:06:07)
Got it.
Adam Frank
(00:06:07)
It decreases.

(00:06:08)
There’s a really interesting point. Close to the star, temperatures are really high. The only thing that can condense, that can freeze out, is going to be stuff like metals. That’s why you find Mercury is this giant ball of iron, basically. Then as you go further out, stuff, the gas gets cooler. And now you can start getting things like water to freeze. There’s something we call the Snow Line, which is somewhere in our solar system, out around between Mars and Jupiter. That’s the reason why the giant planets in our solar system, Jupiter, Saturn, Uranus, and Neptune, all have huge amounts of ice in them, or water and ice.

(00:06:47)
Actually, Jupiter and Saturn don’t have so much, but the moons do. The moons have so much water in them that there’s oceans. We’ve got a number of those moons have got more water on them than there’s water on Earth.
Lex Fridman
(00:06:58)
Do you think it’s possible to do that kind of simulation to have a stronger and stronger estimate of how likely an Earth-like planet is? Can we get the physics simulation done well enough to where we can start estimating what are the possible Earth-like things that could be generated?

Plate tectonics

Adam Frank
(00:07:17)
Yeah, I think we can. I think we’re learning how to do that now. One part is trying to just figure out how planets form themselves in doing the simulations. That cascade from dust grains up to planetary embryos, that’s hard to simulate because you got to do both the gas, and you got to do the dust and the dust colliding, and all that physics.

(00:07:40)
Once you get up to a planet-sized body, then you have to switch over to almost a different kind of simulation. Often what you’re doing is you’re assuming the planet this this spherical ball, and then you’re doing a 1D, a radial calculation. You’re just asking, “All right, what is the structure of it going to be? Am I going to have a solid iron core, or am I going to get a solid iron core with a liquid iron core out around it?” Like we have on Earth. Then you get a silicate, rocky mantle, and then a crust. All those details, those are beyond being able to do full 3D simulations from Ab Initio, from scratch. We’re not there yet.
Lex Fridman
(00:08:20)
How important are those details, like the crust and the atmosphere, do you think?
Adam Frank
(00:08:24)
Hugely important. I’m part of a collaboration at the University of Rochester, where we’re using the giant laser. Literally, this is called the Laboratory for Laser Energetics. We got a huge grant from the NSF to use that laser to slam tiny pieces of silica to understand what conditions are like at the center of the Earth. Or even more importantly, the center of Super-Earths.

(00:08:47)
This is what’s wild. The most common kind of planet in the universe, we don’t have in our solar system. Which is amazing, right? We’ve been able to study or observe enough planets now to get a census. We have an idea of whose average, whose weird. Our solar system’s weird, because the average planet has a mass somewhere between a few times the mass of the Earth, to maybe 10 times the mass of the Earth. That’s exactly where there are no planets in our solar system.

(00:09:20)
The smaller ones of those we call Super-Earths, the larger ones we call Sub-Neptunes. They’re anybody’s guess. We don’t really know what happens to material when you’re squeezed to those pressures, which is millions, tens of millions of times the pressure on the surface of the Earth. Those details really will matter of what’s on in there, because that will determine whether or not you have, say for example, plate tectonics.

(00:09:44)
We think plate tectonics may have been really important for life on Earth, for the evolution of complex life on Earth. It turns out, and this is the next generation where we’re going with the understanding the evolution of planets and life. It turns out that you actually have to think hard about the planetary context for life. You can just be like, “Oh, there’s a warm pond,” and then some interesting chemistry happens in the warm pond. You actually have to think about the planet as a whole and what it’s gone through in order to really understand whether a planet is a good place for life or not.
Lex Fridman
(00:10:16)
Why do you think plate tectonics might be useful for the formation of complex life?
Adam Frank
(00:10:21)
There’s a bunch of different things. One is that the Earth went through a couple of phases of being a snowball planet. We went into a period of glaciation where pretty much the entire planet was under ice. The oceans were frozen.

(00:10:36)
Early on in Earth’s history, there was barely any land. We were actually a water world, with just a couple of Australia-sized cratons they called them, proto-continents.

(00:10:48)
We went through these snowball Earth phases. If it wasn’t for the fact that we had an active plate tectonics, which had a lot of vulcanism on it, we could have been locked in that forever. Once you get into a snowball state, a planet can be trapped there forever. Which is maybe you already had life formed, but then because it’s so cold, you may never get anything more than just microbes.

(00:11:10)
What plate tectonics does, because it fosters more vulcanism, is that you’re going to get carbon dioxide pumped into the atmosphere, which warms the planet up and gets you out of the snowball Earth phase. But even more, there’s even more really important things.

(00:11:26)
I just finished a paper where we were looking at something called the Hard Steps Model, which is this model that’s been out there for a long time that purports to say intelligent life in the universe will be really rare. It made all these assumptions about the Earth’s history, particularly about the history of life and the history of the planet have nothing to do with each other. It turns out, and as I was doing the reading for this, that Earth probably, early on, had a more mild form of plate tectonics, and then somewhere about a billion years ago, it ramped up.

(00:11:54)
That ramping up changed everything on the planet, because here’s a funny thing. The Earth used to be flat. All the Flat Earthers out there can get excited for one second.
Lex Fridman
(00:12:04)
Clip it. It still is.
Adam Frank
(00:12:08)
What I mean by that is that there really weren’t many mountain ranges. The beginning of, I think the term is orogenesis, mountain building, the true Himalayan-style giant mountains, didn’t happen until this more robust form of plate tectonics, where the plates are really being driven around the planet. That is when you get the crusts hitting each other, and they start pushing into these Himalayan- style mountains.

(00:12:30)
The weathering of that, the erosion of that puts huge amounts of nutrients, things that microbes want to use, into the oceans. And then what we call the net primary productivity, the bottom of the food chain, how much sugars they are producing, how much photosynthesis they are doing shot up by a factor of almost 1000. The fact that you had plate tectonics supercharged evolution in some sense. We’re not exactly sure how it happened, but it’s clear that the amount of life, the amount of living activity that was happening really got a boost from the fact that something there was this new vigorous form of plate tectonics.
Lex Fridman
(00:13:10)
It’s nice to have turmoil. In terms of temperature, in terms of surface geometries, in terms of the chemistry of the planet, turmoil.
Adam Frank
(00:13:20)
Yeah, that’s actually really true. Because what happens is, if you look at the history of life … That’s an excellent point that you’re bringing up. If you look at the history of life on Earth, we get abiogenesis somewhere around at least 3.8 billion years ago. That’s the first microbes. They take over enough that they really do, you get a biosphere. You get a biosphere that is actively changing the planet.

(00:13:40)
But then you go through this period they called the Boring Billion, where it’s a billion years and it’s just microbes. Nothing’s happening, it’s just microbes. The microbes are doing amazing things. They’re inventing fermentation. Thank you very much, we appreciate that. But it’s not until you get probably these continents slamming into each other, you really get the beginning of continents forming and driving changes that evolution has to respond to. That on a planetary scale, this turmoil, this chaos is creating new niches, as well as closing other ones. Biology, evolution has to respond to that.

(00:14:15)
Somewhere around there is when you get the Cambrian Explosion. It’s when suddenly every body plan … Evolution goes on an orgy, essentially. Yeah. It does look like that chaos or that turmoil was actually very helpful to evolution.

Extinction events

Lex Fridman
(00:14:31)
I wonder if there is some extremely elevated levels of chaos, almost like catastrophes behind every leap of evolution. You’re not going to have leaps. In human societies, we have an Einstein that comes up with a good idea. But it feels like on an evolutionary timescale, you need some real big drama going on for the evolutionary system to have to come up with a solution to that drama. An extra complex solution to that drama.
Adam Frank
(00:15:01)
Well, I’m not sure if that’s true. I don’t know if it needs to be an almost extinction event.
Lex Fridman
(00:15:05)
Right.
Adam Frank
(00:15:05)
Because it’s certainly true that we have gone through almost extinction events. We’ve had five mass extinctions. But you don’t necessarily see that there was this giant evolutionary leap happening after those.

(00:15:18)
With the comet impact, the K-T Boundary, certainly lots of niches opened up. That’s why we’re here, because our ancestors were little basically rodents, rats living under the footsteps of the dinosaurs. It was that comet impact that opened the route for us. That still took another 65 million years. It was like this thing immediately happened.

(00:15:42)
But what we found with this Hard Steps Paper, because the whole idea of the Hard Steps Paper was it was one of these anthropic reasoning kinds of things. Where Brandon Carter said, “Oh, look. The intelligence doesn’t show up on Earth until about almost close to when the end of the Sun’s lifetime.” He’s like, “Well, there should be no reason why the Sun’s lifetime and the time for evolution to produce intelligence should be the same.” He goes through all this reasoning, anthropic reasoning. He ends up with the idea that, “Oh, it must be that the odds of getting intelligence are super-low, and so that’s the hard step.”

(00:16:21)
There was a series of steps in evolution that were very, very hard. Because of that, you can calculate some probability distributions. Everybody loves a good probability distribution, and they went a long way with this. But it turns out that the whole thing is flawed because, when you look at it, of course the timescale for the Sun’s evolution and the timescale for the evolution on life are coupled, because the timescale for evolution of the Earth is coupled, is about the same timescale as the evolution of the Sun. It’s billions of years. The Earth evolves over billions of years.

(00:16:53)
Life and the Earth co-evolve. That’s what Brandon Carter didn’t see is that actually, the fate of the Earth the fate of life are inextricably combined. This is really important for astrobiology, too. Life doesn’t happen on a planet, it happens to a planet. This is something that David Grinspoon and Sara Walker both say, and I agree with this. It’s a really nice way of putting it.

(00:17:19)
Plate tectonics, the evolution of oxygen, of an oxygen atmosphere, which only happened because of life. These things, these are things that are happening where life and the planet are sloshing back-and-forth. Rather than, to your point about do you need giant catastrophes, maybe not giant catastrophes. But what happens is, as the Earth and life are evolving together, windows are opening up, evolutionary windows.

(00:17:46)
For example, life put oxygen into the atmosphere. When life invented this new form of photosynthesis about 2.5 billion years ago, that broke water apart to work to do its chemical shenanigans. It broke water apart and pushed oxygen into the atmosphere. That’s why there’s oxygen in the atmosphere. It’s only because of life.

(00:18:07)
That opened up huge possibilities, new spaces for evolution to happen. But it also changed the chemistry of the planet forever. The introduction of oxygen photosynthesis changed the planet forever, and it opened up a bunch of windows for evolution that wouldn’t have happened otherwise. Like for example, you and I, we need that amount of oxygen. Big-brained creatures need an oxygen-rich atmosphere because oxygen is so potent for metabolism. You couldn’t get intelligent creatures 100 million years after the planet formed.

Biosphere

Lex Fridman
(00:18:41)
So really, on a scale of a planet when there’s billions and trillions of organisms on a planet, they can actually have planetary scale impact.
Adam Frank
(00:18:53)
Yeah.
Lex Fridman
(00:18:53)
The chemical shenanigans of an individual organism when scaled out to trillions can actually change a planet.
Adam Frank
(00:18:59)
Yeah. We know this for a fact now.

(00:19:00)
There was this thing, Gaia Theory, which James Lovelock introduced in the ’70s. And then, Lynn Margulis, the Biologist Lynn Margulis together. This Gaia Theory was the idea that life takes over a planet, life hijacks a planet in a way that the sum total of life creates these feedbacks between the planet and the life, such that it keeps the planet habitable. It’s kind of a homeostasis.

(00:19:29)
I can go out … Right now outside, it’s 100-degrees. And I go outside, but my internal temperature is going to be the same. I can go back to Rochester, New York in the winter, and it’s going to be zero-degrees, but my internal temperature is going to be the same. That’s homeostasis.

(00:19:42)
The idea of Gaia Theory was that life, the biosphere exerts this pressure on the planet or these feedbacks on the planet, that even as other things are changing, the planet will always stay in the right kinds of conditions for life. Now when this theory came out, it was very controversial. People were like, “Oh my God, what are you, smoking weed?” There were all these Gaian Festivals with Gaian dances. It became very popular in the New Age community.

(00:20:09)
But Lovelock actually, they were able to show that no, this has nothing to do with the planet being conscious or anything. It was about these feedbacks, that the biology, the biosphere can exert these feedbacks. We’re still unclear whether there are true Gaian feedbacks, in the sense that the planet can really exert complete control. But it is absolutely true that the biosphere is a major player in Earth’s history.
Lex Fridman
(00:20:35)
The biosphere fights for homeostasis on Earth.
Adam Frank
(00:20:39)
Okay. What I would say right now is I don’t know if I can say that scientifically. I can certainly say that the biosphere does a huge amount of the regulation of the planetary state. And over billions of years, has strongly modified the evolution of the planet. A true Gaian feedback would be exactly what you said.

(00:20:57)
The biosphere is somewhere … Sara Walker, and David Grinspoon, and I actually did a paper on this about the idea of planetary intelligence, or cognition across a planetary scale. I think that actually is possible. It’s not conscious, but there is a cognitive activity going on. The biosphere, in some sense, knows what is happening because of these feedbacks. It’s still unclear whether we have these full Gaian feedbacks, but we certainly have semi-Gaian feedbacks.

(00:21:24)
If there’s a perturbation on the planetary scale, temperature, insulation, how much sunlight’s coming in, the biosphere will start to have feedbacks that will damp that perturbation. Temperature goes up, the biosphere starts doing something, temperature comes down.

Technosphere

Lex Fridman
(00:21:39)
Now I wonder if the technosphere also has a Gaian feedback or elements of a Gaian feedback? Such that the technosphere will also fight to some degree for homeostasis. Open question, I guess.
Adam Frank
(00:21:51)
Well, I’m glad you asked that question. Because that paper that David, and Sara, and I wrote, what we were arguing was is that over the history of a planet … When life first forms, 3.8 billion years ago, it’s thin on the ground. You’ve got the first species, these are all microbes. There are not enough of them to exert any kind of these Gaian feedbacks. We call that an immature biosphere. But then as time goes on, as life becomes more robust and it begins to exert these feedbacks keeping the planet in the place where it needs to be for life, we call that a mature biosphere. I’m sure later on, we’re going to talk about definitions of life and such. There’s this great term called autopoiesis that Francisco Varela, the Neurobiologist Francisco Varela came up with. He said, “One of the defining things about life is this property of autopoiesis,” which means self-creating and self-maintaining. Life does not create the conditions which will destroy itself. It’s always trying to keep itself in a place where it can stay alive. The biosphere, from this Gaian perspective, has been autopoietic for billions of years.

(00:23:02)
Now we just invented this technosphere in the last couple of hundred years. What we were arguing in that paper is that it’s an immature technosphere. Because right now, with climate change and all the other things we’re doing, the technosphere right now is destroying the conditions under which it needs to maintain itself. The real job for us if we’re going to last over geological timescales, if we want a technosphere that’s going to last tens of thousands, hundreds of thousands, millions of years, then we’ve got to become mature. Which means to not undermine the conditions, to not subvert the conditions that you need to stay alive. As of right now, I’d say we’re not autopoietic.
Lex Fridman
(00:23:44)
Wow. I wonder if we look across thousands, tens of thousands, hundreds of thousands of years, that the technosphere should create perturbations as a way for developing greater and greater defenses against perturbations. Which sounds like a ridiculous statement. But basically, go out and play in the yard and hurt yourself, to strengthen. Or drink water from the pond.
Adam Frank
(00:24:13)
From the pond. Yeah, right. Get sick a few times.
Lex Fridman
(00:24:16)
To strengthen the immune system.
Adam Frank
(00:24:18)
Yeah. Well, you know it’s interesting with the technosphere, we can talk about this more. We’re just emerging as a technosphere, in terms of as an interplanetary technosphere. That’s really the next step for us. David Grinspoon talks about it. I love this idea of anti-accretion. This amazing thing that, for the first time over the entire history of the planet, stuff is coming off the planet. It used to be everything just fell down, all the meteorites fell down. But now we’re starting to push stuff out. The idea of planetary defense or such, we are actually going to start exerting perturbations on the solar system as a whole. We’re going to start engineering, if we make it. I always like to say that if we can get through climate change, the prize at the end is the solar system. We’ll be literally engineering the solar system.

(00:25:06)
But what you can think of right now with what’s happening with the Anthropocene, the great acceleration that is the technosphere, is the creation of it, that is a giant perturbation on the biosphere. The technosphere sits on top of the biosphere, and if the technosphere undermines the biosphere for its own conditions of habitability, then you’re in trouble. The biosphere is not going away. There’s nothing we could do. The idea that we have to save the Earth is a little ridiculous. The Earth is not a furry little bunny that we need to protect. But it’s the conditions for us. Humanity emerged out of the Holocene, the last 10,000 years interglacial period. We can’t tolerate very different kinds of Earths. That’s what I mean about a perturbation.

Emergence of intelligence

Lex Fridman
(00:25:53)
Before we forget, I got to ask you about this paper.
Adam Frank
(00:25:55)
Right.
Lex Fridman
(00:25:56)
It’s pretty interesting. There’s an interesting table here about hard steps. Abiogenesis, glucose fermentation to propionic acid, all kinds of steps, all the way to homo sapiens, animal intelligence, land ecosystems, endoskeletons. Eye precursor, so formation of the eye.
Adam Frank
(00:26:13)
Yeah.
Lex Fridman
(00:26:13)
Complex multicellularity.
Adam Frank
(00:26:17)
That’s definitely one of the big ones.
Lex Fridman
(00:26:18)
Yeah. Interesting. What can you say about this chart? There are all kinds of papers talking about, what, the difficulty of these steps?
Adam Frank
(00:26:26)
Right. This was the idea. What Carter said was, “We’re using anthropic reasoning.” He said, “There must be a few very hard steps for evolution to get through to make it to intelligence.” Some steps are going to be easy, so every generation, you roll the dice. Yeah, it won’t take long for you to get that step. But there must be a few of them, and he said you could even calculate how many there were, five, six, in order to get to intelligence.

(00:26:54)
This paper here, this plot is all these different people who’ve written all these papers. This is the point, actually. You can see all these papers that were written on the hard steps. Each one proposing a different set of what those steps should be. There’s this other idea from biology of the major transitions in evolution, MTEs, that those were the hard steps.

(00:27:13)
But what we actually found was that none of those are actually hard. The whole idea of hard steps, that there are hard steps, is actually suspect. What’s amazing about this model is it shows how important it is to actually work with people who are in the field. Brandon Carter was a brilliant physicist, the guy who came up with this. And then lots of physicists and astrophysicists like me have used this. But the people who actually study evolution and the planet were never involved.

(00:27:43)
If you went and talked to an evolutionary biologist or a bio-geophysicist, they’d look at you when you explained this to them and they’d be like, “What? What are you guys doing?” It turns out, none of the details, or none of the conceptual structure of this matches with what the people who actually study the planet and its evolution.
Lex Fridman
(00:28:06)
Is it mostly about the fact that there’s not really discrete, big steps? Is this a gradual, continual kind of process?
Adam Frank
(00:28:12)
Well, there’s two things. The first most important one was that the planet and the biosphere have evolved together.
Lex Fridman
(00:28:16)
Together.
Adam Frank
(00:28:17)
That’s something that most bio-geophysicists completely accept. It was the first thing that Carter rejected. He said, “No, that’s probably not possible.” And yet, if he’d only had more discussions with this other community, he would have seen, no, there are actually windows that open up.

(00:28:34)
Then the next thing is this idea of whether a step is hard or not. Because for hard, what we mean by a hard step is, like I said, every time there’s a generation, every time there’s a next generation born, you’re rolling the dice on whether this mutation will happen. The idea of something being a hard step, there’s two ways in which something might even appear as a hard step and not be. Or actually not be a hard step at all.

(00:28:56)
One is that you see something that has occurred in evolution that has only happened once. Let’s take the opposite, we see something that’s happened multiple times. Like wings, lots of examples of wings over lots of different evolutionary lineages. Making wings is not a hard step.

(00:29:12)
There’s certain other things that people say, “No, that’s a hard step.” Oxygen, the oxygen photosynthesis. But they tend to be so long ago that we’ve lost all the information. There could be other things in the fossil record that made this innovation, but they’re just gone now so you can’t tell, so there’s information loss.

(00:29:32)
The other thing is the idea of pulling up the ladder. That somebody, some species makes the innovation, but then it fills the niche and nobody else can do it again. Yeah, it only happened once but it happened once because basically, the creature was so successful it took over, and there was no space for anybody else to evolve it.

(00:29:49)
Yeah. The interesting thing about this was seeing how much, once you look at the details of life’s history on Earth, how it really shifts you away from this hard steps model. It shows you that those details, as we were talking about with do you have to know about the planet, do you have to know about plate tectonics? Yeah, you’re going to have to.
Lex Fridman
(00:30:07)
To be fair to Carter on the first point, it makes it much more complicated if life and the planet are co-evolving. Because it would be nice to consider the planet as a static thing that sets the initial conditions.
Adam Frank
(00:30:23)
Yeah.
Lex Fridman
(00:30:24)
And then we can, from an outside perspective, analyze planets based on the initial conditions they create. Then there’s a binary yes or no at will it create life. But if they co-evolve, it’s a really complex dynamical system, the way everything is … Because it’s much more difficult from the perspective of settee. Of looking out there and trying to figure out which ones are actually producing life.
Adam Frank
(00:30:50)
But I think we’re at the point now, now there may be other kinds of principles that actually … Co-evolution actually has its own. Not deterministic, you’re done with determinism.
Lex Fridman
(00:30:59)
Yeah.
Adam Frank
(00:31:00)
But complex systems have patterns.
Lex Fridman
(00:31:03)
Yeah.
Adam Frank
(00:31:03)
Complex systems have constraints. That’s actually what we’re going to be looking for, are constraints on them.

(00:31:10)
Again, nothing against Carter. It was a brilliant idea. But it just goes to show you … I’m a theoretical physicist. Give me a simplified model, with dynamical equations and some initial conditions, I’m very happy. But there’s this great XTC comic, where somebody’s working something out on the board, and this physicist is looking over and saying, ” Oh, oh, I just wrote down an equation for that. I solved your problem. Do you guys even have a journal for this?” The subtitle is Why Everybody Hates Physicists.
Lex Fridman
(00:31:37)
Yeah.
Adam Frank
(00:31:38)
Sometimes that approach totally works.
Lex Fridman
(00:31:40)
Yeah.
Adam Frank
(00:31:40)
Sometimes physicists, we can be very good at zooming in on what is important and casting the details aside so you can get to the heart of an issue. That’s very useful sometimes. Other times, it obfuscates. Other times, it clouds over actually what you needed to focus on, especially when it comes to complexity.

Drake equation

Lex Fridman
(00:32:02)
Speaking of simplifying everything down to an equation, let’s return back to the question of how many alien civilizations are out there and talk about the Drake Equation.
Adam Frank
(00:32:12)
Yeah.
Lex Fridman
(00:32:12)
Can you explain the Drake Equation?
Adam Frank
(00:32:15)
People have various feelings about the Drake Equation. It can be abused. The story actually is really interesting.

(00:32:23)
Frank Drake in 1960 does the first ever astrobiological experiment. He gets a radio telescope, points it at a couple of stars, and listens for signals. That was the first time anybody had done any experiment about any kind of life in the history of humanity. He does it, and he’s waiting for everybody to make fun of him. Instead, he gets a phone call from the government and says, “Hey, we want you to do a meeting on interstellar communications.” He’s like, “Okay.”

(00:32:51)
They organized a meeting with just eight people. A young Carl Sagan is going to be there as well. The night before, Drake has to come up with an agenda. How do you come up with an agenda for a meeting on a topic that no one’s ever talked about before? What he does, what’s so brilliant about the Drake Equation, is he breaks the problem of how many civilizations are there out there into a bunch of sub-problems. He breaks it into seven sub-problems. Each one of them is a factor in an equation that, when you multiply them all together, you get the number of civilizations out there that we could communicate with.

(00:33:28)
The first term is the rate at which stars form. The second term is the fraction of those stars that have plants, F-sub-P. The next term is the number of planets in the habitable zone, the place where we think life could form. The next term after that is the fraction of those planets where actually an abiogenesis event, life forms, occurs. The next one is the fraction of planets on which you start to get intelligence. After that, it’s the fraction of planets where that intelligence goes on to create a civilization. Then finally, the last term, which is the one that we really care about, is the lifetime, have a civilization and how long does it last.
Lex Fridman
(00:34:08)
When you say we, we humans?
Adam Frank
(00:34:09)
We humans, because we’re staring at multiple guns pointing at us.
Lex Fridman
(00:34:13)
Yeah.
Adam Frank
(00:34:14)
Nuclear war, climate change, AI. How long in general does civilizations last?

(00:34:20)
Now each one of these terms, what was brilliant about what he did was, what he was doing was he was quantifying our ignorance. By breaking the problem up into these seven sub-problems, he gave astronomers something to do. This is always with a new research field, you need a research program or else you just have a bunch of vague questions. You don’t even know really what you’re trying to do.

(00:34:41)
The star people could figure out how many stars were forming per year. The people who were interested in planets could go out and find techniques to discover planets, et cetera, et cetera.
Lex Fridman
(00:34:50)
These are their own fields. Essentially by creating this equation, he’s launching new fields.
Adam Frank
(00:34:56)
Yeah. That’s exactly … He gave astrobiology, which wasn’t even a term then, a roadmap. “Okay, you guys go do this, you go do that.”
Adam Frank
(00:35:03)
And then, a roadmap like, “Okay, you guys go do this, you go do that, you go do that.” And it had such far-reaching effect on astrobiology because it did break the problem up in a way that gave useful marching orders for all these different groups. For example, it’s because of the Drake equation in some sense that people who were involved in SETI pushed NASA to develop the technologies for planet hunting. There was this amazing meeting in 1978, two meetings, 1978 and 1979, that were driven in some part by the people who were involved in SETI getting NASA together to say, “Look, okay, look, what’s the roadmap for us to develop technologies to find planets?”

(00:35:45)
So, the Drake equation is absolutely foundational for astrobiology, but we should remember that it’s not a law of nature. It’s not equal to MC squared. And so, you can see it being abused in some sense. Yeah, it’s generated a trillion papers. Some of those papers are good, I’ve written some of those. And some of those papers are bad, I’m not sure where my paper fits in on those. I’m saying one should be careful about what you’re using it for. But in terms of understanding the problem that astrobiology faces, this really broke it up in a useful way.

Exoplanets

Lex Fridman
(00:36:20)
We could talk about each one of these, but let’s just look at exoplanets.
Adam Frank
(00:36:24)
Yeah.
Lex Fridman
(00:36:25)
So, that’s a really interesting one. I think when you look back hundreds of years from now, was it in the 90s when they first detected the first-
Adam Frank
(00:36:32)
Yeah. ’92 and ’95. ’95 to me was really, that was the discovery of the first planet orbiting a sun-like star. To me, that was the water, the dam being broken.
Lex Fridman
(00:36:40)
I think that’s one of the greatest discoveries in the history of science.
Adam Frank
(00:36:45)
I agree. I agree.
Lex Fridman
(00:36:46)
Right now, I guess nobody’s celebrating it too much because you don’t know what it really means. But I think once we almost certainly will find life out there, it will obviously allow us to generalize across the entire galaxy of the entire universe. So, if you can find life on a planet, even in the solar system, you can now start generalizing across the entire universe.
Adam Frank
(00:37:12)
You can, all you need is one. Right now, our understanding of life, we have one example. We have N equals one example of life. So, that means we could be an accident. It could be that we’re the only place in the entire universe where this weird thing called life has occurred. Get one more example and now you’re done, because if you have one more example, now you don’t have to find all the other examples. You just know that it’s happened more than once, and now you are from a Bayesian perspective, you can start thinking like, “Yeah. Life is not something that’s hard to make.”
Lex Fridman
(00:37:43)
Well, let me get your sense of estimates for the Drake equation. You’ve also written a paper expanding on the Drake equation, but what do you think is the answer?
Adam Frank
(00:37:51)
So, there was this paper we wrote, Woody Sullivan and I in 2016, where we said, “Look, we have all this exoplanet data now.” So, the thing that exoplanet science and the exoplanet census I was talking about before have nailed is F sub P, the fraction of stars that have planets, it’s one. Every fricking star that you see in the sky hosts a family of worlds. I mean, it’s mind-boggling because those are all places, right? They’re either gas giants, probably with moons, so the moons are places you can stand and look out. Or they’re like terrestrial worlds where even if there’s not life, there’s still snow falling and there’s oceans washing up on shorelines.

(00:38:33)
It’s incredible to think how many places and stories there are out there. So, the first term was F sub P, which is how many stars have planets. The next term is how many planets are in the habitable zone on average, and it turns out to be one over five, so around 0.2. So, that means you just count five of them go out at night and go one, two, three, four, five. One of them has an Earth-like planet in the habitable zone, like, whoa.
Lex Fridman
(00:39:00)
So, what defines a habitable zone?

Habitable zones

Adam Frank
(00:39:02)
Habitable zone is an idea that was developed in the 1958 by the Chinese American astronomer, Xu Sheng, and it was a brilliant idea. It said, “Look, I can do the simple calculation. If I take a planet and just stick it at some distance from a star of what’s the temperature of the planet? What’s the temperature of the surface?” So now, give it a standard Earth-like atmosphere and ask, “Could there be liquid water on the surface?” We believe that liquid water is really important for life. There could be other things that’s happening fine, but if you were to start off trying to make life, you’d probably choose water as your solvent for it.

(00:39:41)
So basically, the habitable zone is the band of orbits around a star where you can have liquid water on the surface. You could take a glass of water, pour it on the surface, and it would just pull up. It wouldn’t freeze immediately, which would happen if your planet is too far out and it wouldn’t just boil away if your planet’s too close in. So, that’s the formal definition of the habitable zone. So, it’s a nice strict definition, there’s probably way more going on than that, but this is a place to start.
Lex Fridman
(00:40:07)
Well, we should say it’s a place to start, I do think it’s too strict of a constraint.
Adam Frank
(00:40:11)
I would agree.
Lex Fridman
(00:40:12)
We’re talking about temperature where water can be on the surface. There’s so many other ways to get the aforementioned turmoil where the temperature varies, whether it’s volcanic, so interaction of volcanoes and ice and all of this on the moons of planets that are much farther away, all this kind of stuff.
Adam Frank
(00:40:33)
Yeah. Well, for example, we know in our own solar system we have, say Europa, the moon of Jupiter, which has got a hundred-mile-deep ocean under 10 miles of ice. That’s not in the habitable zone, that is outside the habitable zone, and that may be the best place. It’s got more water than Earth does, all of its oceans. It’s twice as much water on Europa than there is on Earth. So, that may be a really great for life to form, and it’s outside the habitable zone. So, the habitable zone is a good place to start and it helps us. And there’s reasons why you do want to focus on the habitable zone, because like Europa, I won’t be able to see from across telescopic distances across light years.

(00:41:12)
I wouldn’t be able to see life on Europa because it’s under 10 miles of ice. So, with the important thing about planets in the habitable zone is that we’re thinking they have atmospheres. Atmospheres are the things we can characterize across 10, 50 light years and we can see biosignatures as we’re going to talk about. So, there is a reason why the habitable zone becomes important for the detection of extra solar life.
Lex Fridman
(00:41:37)
But for me, when I look up at the stars, it’s very likely that there’s a habitable planet or moon in each of the stars, habitable defined broadly.
Adam Frank
(00:41:47)
Yeah, I think that’s not unreasonable to say, especially since the formal definition, you get one in five, right? One in five is a lot, there’s a lot of stars in the sky. So yeah, saying that in general, when I look at a star, there’s a pretty good chance that there’s something habitable orbiting it. It is not a unreasonable scientific claim.

Fermi Paradox

Lex Fridman
(00:42:06)
To me, it seems like there should be alien civilizations everywhere. Why the Fermi paradox? Why haven’t we seen them?
Adam Frank
(00:42:17)
Okay, the Fermi paradox. I love talking about the Fermi paradox because there is no Fermi paradox. Dun dun, dun dun. Yeah, so the Fermi paradox, let’s talk a about the Fermi paradox and the history of it. So, Enrico Fermi, it’s 1950, he’s walking with his friends at Los Alamos nuclear weapons lab to the Cantina, and there had been this cartoon in the New Yorker, they all read the New Yorker. And the cartoon was trying to explain why there had been this rash of garbage cans being disappearing in New York. And this cartoon said, “Oh, it’s UFOs.” Because it’s 1950, the first big UFO craze happened in ’47.

(00:42:55)
So, they were laughing about this as they’re walking, and they started being physicists, started talking about interstellar travel, interstellar propulsion. Conversation goes on for a while, conversation turns to something else, they’ve gone to other things. About 40 minutes later, over lunch, Fermi blurts out, “Well, where is everybody?” Typical Fermi sort of thing. He’d done the calculation in his head and he suddenly realized that, look, if intelligence is common, that even traveling at sub lights speeds a civilization could cross, hop from one star system to the other and spread it out across the entire galaxy in a few hundred thousand years.

(00:43:34)
And he realized this, and so he was like, “Why aren’t they here now?” And that was the beginning of the Fermi paradox. It actually got picked up as a formal thing in 1975 in a paper by Hart where he actually went through this calculation and showed and said, “Well, there’s nobody here now, therefore, there’s nobody anywhere.” Okay, so that is what we will call the direct Fermi paradox, why aren’t they here now? But something happened after SETI began, where people started to, there was this idea of the great silence. People got this idea in their head that like, “Oh, we’ve been looking for decades now for signals of extra-terrestrial intelligence that we haven’t found any. Therefore, there’s nothing out there.

(00:44:12)
So, we’ll call that the indirect Fermi paradox and there absolutely is no indirect Fermi paradox for the most mundane of reasons, which is money. There’s never been any money to look. SETI was always done by researchers who were scabbing some time, some extra time from their other projects to look a little bit at the sky where the telescope, telescopes are expensive. So, Jason Wright, one of my collaborators, he and his students did a study where they looked at the entire search space for SETI, and imagine that’s an ocean. All the different stars you have to look at, the radio frequencies you have to look at, how when you look, how often you look.

(00:44:49)
Then they summed up all the SETI searches that had ever been done, they went through the literature. And what they found was if that search space, if the sky is an ocean and you’re looking for fish, how much of the ocean have we looked at, and it turns out to be a hot tub. That’s how much of the ocean that we’ve looked up. We’ve dragged a hot tub’s worth of ocean water up and there was no fish in it, and so now are we going to say, “Well, there’s no fish in the ocean.” So, there is absolutely positively no indirect Fermi paradox, we just haven’t looked, but we’re starting to look. So finally, we’re starting to look, that’s what’s exciting.

(00:45:25)
The direct Fermi paradox, there are so many ways out of that. There’s a book called 77 Solutions to the Fermi Paradox that you can pick your favorite one. It just doesn’t carry a lot of weight because there’s so many ways around it. We did an actual simulation, my group, Jonathan Carroll, one of my collaborators, we actually simulated the galaxy and we simulated probes moving at sub light speed from one star to the other, gathering resources heading to the next one. And so, we could actually track the expansion wave across the galaxy, have one IA biogenesis event, and then watch the whole galaxy get colonized or settled. And it is absolutely true that wave crosses, Hart was right, Fermi was right, that wave crosses very quickly. But civilizations don’t last forever, so one question is when did they visit? When did they come to Earth? So, if you give civilizations a finite lifetime, let them last 10,000, 100,000 years, what you find is you now have a steady state. Civilizations are dying, they’re coming back, they’re traveling between the stars. What you find then is you can have big holes opened up. You can have regions of space where there is nobody for millions of years. And so, if we’re living in one of those bubbles right now, then maybe we revisited but we revisited 100 million years ago.

(00:46:39)
And there was a paper that Gavin Schmidt and I did that showed that if there was a civilization, whether it was dinosaurs or aliens that was here a 100 million years ago, there’s no way to tell, there’s no record left over, the fossil record is too sparse. The only way maybe you could tell is by looking at the isotopic strata to see if there was anything reminiscent of an industrial civilization. But the idea that you’d be able to find iPhones or toppled buildings after 100 million years is there’s no way.
Lex Fridman
(00:47:09)
So, if there was an alien camp here, an alien village, a small civilization, maybe even large civilizations?
Adam Frank
(00:47:17)
Even a large civilization, even if it was-
Lex Fridman
(00:47:19)
100 million years ago?
Adam Frank
(00:47:20)
And it lasted 10,000 years, fossil record’s not going to have it. Yeah, the fossil record is too sparse, most things don’t fossilize.
Lex Fridman
(00:47:28)
Yeah.
Adam Frank
(00:47:28)
And 10,000 years is a blink in the eye of geological time. So, Gavin called this the Silurian Hypothesis after the Doctor who episode with the lizard creatures, the Silurians. And so, that paper got a lot of press, but it was an important idea, and this was really Gavin’s, I was just helping with the astrobiology. That to recognize that like, “Yeah, we could have been visited a long time ago there just would be no record.” Yeah, it’s mind-blowing.
Lex Fridman
(00:47:56)
It’s really mind-blowing.
Adam Frank
(00:47:57)
Yeah.
Lex Fridman
(00:47:57)
And it’s also a good reminder that intelligent species have been here for a very short amount of time.
Adam Frank
(00:48:05)
Very short amount of time. Yeah. This is not to say that there was, so I was on Joe Rogan for exactly this paper, and I had to always emphasize, we’re not saying there was a Silurian, but we’re just saying that if there was, that’s why I love Gavin’s question. Gavin’s question was just like, “How could you tell”? It was a very beautifully scientific question. That’s what we were really showing is that unless you did a very specific kind of search, which nobody’s done so far, there’s not an obvious way to tell that there could have been civilizations here earlier on.
Lex Fridman
(00:48:40)
I’ve actually been reading a lot about ancient civilizations, and it just makes me sad how much of the wisdom of that time is lost and how much guessing is going on, whether it’s in South America, what happened in the jungle.
Adam Frank
(00:48:57)
Like the Amazon, that was the conquistadors came and wiped everybody out, and especially just even the plague may have decimated. So yeah, how much of that civilization.
Lex Fridman
(00:49:09)
And there’s a lot of theories, and because of archaeology only looks at cities, they don’t really know the origins of humans.
Adam Frank
(00:49:19)
Yeah.
Lex Fridman
(00:49:19)
And there’s a lot of really interesting theories, and there are of course controversial and there’s a lot of controversial people in every discipline, but archaeology is a fascinating one because we know so little. They’re basically storytellers, you’re assembling the picture from just very few puzzle pieces, and it’s fascinating. It’s humbling and it’s sad that there could be entire civilizations, ancient civilizations that are either almost entirely or entirely lost.
Adam Frank
(00:49:48)
Yeah. Well, the indigenous peoples of North America, there could have been millions and millions. We get this idea that like, oh, the Europeans came and it was empty. But it may have only been empty because the plague gets swept up from what happened in Mesoamerica, and they didn’t really build cities. They didn’t build wooden or stone cities, they built wooden cities.
Lex Fridman
(00:50:13)
Everybody seems to be building pyramids and they’re really damn good at it. I don’t know-
Adam Frank
(00:50:17)
What it is up with a pyramid. Why does that apply? What archetype in our brain is that?
Lex Fridman
(00:50:22)
And it is also really interesting, speaking of archetypes, is that independent civilizations formed and they had a lot of similar dynamics like human nature when it builds up hierarchies in a certain way, it builds up myths and religions in a certain way, it builds pyramids in a certain way. It goes to war, all this kind of stuff independently, which is fascinating.
Adam Frank
(00:50:48)
Santa Fe Institute, the stuff the Santa Fe Institute does on these as complex systems, the origin of hierarchies and such. Very cool.
Lex Fridman
(00:50:55)
Yeah, Santa Fe folks, complexity in general is really cool.
Adam Frank
(00:50:59)
Really cool.

Alien civilizations

Lex Fridman
(00:51:00)
What phenomena emerge when a bunch of small things get together and interact? Going back to this paper, a new empirical constraint on the prevalence of technological species in the universe. This paper that expands on the Drake equation, what are some interesting things in this paper?
Adam Frank
(00:51:16)
Well, so the main thing we were trying to do with this paper is say, “Look, we have all of this exoplanet data.” It’s got to be good for something, especially since two of the terms that have been nailed down empirically are two terms in the Drake equation. So, F sub P, that’s the second term, fraction of stars that have planets, and then N sub E, the average number of planets in the habitable zone. Those are the second and third term in the Drake equation. So, what that means is all the astronomical terms have been nailed. And so, we said, “Okay, how do we use this to do something with the Drake equation?”

(00:51:46)
And so, we realized is, “Well, okay, we got to get rid of time.” The lifetime thing, we can’t say anything about that, but if we don’t ask how long do they last but instead ask, “What’s the probability that there have been any civilizations at all?” No matter how long they lasted, I’m not asking whether they exist now or not, I’m just asking in general about probabilities to make a technological civilization anywhere and at any time in the history of the universe and that we were able to constrain. And so, what we found was basically that there have been 10 billion trillion habitable zone planets in the universe. And what that means is that those are 10 billion trillion experiments that have been run.

(00:52:35)
And the only way that we’re this whole process from a biogenesis to a civilization has occurred is if every one of those experiments failed. So therefore, you could put a probability, we called it the pessimism line. We don’t really know what nature sets for the probability of making intelligent civilizations, but we could set a limit using this. We could say, “Look, if the probability per habitable zone planet is less than 10 to the minus 22, 1 in 10 billion trillion, then yeah, we’re alone.” If it’s anywhere larger than that, then we’re not the first, it’s happened somewhere else. And to me, that was mind-blowing. It doesn’t tell me there’s anybody nearby, the galaxy could be sterile.

(00:53:17)
It just told me that unless nature’s really has some bias against civilizations, we’re not the first time this has happened. This has happened elsewhere over the course of cosmic history.
Lex Fridman
(00:53:29)
10 billion trillion experiments.
Adam Frank
(00:53:33)
Yeah, that’s a lot of experiments.
Lex Fridman
(00:53:35)
That’s a lot.
Adam Frank
(00:53:35)
Right.
Lex Fridman
(00:53:35)
1,000 is a lot.
Adam Frank
(00:53:36)
Yeah.
Lex Fridman
(00:53:36)
100 is a lot.
Adam Frank
(00:53:39)
Yeah.
Lex Fridman
(00:53:40)
If we, normal humans saw 100 experiments, and we knew that at least one time there was a successful human civilization built we would say for sure, in 100 you’ll get another one.
Adam Frank
(00:53:55)
Yeah. So, that’s why these kinds of arguments you have to be careful of what they can do. But I felt like what this paper showed was that the burden of proof is now on the pessimists. So, that’s why we called it the pessimism line. Throughout history, there’s been alien pessimists and alien optimists, and they’ve been yelling at each other, that’s all they had to go with. And with Giordano Bruno in 1600, they burned the guy at the stake for being an alien optimist. But nobody really knew what pessimism or optimism meant. We thought this was like the plank length, this was the plank length of astrobiology.

(00:54:27)
Gave you an actual number that if you could somehow calculate what the probability of forming a technological civilization was, this thing shows you where the limit is. As long as you’re above 10 to the minus 22, then you actually absolutely, it has occurred in the history. Other civilizations have occurred in the history of the universe.
Lex Fridman
(00:54:47)
So, to me, at least, the big question is FE, which is basically a biogenesis. How hard is it for life to originate in a planet? Because all the other ones seem very likely, everything seems very likely. The only open question to me is how hard is it for life to originate?
Adam Frank
(00:55:03)
There’s lots of ways to, again, we don’t know unless we look, and you had Sarah Walker around not too long ago, she’s very interested in origins of life. So, lots of people are working on this. But I think it’s hard looking at the history of the Earth, and again, you can do Bayesian arguments on this. But yeah, forming life I don’t think is hard. Getting basic biology started, I don’t think is hard. It’s still wild, it’s an amazing process that actually I think requires some deep rethinking about how we conceptualize what life is and what life isn’t. That’s one of the things I like about Sarah’s work, we’re pursuing on a different level about life as the only system that uses information. But still, regardless of all those kinds of details, life is probably easy to make. That’s my gut feeling.
Lex Fridman
(00:55:55)
Day by day, this changes for me, but I just see that once you create bacteria, it is off to the races. You’re going to get complex life as long as you have enough time. That boring billion, but I just can’t imagine a habitable planet not having a couple of billion to spare.
Adam Frank
(00:56:15)
Yeah, a couple billion years to spare. There is a mystery there about why did it take so long with the Cambrian explosion, but that may be again, about these windows. That it couldn’t happen until the window, the planet and the life had evolved together enough that they together opened the window for the next step. Intelligent life and how long intelligent, technological civilizations, I think there’s a big question about how long those last. And I’m hopeful, but in terms of just, I think life is absolutely going to be common, pretty common in the universe.
Lex Fridman
(00:56:52)
Yeah. I think, again, if I were to bet everything, even in advanced civilizations are common. So, to me then the only explanation is the L. Our galaxy is a graveyard of civilizations.
Adam Frank
(00:57:09)
Yeah. You think about it, we’ve only been around, truly when we think about in Drake’s definition, you had to have radio telescopes, that’s been 100 years. And if we got another 10,000, 100,000 years of history, for us, it’d be pretty amazing. But that still, that wouldn’t be long enough to really pop up the number of civilizations in the galaxy. So, you really need it to be hundreds of millions of years. And that raises a question, which I am very interested in, which is how do we even talk about, I call it the billion-year civilization. How do we even begin to hypothesize or think about in any kind of systematic way, what happens to a technological civilization across hundreds of millions to a billion years?
Lex Fridman
(00:57:52)
Yeah. How do you even simulate the trajectories as civilizations can take across that kind of timescale?
Adam Frank
(00:57:58)
Yeah.
Lex Fridman
(00:57:58)
When all the data we have is just for the 10,000 years or so, 20,000 years that humans have been building civilizations.
Adam Frank
(00:58:06)
Yeah.
Lex Fridman
(00:58:08)
And I don’t know what you put it at, but maybe 100 years that we’ve been technological?
Adam Frank
(00:58:12)
And we’re ready to blow ourselves to bits or drive ourselves off the planet. Yeah, no, it’s really interesting. But there’s got to be a way that I think that’s really a frontier. So, you had David Kipping on not too long ago, and David and I did a paper and Caleb Scharf, David really drove this. Where it was a Bayesian calculation to ask the question, “If you were to find a detection, if you were to find a signal or a techno signature, would that come from a civilization that was younger your age or older?” And you could see, this is not hard to do, but it was great. The formalism, the formalism was hard. It’s intuitive, but the formalism was hard to show that, yeah, they’re older, probably much older.

(00:58:49)
So, that means you really do need to think about like, “Okay, how do billion-year civilizations manifest themselves? What signatures will they leave?” And yeah, what’s so cool about it, it’s so much fun because you have to imagine the unimaginable. Obviously biological evolution can happen on those kinds of timescales, so you wouldn’t even really be the same thing you started out as. But social forms, what kind of social forms can you imagine that would be continuous over that? Or maybe they wouldn’t be continuous, should get they drop out, they destroy themselves, and then they come back. So, maybe it’s a punctuated evolution, but this is the fun part we have to work this out.
Lex Fridman
(00:59:31)
Well, one way to approach that question is what are the different ways to achieve homeostasis is you get greater and greater technological innovation. So, if you expand out into the universe and you have up to Kardashev scale, what are the ways you can avoid destroying yourself? Just achieve stability while still growing. That’s an interesting question, I think it’s probably simulatable?
Adam Frank
(01:00:00)
Could be, agent-based modeling you could do it with. So, our group has used agent-based modeling to do something like the Fermi paradox that was agent-based modeling. But you can also do this. People at Santa Fe have done this, other groups have done this to do use agent-based modeling to track the formation of hierarchies, the formation of stable hierarchies. So, I think it’s actually very doable, but understanding the assumptions and principles that are going into it and what you can extract from those, that is what is the frontier.

Colonizing Mars

Lex Fridman
(01:00:32)
Do you think if humans colonize Mars, the dynamic between the civilization on Earth and Mars will be fundamentally different than the dynamic between individual nations on Earth right now? That’s a thing to load into the agent-based simulation we’re talking about.
Adam Frank
(01:00:50)
Yeah. If we settle it, Mars will very quickly want to become its own nation.
Lex Fridman
(01:00:53)
Well, no, there’s already going to be nations on Mars that’s guaranteed-
Adam Frank
(01:00:58)
Yeah. And they’re there on-
Lex Fridman
(01:00:59)
2 million people. The moment you have 1 million people, there’s going to be two tribes.
Adam Frank
(01:01:03)
Right.
Lex Fridman
(01:01:04)
And then they’re going to start fighting.
Adam Frank
(01:01:06)
Right.
Lex Fridman
(01:01:06)
And the question is, interplanetary fighting. How quickly does that happen and does it have a different nature to it because of the distances?
Adam Frank
(01:01:14)
Are you a fan of The Expanse? Have you watched The Expanse? Great show, I highly recommend to everybody. It’s based on a series of books that are excellent. It’s on Prime, six seasons, and it’s basically about the settled solar system. It takes place about 300 years from now, and the entire solar system is settled, and it is the best show about interplanetary politics. The first season, actually, the journal, what was it? Foreign Affairs said the best show on TV about politics it takes place is interplanetary. So yeah, I think human beings being human beings, yes, there will be warfare and there will be conflict.

(01:01:49)
And I don’t think it’ll be necessarily all that different because really I think within a few hundred years we will have lots of people in the solar system, and it doesn’t even have to be on Mars. We did a paper where we look based on, because I always wanted to know about whether an idea in The Expanse was really possible. In The Expanse, the asteroid belt, what they’ve done is they have colonized the asteroid belt by hollowing out the asteroids and spinning them up and living on the inside because they have the Coriolis force. And I thought like, “Wow, what a cool idea.”

(01:02:17)
And when I ran the blog for NPR, actually talked to the guys and said, “Did you guys calculate this to see whether it’s possible?” Sadly, it’s not possible. The rock is just not strong enough that if you tried to spin it up to the speeds you need to get one third gravity, which is what I think the minimum you need for human beings. The rock would just fall apart, it would break. But we came up with another idea, which was that if you could take small asteroids, put a giant bag around them, a nanofiber bag and spin those up, it would inflate the bag. And then even a small couple of kilometer wide asteroid would expand out to, you could get a Manhattan’s worth of material inside.

(01:02:54)
So, forget about even colonizing Mars space stations or space habitats with millions of people in them. So anyway, the point is that I think within a few hundred years, it is not unimaginable that there will be millions, if not billions of people living in the solar system.
Lex Fridman
(01:03:11)
You think most of them will be in space habitats versus on Mars on the planetary surface?
Adam Frank
(01:03:16)
It’s a lot easier on some level. It depends on how with nanofabrication and such, but getting down to gravity well is hard. So, there’s a certain way in which it’s a lot easier to build real estate out of the asteroids, but we’ll probably do both. I think what’ll happen is the next, should we make it through climate change and nuclear war and all the other, and AI? The next 1,000 years of human history is the solar system. And so, I think we’ll settle every nook and cranny we possibly can, and what I love about, what’s hopeful about it is this idea you’re going to have all of these pockets, and I’m sure there’s going to be a Mormon space habitat.

(01:03:57)
Whatever you want, a libertarian space habitat, everybody’s going to be able to create, there’ll be lots of experiments in human flourishing. And those kinds of experiments will be really useful for us to figure out better ways for us to interact and have maximum flourishing, maximum wellness, maximum democracy, maximum freedom.
Lex Fridman
(01:04:15)
Do you think that’s a good backup solution to go out into space, so to avoid the possibility of humans destroying themselves completely here on Earth?
Adam Frank
(01:04:24)
Well, I think I want to be always careful with that, because like I said, it’s centuries that we’re talking about. So, the problem with climate change, and same thing with nuclear war, it’s breathing down our necks now. So, trying to establish a base on Mars it’s going to be so hard that it is not even going to be close to being self-sufficient for a couple a century at least. So, it’s not like a backup plan now, we have to solve the problem of climate change, we have to deal with that. There’s still enough nuclear weapons to really do horrific things to the planet for human beings.

(01:04:59)
So, I don’t think it’s a backup plan in that way, but I do think, like I said, it’s the prize. If we get through this, then we get the entire solar system to play around and experiment with and do really cool things with.
Lex Fridman
(01:05:11)
Well, I think it could be a lot less than a couple of centuries if there’s a urgency, a real urgency, like a catastrophe. Maybe a small nuclear war breaks out where it’s like, holy shit, this is for sure a bigger one is looming. Maybe if geopolitically the war between China and the United States escalates where there’s this tension that builds and builds and builds and it becomes more obvious that we need to really, really [inaudible 01:05:39].
Adam Frank
(01:05:39)
Yeah. I think my only dilemma with that is that I just think that a self-sufficient base is so far away. That say you start doing that and then there is a full-scale nuclear exchange that base is, it’s not going to last because the self-sufficiency requires a kind of economy. Literally a material economy that we are so far from with Mars that we are centuries from. Like I said, three centuries, which is not that long, two to three centuries. Look at 1820, nobody had traveled faster than 60 miles an hour unless they were falling off a cliff. And now we routinely travel at 500 miles an hour, but it is centuries long.

(01:06:17)
So, that’s why I think we’d be better off trying to solve these problems than I just think the odds that we’re going to be able to create a self-sufficient colony on Mars before that threat comes to head is small. So, we’d have to deal with the threat.
Lex Fridman
(01:06:35)
That’s an interesting scientific and engineering question of how to create a self-sufficient colony on Mars or out in space as a space habitat where Earth entirely could be destroyed, you could still survive.
Adam Frank
(01:06:47)
Yeah. Because it’s really what about, thinking about complex systems? A space habitat would have to be as robust as an ecosystem. As the kind of thing, you go out and you see a pond with all the different webs of interactions. That’s why I always think that if this process of going out into space will help us with climate change and with thinking about making a long-term sustainable version of human civilization. Because you really have to think about these webs, the complexity of these webs and recognize the biosphere has been doing this forever. The biosphere knows how to do this.

(01:07:23)
And so, A, how do we build a vibrant, powerful technosphere that also doesn’t mess with the biosphere, mess with the biosphere’s capacity to support our technosphere? So, by trying to build space habitats, in some sense, you’re thinking about building a small-scale version of this. So, I think the two problems are going to feedback on each other.
Lex Fridman
(01:07:44)
Well, there’s also the other possibility of the movie Darren Aronofsky’s Postcard from Earth, where we can create this life gun that just shoots as opposed to engineering everything. Basically, seeding life on a bunch of places and letting life do its thing, which is really good at doing it seems like. So, as opposed to with a space habitat, you basically have to build the entire biosphere and technosphere, the whole thing-
Adam Frank
(01:08:13)
The whole thing.
Lex Fridman
(01:08:13)
… by yourself. If you just, hey, the aforementioned cockroach with some bacteria, place it in Europa, I think you’d be surprised what happens.
Adam Frank
(01:08:25)
Yeah.
Lex Fridman
(01:08:25)
Honestly, if you put a huge amount of bacteria, a giant number of organisms from Earth into on Mars, on some of these moons of the other planets in the solar system, I feel like some of them would actually find a way to survive.
Adam Frank
(01:08:45)
The moon is hard, the moon may be really hard. But I wonder if somebody must’ve done these experiments. Because we know they’re extremophiles, we know that you can go down 10 miles below the Earth’s surface. And there are things where there’s no sunlight, the conditions are so extreme and there’s lots of microbes having a great time living off the radioactivity in the rocks. But they had lots of time to evolve to those conditions, so I’m not sure if you dumped a bunch of bacteria, so somebody must’ve done these experiments. How fast could microbial evolution occur in under harsh conditions that you maybe get somebody who figures out, ” Okay, I can deal with this.”

(01:09:30)
I think the Moon’s too much because it’s so sterile. But Mars, I don’t know, maybe. I don’t know, but it’s an interesting idea.
Lex Fridman
(01:09:37)
I wonder if somebody has done those experiments.
Adam Frank
(01:09:39)
Yeah, you think somebody would, let’s take a bunch of microbes-
Lex Fridman
(01:09:43)
The harshest possible condition of all different kinds, temperature, all this kind of stuff.
Adam Frank
(01:09:46)
Right, pressure, salinity, and then just dump a bunch of things that are not used to it, and then just see, does everybody just die? That’s it.
Lex Fridman
(01:09:55)
The thing about life, it flourishes in a non-sterile environment where there’s a bunch of options for resources, even if the condition is super harsh-
Lex Fridman
(01:10:03)
… Options for resources, even if the condition is super harsh. In the lab, I don’t know if you can reconstruct harsh conditions plus options for survival. You know what I mean? You have to have the huge variety of resources that are always available on a planet somehow, even when it’s a super harsh condition. So that’s actually not a trivial experiment and if somebody did that experiment in the lab, I’d be a little bit skeptical because I could see bacteria doesn’t survive in this kind of temperature. But then I’d be like, “I don’t know. I don’t know.”
Adam Frank
(01:10:38)
Right. Are there other options? Is the condition rich enough?
Lex Fridman
(01:10:41)
Rich enough, yeah.
Adam Frank
(01:10:42)
There’s an alternative view though, which is, there’s this great book by Kim Stanley Robinson called Aurora. So there’s been 1,000,000 sentry ship stories where Earth sends out a generation ship or sentry ship, and it goes to another planet and they land and they colonize. And on this one, they get all the way there and they think the planet’s going to be habitable. And it turns out that it’s not habitable for earth life. There’s bacteria or prions actually, that just kill people in the simplest way. And the important thing about this book was the idea that life is actually very tied to its planet. It may not be so easy. I just thought it was a really interesting idea. I’m not saying necessarily supporting it, but that actually, life reflects the planetary conditions… Not the planetary, the planet itself, the whole lineage, the whole history of the biosphere. And it may not be so easy to just be like, “Oh, just drop it over here and it’ll…”

(01:11:35)
Because the bacteria, even though they’re individual examples of life, and I believe this the true unit of life, it’s not DNA, it’s not a cell, it’s the biosphere. It’s the whole community.
Lex Fridman
(01:11:46)
Yeah. That’s actually an interesting field of study is how when you arrive from one planet to another… So we humans arrive to a planet that has a biosphere, maybe a technosphere, what is the way to integrate without killing yourself or-
Adam Frank
(01:12:06)
Or the other one?
Lex Fridman
(01:12:06)
Or the other one? Let’s stick to biology. That’s an interesting question. I don’t know if we have a rigorous way of investigating that.
Adam Frank
(01:12:18)
Because everything on life has the same lineage. We all come from LUCA, the last universal common ancestor. And what you see is often in science fiction, people will do things like, “Oh, well, it’s okay,” because that metabolism, that biochemistry is so different from ours that we can coexist because they don’t even know each other.
Lex Fridman
(01:12:35)
Right.
Adam Frank
(01:12:37)
And then the other version is you get there, you land, and instantly, the nose bleeds and you’re dead. So it’s-
Lex Fridman
(01:12:43)
Unfortunately, I think it’s the latter.
Adam Frank
(01:12:44)
Yeah, it feels like the alien kind of thing.

Search for aliens

Lex Fridman
(01:12:48)
So as we look out there, according to the Drake equations we just discussed, it seems impossible to me that there’s not civilizations everywhere. So how do we look at them, this process of SETI?
Adam Frank
(01:12:59)
I have to put on my scientist hat and just say, my gut feeling is that dumb life, so to speak, is common. I can see ways in which intelligent civilizations may be sparse, but until… We got to go look, it’s all armchair astronomy.
Lex Fridman
(01:13:15)
That’s from a rigorous scientific perspective. From my bro science perspective, it seems, again, smoking the aforementioned weed-
Adam Frank
(01:13:24)
Smoking the weed, yeah. After the bong hit, it seems so.
Lex Fridman
(01:13:28)
Honestly, it really just seems impossible to me that there’s not potentially dead, but advanced civilizations everywhere in our galaxy.
Adam Frank
(01:13:37)
Yeah, yeah. The potentially dead part, I think, right. It could be that making civilizations is easy, they just don’t last long. So when we went out there, we’d find a lot of extinct civilizations.
Lex Fridman
(01:13:45)
Extinct civilizations. Yeah. Apex predators don’t survive. They get better, better, better.
Adam Frank
(01:13:51)
Right.
Lex Fridman
(01:13:51)
And they die, kill themselves all somehow. Anyway. So just how do we find them?
Adam Frank
(01:13:56)
Yeah. So SETI, Search for Extraterrestrial Technology is a term that I am not fond of using anymore. Some people in my field are. So I’m sorry folks, but what I really like is the idea of technosignatures because I think to me, SETI is the… First of all, intelligence. We’re not really looking for intelligence. We’re looking for technology, and SETI, the classic idea of SETI is the radio telescopes and contact, Jodie Foster with the headphones. That whole thing is still part, it’s still active, there’s still great things going on with it, but suddenly, this whole new window opened up. When we discovered exoplanets, we now found a new way to look for intelligence civilizations or life in general in a way that doesn’t have any of the assumptions that had to go into the classic radio SETI. And specifically, what I mean is we’re not looking for somebody sending us a beacon. You really needed that with a classic model, for a bunch of different reasons. You have to assume they wanted to be found and they were sending you a super powerful beacon.

(01:14:56)
Now, because we know exactly where to look and we know exactly how to look, we can just go about looking for passive signatures of the civilization, going about its civilizationing business, without asking whether they want to be contacted or not. So this is what we call a biosignature or a technosignature. It is an imprint in the light from the planet of the activity of a biosphere or a technosphere, and that’s really important. That is why the whole Gaia idea ends up being astrobiological, that biospheres and technospheres are so potent, they change the entire planet, and you can see that from 20 light years.

(01:15:36)
So let’s give an example of a biosignature to start off with, which would be a signature of a biosphere, oxygen. Right? On earth at least, we know that oxygen is only in the atmosphere because life put it there. If life went away, the oxygen, and particularly oxygen and methane, that pair, they would disappear very quickly. They’d react away. They’d all be gone. So if you find a planet with oxygen and methane, that’s a good bet that there’s a biosphere there. Okay, what about technospheres? Technospheres, so I’m the principal investigator on the first grant NASA has ever given to do these exoplanet technosignatures. For reasons we can talk about, NASA had gotten pretty gun-shy about funding anything about intelligent life, but okay. What’s an example of a technosignature? Well, one could be atmospheric, “Pollution.” I’m going to put, “Pollution,” in quotes here because it doesn’t have to be pollution, but gases like chlorofluorocarbons.

(01:16:33)
So we dumped a huge amount of chlorofluorocarbons into the atmosphere by mistake. It was affecting the ozone, but we put so much in there that actually, this is one of the things we did, we did a paper where we showed, you could detect it across interstellar distances. You could look at the atmosphere, look at the light coming from a distant planet, pass the light through a spectrograph and see the spectral lines, the fingerprint, the spectral fingerprint of chlorofluorocarbons in an atmosphere. And that would for sure tell you that there was a technological civilization there, because there’s no other way to make chlorofluorocarbons except through some kind of industrial process.
Lex Fridman
(01:17:11)
So in the case of the biosphere, you’re looking for anomalies in the spectrograph?
Adam Frank
(01:17:17)
I wouldn’t necessarily call these anomalies. For biosignature, I’m looking for things that a geosphere, right? That just rock and air wouldn’t produce on its own.
Lex Fridman
(01:17:28)
What kind of chemicals would life produce?
Adam Frank
(01:17:29)
Right. And that’s the interesting thing. So we can use earth as an example. We can say, look, oxygen. We know there would be no oxygen in the atmosphere if it wasn’t for dimethyl sulfide, which is a compound that phylloplankton dump into the atmosphere, a lot of it, that’s sometimes mentioned. And there was a paper that somebody wrote where it was like, “Well, we’re not saying we see it, but there’s a bunch of noise in the spectra right there.” So there’s a whole list of things that earth has done that are in the atmosphere that might be biosignatures, but now we’re reaching an interesting point. The field has matured to the point where we can start asking about agnostic biosignatures, things that have nothing to do with earth’s history, but we think that would still be indications of this weirdness we call life. What is it in general that life does that leaves an imprint?

(01:18:20)
So one of these things could be the structure of the network of chemical reactions that biology always produces very different chemical networks, who’s reacting with who, than just rock and water. So there’s been some proposals for networked biosignatures. Information theory, you can try and look at the information that is in the different compounds that you find in the atmosphere, and maybe that information shows you like, “Oh, there’s too much information here. There must’ve been biology happening. It’s not just rock.” Same thing for techno. That’s what we’re working on right now, for technosignatures as well.
Lex Fridman
(01:18:58)
So how do you detect technosignatures?
Adam Frank
(01:19:01)
Okay. So with technosignatures, I gave the example of chlorofluorocarbons. So that would be an example of, and again, that one is a non-agnostic one, because we sort of like, “Oh, we produced chlorofluorocarbons. Maybe they will.” And there’s solar panels. The glint off of solar panels will produce the way the light is reflected off of solar panels, no matter what it’s made out of actually. There was a paper that Manasvi Lingam and Avi Loeb did in… I think it was 2017. We’ve just followed up on it. That actually could act as a technosignature. You’d be able to see in the reflected light this big jump that would occur because of… City lights, artificial illumination. If there’s really large scale cities like Coruscant and Star Wars or Trantor in the foundation, those city lights would be detectable, the spectral imprint of those across 20, 30 light years.

(01:19:55)
So our job in this grant is to develop the first ever library of technosignatures. Nobody’s really ever thought about this before. So we’re trying to come up with all the possible ideas for what a civilization might produce that could be visible across interstellar distances. And are these good ones or are these ones going to be hard to detect or such?
Lex Fridman
(01:20:17)
City lights. So if a planet is all lit up with artificial light across 20 to 30 light years, we can see it.
Adam Frank
(01:20:25)
Yeah. If you looked at earth at night from a distance, looked at spectra and you had sensitive enough instruments, you’d be able to see all the sodium lights and the reflected light off of. They bounce off the ground, the light bounces off the ground. So you’d convolve the sodium lamps with the reflected spectra from the ground. And yeah, you’d be able to see that there’s city lights. Now, increase that by a factor of 1,000 if you had a trantor, and you’d be able to detect that across interstellar distances. Thomas Beatty did this work, who’s now working with us.
Lex Fridman
(01:20:56)
What do you think is the most detectable thing about earth?
Adam Frank
(01:21:01)
Wow, this is fun. We just have Sophia Sheikh, who’s part of our collaboration, just did a paper. We did earth from earth. If you were looking at earth with earth technology for a bunch of different technosignatures, how close would you have to be to be able to detect them? And most of them turn out to be… You’d have to be pretty close, at least out to the Oort cloud, but actually, it is our radio signatures still, that is still most detectable.
Lex Fridman
(01:21:23)
By the way, when you said you had to be pretty close and then you said the Oort cloud, that’s not very close. But you mean from an interstellar perspective.
Adam Frank
(01:21:29)
Interstellar distance, because we really want to know is I’m sitting here on earth, I’m looking at these exoplanets, the nearest star is four light years away. So that’s the minimum distance. So if I’m looking at exoplanets, what kind of signals could I see?
Lex Fridman
(01:21:44)
What is detectable about earth with our current technology from our nearest solar system?
Adam Frank
(01:21:49)
Oh my God, there’s all kinds of stuff. Well, like the chlorofluorocarbons, you can see earth’s pollution, and I think city lights, you had to be within the solar system.
Lex Fridman
(01:22:01)
If they do direct imaging of earth-
Adam Frank
(01:22:04)
They’re going to need much more powerful, but let me tell you, let’s talk about direct imaging for a moment because I just have to go on, this is such a cool idea. So what we really want, and the next generation of space telescopes and such is we’re trying to do direct imaging. We’re trying to get an image of a planet separated from its star to be able to see the reflected light or the actual emission from the planet itself.
Lex Fridman
(01:22:24)
By the way, just to clarify, direct imaging means literally a picture?
Adam Frank
(01:22:29)
A picture, but the problem is that even with the thing that’s going to come after JWST, it’s going to be a pixel. You’re not going to get any kind of resolution. You’ll be able to get the light from it, which you’ll be able to pass through a spectrograph, but you’re not going to be able to take a picture. But there is this idea called the solar gravity lens telescope, I think that’s what it is. And the idea is insane. So their general relativity says, “Look, massive bodies distort space. They actually curve space-time.” So the sun is a massive body, and so that means that the light passing through the sun gets focused like a lens. So the idea is to send a bunch of telescopes out into the Oort cloud, and then look back towards the sun towards an exoplanet that is behind… Not directly behind the sun, but is in the direction of the sun.

(01:23:16)
And then let the sun act like a lens and collect, focus the light onto the telescope and you would be able to get, and they’ve done… It’s amazing. This idea is insane. They’d be able to get, if everything works out, 24 kilometer resolution. You’d be able to see Manhattan on an exoplanet. And this thing, it sounds insane, but actually, NASA, the team has already gotten through three levels of NASA… There’s the NASA program for, “Give us your wackiest idea.” And then the ones that survive that are like, “Okay, tell us whether that wacky idea is even feasible?” And they’re marching along. And the idea is that they even have plans for how you’d be able to get these probes out into the Oort cloud on relatively fast time scales. You need to be about 500 times as far from the sun as earth is, but right now, the idea seems to hold together.

(01:24:10)
So probably when I’ll be dead, but when you’re an old man, it’s possible that something like this… Could you imagine having that kind of resolution, a picture of an exoplanet down to kilometers? So I’m very excited about that [inaudible 01:24:26].
Lex Fridman
(01:24:25)
I can only imagine having a picture like that, and then there’s some mysterious artifacts that you’re seeing.
Adam Frank
(01:24:33)
Yeah.
Lex Fridman
(01:24:34)
It’s both inspiring and almost heartbreaking that we can see. I think we would be able to see a civilization where there’s a lot of scientists agree that this is very likely something and then we can’t-
Adam Frank
(01:24:49)
We can’t get there. But again, this is the thing about being long-lived. We’ve got to get to the point where we’re long-lived enough that… Let’s imagine that we find, say 10 light years away, we find a planet that looks like it’s got technosignatures. Righ? It doesn’t end there. That would be the most important discovery in the history of humanity, and it wouldn’t be like, “Well, okay, we’re done.” The first thing we do is we build bigger telescopes to try and do those imaging. And then the next thing after that, we plan a mission there. We would figure out, with Breakthrough Starshot, there was this idea of trying to use giant lasers to propel small spacecrafts, light sails, almost to the speed of light. So they would get there in 10 years and take pictures. So if we actually made this discovery, there would be the impulse. There would be the effort to actually try and send something to get there.

(01:25:42)
Now, we probably couldn’t land, so maybe we take 30 years to build, 10 years to get there, 10 years to get the picture back. Okay, you’re dead, but your kids are… You know what I mean? So it becomes now this multi-generational project. How long did it take to build the pyramids? How long did it take to build the giant cathedrals? Those were multi-generational projects, and I think we’re on the cusp of that kind of project.
Lex Fridman
(01:26:07)
I think that would probably unite humans.
Adam Frank
(01:26:09)
I think it would play a big role. I think it would be helpful. Human beings are a mess, let’s face it. That’s why I always say to people, discovery of life, of any kind of life, even if it was microbial life, it wouldn’t matter, that to know that we’re not an accident, to know that there is probably… If we found one example of life, we’d know that we’re not an accident and there’s probably lots of life and that we’re a community. We’re part of a cosmic kind of community of life, and who knows what life has done? All bets are off with life.
Lex Fridman
(01:26:36)
Since we’re talking about the future of telescopes, let’s talk about our current super sexy, awesome telescope, the James Webb Space Telescope, that I still can’t believe actually worked.
Adam Frank
(01:26:46)
I can’t believe it worked either. I was really skeptical. I was like, “Okay, guys. All right, sure.”
Lex Fridman
(01:26:51)
We only got one shot for this incredibly complicated piece of hardware to unfold. So what kind of stuff can we see with it? I’ve been just looking through different kinds of announcements that have been detected. There’s been some direct imaging-
Adam Frank
(01:27:06)
Yes, like a single pixel.
Lex Fridman
(01:27:07)
The kinds of exoplanets were able to direct image I guess would have to be hot.
Adam Frank
(01:27:13)
Hot, reasonably far away from the star. I think JWST is really at the hairy edge of being able to do much with this. What’s more important I think, for JWST is the spectra. And the problem with spectra is that there’s not sexy pictures. It’s like, “Hey, look at this wiggly line,” but be able to find and characterize atmospheres around terrestrial exoplanets is the critical next step. That’s where we are right now. In order to look for life, we need to find planets with atmospheres. And then we need to be able to do this thing called characterization, where we look at the spectral fingerprints for what’s in the atmosphere. Is there carbon? Is there carbon dioxide? Is there oxygen? Is there methane? And that’s the most exciting thing.

(01:27:54)
For example, there was this planet K2-18b, which they did a beautiful job getting the spectra, and the spectra indicated it may be an entirely new kind of habitable world called a hycean world, hycean meaning hydrogen ocean world. And that is a kind of planet that it would be in the super earth, sub-Neptune domain we were talking about, maybe eight times the mass of the earth. But it’s got a layer of hydrogen, of an atmosphere of hydrogen. Hydrogen is an amazing greenhouse gas. So hydrogen will keep the planet underneath it warm enough that you could get liquid water, you can get a giant ocean of liquid water, and that’s an entirely different kind of planet. That could be habitable planet. It could be a 60 degree warm ocean.

(01:28:41)
So the data that came out of JWST for that planet was good enough to be able to indicate like, “Oh yeah, you know what? From what we understand with the models, this looks like it could be a hycean world.”
Lex Fridman
(01:28:54)
And it’s 120 light years away from earth.
Adam Frank
(01:28:57)
And so isn’t that amazing? It’s 120 light years away, but we can see into the atmosphere. We can see to the atmosphere so well that we can be like, “Oh, look, methane.” Methane was a five sigma detection. You knew that the data were so good that it was the gold standard of science.

Alien megastructures

Lex Fridman
(01:29:13)
What about detecting maybe through direct imaging or in other ways, megastructures, that the civilizations build?
Adam Frank
(01:29:24)
You know what’s great about megastructures is first of all, it’s fun to say, who doesn’t want to say megastructure? Alien, megastructure, right? Every morning, I’m looking for an opportunity to say that. So the err example of this is the Dyson sphere, which is amazing because it was literally 1960 that this idea came up.
Lex Fridman
(01:29:39)
Can you explain the Dyson sphere?
Adam Frank
(01:29:40)
Yeah, the Dyson sphere. So Freeman Dyson, one of the greatest physicists ever, who was very broad-minded and thought about a lot of different things. He recognized that as civilizations progress, what they’re going to need is ever more energy to do ever more amazing things. And what’s the best energy source in a solar system? It’s the star. Right? So if you surrounded the star with solar collecting machines, sunlight collecting machines… Anyway, the limit of this would actually build a sphere, an actual sphere around your star that had all solar panels on the inside. You could capture every photon the star produced, which is this insane amount of light. You would have enough power now to do anything to re-engineer your solar system. So that was a Dyson sphere.

(01:30:25)
It turns out that a Dyson sphere doesn’t really work, it’s unstable, but a Dyson swarm, and that’s really what he meant, this large collection of large orbiting structures that were able to collect light.
Lex Fridman
(01:30:37)
So he didn’t actually mean a rigid sphere structure.
Adam Frank
(01:30:42)
Right.
Lex Fridman
(01:30:42)
He basically meant a swarm. So like you said, then the limit basically starts to look-
Adam Frank
(01:30:48)
People started to say, “Yeah, it was like a sphere.” And we actually almost thought we might’ve found one of these back with a Bajoyan star. The way we detect planets is through the transit method where the planet passes in front of the star and there’s a dip in the starlight. It’s a little eclipse basically, and we know exactly what they should look like. And then with this one star, there were these really weird transits where it was like this little dragon’s tooth, and then there’d be another one and another one and another one, and then nothing, and then three more. And in the paper that was written about this, they went through the list of, it could be comets, it could be chunks of a broken up planet, and it could also be an alien megastructure. And of course, the news picked up on this and everybody’s newsfeed the next day, “Alien megastructures discovered.”

(01:31:31)
Turns out, sadly, they were not alien megastructures. They were probably gas or dust clouds, but it raised the possibility like, “Oh, these are observable.” And people have worked out the details of what they would look like. You don’t really need direct imaging. You can do transits, right? They’re big enough that when they pass in front of the star, they’re going to produce a little blip of light because that’s what they’re supposed to. They’re absorbing starlight. So people have worked out like, “Well, a square one or a triangular one.”
Lex Fridman
(01:31:55)
But that wouldn’t be a distance sphere. That would be like one object.
Adam Frank
(01:31:58)
One object, right. If it’s a swarm, you’d expect the light to be blinking in and out as these things pass in front of… If you’ve got thousands of these, much of the time, they’ll be blotting out the star. Sometimes they won’t be. Right? And so you’re going to get an irregular transit signal.
Lex Fridman
(01:32:15)
One you wouldn’t expect from a star that doesn’t have anything.
Adam Frank
(01:32:18)
Exactly. Or just a planet or a couple of planets. There’d be so many of these that it would be like, “Beep, beep, blip, blip, blip, blip, blip.”
Lex Fridman
(01:32:24)
And that usually doesn’t happen in a star system because there’s only just a handful of planets.
Adam Frank
(01:32:31)
That’s exactly what it is. Everything’s coagulant. In a stable solar system, you get a handful of planets, five, 10, that’s it probably, and nothing else. So if now suddenly you see lots of these little micro transits telling you there’s something else that’s big enough to create a transit, but too many of them, and also, within a regular shape, the transit itself, that these could be megastructures.
Lex Fridman
(01:32:54)
How many people are looking for megastructures now?
Adam Frank
(01:32:58)
Well, the main groups looking for megastructures are again, Jason Wright at Penn State, and collaborators. The way they’re looking for it though is for infrared light because the second law of thermodynamics says, “Look, if you capture all of this starlight, your thing’s going to warm up and emit an infrared.” It’s going to be waste heat, waste heat and waste light from this.
Lex Fridman
(01:33:22)
That feels like a louder, clearer way to detect it.
Adam Frank
(01:33:25)
Right. And that’s actually why Dyson proposed it. He wasn’t really proposing it because he was saying, “This is what civilizations are going to do.” He proposed it because he was like, “Oh, we want to start looking for alien civilizations. Here’s something that would have a detectable signature.” So Jason and company have done pretty good searches, and recently, they made news because they were able to eliminate a lot of places. “No, these are not Dyson Spheres,” but they did have a couple that were anomalous enough that they’re like, “Well, this is what it would look like.” It’s not a detection. They were saying they would never say it’s a detection, but they were not non-detections.
Lex Fridman
(01:34:00)
And they’re potential candidates.
Adam Frank
(01:34:01)
Potential candidates, yeah.
Lex Fridman
(01:34:02)
Love it. We have megastructure candidates. That’s inspiring. What other megastructures do you think that could be? So Dyson Sphere is about capturing the energy of a star.
Adam Frank
(01:34:12)
Yeah.
Lex Fridman
(01:34:13)
There could be other-
Adam Frank
(01:34:14)
Well, there’s something called the Clark Belt. So we have a bunch of satellites that are in geosynchronous orbit. Nothing naturally is going to end up in geosynchronous orbit. Geosynchronous orbit is one particular orbit that’s really useful if you want to beam things straight down, or if you want to put a space elevator up. Right? So there’s this idea that if a civilization becomes advanced enough that it’s really using geosynchronous orbit, that you actually get a belt, something that would actually be detectable from a distance via a transit. There’s been a couple papers written about the possibility of these Clark Belts, densely occupied Clark Belts being a megastructure. It’s not as mega as a Dyson swarm, but it’s planetary scale.
Lex Fridman
(01:34:57)
You think it’s detectable, Clark Belt?
Adam Frank
(01:34:58)
It could be. In our list of technosignatures, it would be down there, but it would be… Again, if you had an advanced enough civilization that did enough of this, you’d have a Clark Belt. And the question is whether or not it’s detectable?
Lex Fridman
(01:35:11)
Yeah, probably Dyson sphere is the… That’s the more exciting thing too.
Adam Frank
(01:35:14)
That’s the go-to one. Yeah.
Lex Fridman
(01:35:16)
Speaking of the Dyson Sphere, let’s talk through the Kardashev scales.

Kardashev scale

Adam Frank
(01:35:19)
Right.
Lex Fridman
(01:35:19)
What is the Kardashev scale and where are humans on it?
Adam Frank
(01:35:24)
Right. So the Kardashev scale was at the same time. This is this golden age of SETI, like ’59 to ’65 when it just starts. Frank Drake has done his first experiment. People are like, “Oh my God, this is even possible.” And so people are just throwing out these ideas and as I said in the book, science is conservative. And what I mean by that is it holds onto its best ideas. So Kardashev comes up with this idea that, “Look, if we’re…” Again, it’s always about detectability. “If we’re looking for civilizations, we should think about what are the, “Natural,” stages,” natural in quotes, “That a civilization goes through?” And he was thinking in terms of energy use, like a good physicist. So he said, “Look, the first hurdle in terms of energy or threshold that a civilization will go through is using all the starlight that falls onto a planet.” He called that a type one civilization. In whatever way you’re doing it, some large fraction of the starlight that falls on your planet, you’re using for your own ends.

(01:36:24)
The next would be to use all the starlight there is from that star. Right? So that’s the Dyson sphere. So Dyson had already proposed his idea of the swarm and Kardashev was picking up. So that’s a type two civilization. Type three is galactic scale, a civilization that could use all the starlight in a galaxy. So where are we now? Remarkably, on a log scale. We’re at 0.7 of a type one.
Lex Fridman
(01:36:49)
So we’re not even type one?
Adam Frank
(01:36:50)
No, no, no. We’re not even type one, but according to… There was a paper written by a group that said, “Can we continue on our path? We’ll be at a type one at around 2300.”
Lex Fridman
(01:37:02)
2300. So this is on a log scale?
Adam Frank
(01:37:08)
Yeah.
Lex Fridman
(01:37:08)
So 0.7. So type one is about 10 to the 16th watts. Type two is 10 orders of magnitude larger than that, 10 to the 26th watts, and I think estimate for the galaxy is another 10 orders of magnitude.
Adam Frank
(01:37:20)
Yeah, because there’s a 100,000,000,000 star of order, 100,000,000 stars.
Lex Fridman
(01:37:24)
So that’s a lot.
Adam Frank
(01:37:25)
That’s a lot energy.
Lex Fridman
(01:37:27)
Do you think humans ever get to type one?
Adam Frank
(01:37:30)
I think that there’s a problem with type one, which is that we already know about climate change. The effects of our harvesting energy to do the work of civilization is already changing the climate state, and that’s something that Kardashev couldn’t have recognized. There’s the first law of thermodynamics, which is just about the different forms of energy. Then there’s the second law, which is about when you use that energy, Kardashev wasn’t thinking about the second law. If you get all that energy and you use it, there is waste heat. You don’t get to use it all. Right? Second law tells you that if I have a tank of gasoline, I can only use a certain fraction of the energy in that tank, and the rest is going to go to heating up the engine block. So that second law tells you that you can only use so much energy before the climate state is like, “Uh-oh, sorry, it’s going to change on you.”

(01:38:25)
So there’s a way in which we probably can’t get to a type one without devastating the earth’s climate. The most important thing actually here is probably, this is why space becomes… So the colonization or settlement of space. If we have an idea that we’ve been working on for a while called service worlds, that at some point you probably move a lot of your industry off world. We’ve got Mercury, for example. There’s nothing on Mercury, there’s no life on Mercury. Why don’t you put your energy harvesting there? Because you can’t mess with the biosphere. The biosphere is more powerful than you are. And so there’s limits to how much energy we can harvest to do work on the earth without really adversely affecting the biosphere.
Lex Fridman
(01:39:10)
It does seem that the best response to the climate change is not to use less technology, but to invent better technology and to invent technology that avoids the destructive effects.
Adam Frank
(01:39:25)
This is the frontier where you are, and that was the topic of my last book, Light of the Stars. It’s like you have to do the astrobiology of the Anthropocene. You have to see the transition that we’re going through now of the Anthropocene on a planetary astrobiological framework. And that paper we were talking about with a 10 billion trillion worlds, that was actually in service of the work I was doing for this other book where I wanted to know how often do you go through an… Does every technological civilization trigger its own planetary crisis, its own climate Anthropocene crisis? And the answer we actually came up from doing models was like, yeah, probably. And then the question is, are you smart enough to figure out how to readjust what you’re doing technologically so that all boats rise? You want to figure out how to do this so that the biosphere becomes even more productive and healthy and resilient.

(01:40:15)
So yeah, right. It’s the kind of technology. I think there’s probably absolutely limits on how much energy you can use, but how do you use that energy? And then also, getting off planet eventually. If you want to use 10 times more energy than that, you’re going to not going to do it on world.

Detecting aliens

Lex Fridman
(01:40:33)
So how do we detect alien type one, two, and three civilizations? So we’ve been kind of talking about basically type one civilization detection.
Adam Frank
(01:40:43)
Yeah. Right,
Lex Fridman
(01:40:44)
Maybe with the Dyson sphere, you start to get a little bit more type two, but it feels like if you have a type two civilization, it won’t be just the Dyson sphere.
Adam Frank
(01:40:54)
Right.
Lex Fridman
(01:40:55)
It feels like that. Just for the same reason you mentioned climate change, but now at the star system level, they’re probably expanding, right? So how would you detect a type two?
Adam Frank
(01:41:08)
How about propulsion plumes? Right? If you’re expanding… No, no.
Lex Fridman
(01:41:12)
Yeah, that’s great. That’s great.
Adam Frank
(01:41:12)
I literally just put in a NASA proposal now. Thomas Beatty, who’s joined us, he’s at the University of Wisconsin, has an idea to look for plumes. Right? If you have a solar system-wide civilization and you got space truckers going back and forth from Mars to… They’re doing the insettlest run, they’re accelerating and decelerating the whole way there. If you want to get to Mars in a couple of weeks, you have your fusion drive on the entire way out there. You flip and burn and have it on. So you also always have gravity. You have thrust gravity. So would those plumes be detectable? Because now you’ve got spaceships going all over the place and the odds that the plume is going to cross your field of view could become pretty high. So yeah, I think that’s one idea of looking for large-scale interplanetary, which is like when you’re getting to a type two.

(01:42:11)
Another possibility is looking for the tailings of asteroid mining. This was an idea, it was a group at Harvard Smithsonian, that to be able to look for… If you’re really chewing up asteroids to build space habitats, there’d be dust particles left around and would they look different from just say the dust from just regular collisions?
Lex Fridman
(01:42:30)
So pollution of all different kinds.
Adam Frank
(01:42:32)
Pollution of all different kinds
Lex Fridman
(01:42:33)
And trash also?
Adam Frank
(01:42:34)
Okay, so trash is an interesting idea when you come to the actual solar system. There’s a whole other field of technosignatures, which are things in the solar system. What if somebody came by 1,000,000 years ago and left some stuff? So the earth has been showing biosignatures for billions of years. A species like us, at our level, looking at earth, would’ve been able to know that earth had life on it, had a biosphere for billions of years. So maybe somebody sent something by a half a billion years ago. So this idea of looking say at the Moon for artifacts that have been there for a long time is something that a number of people are doing. We’re just working on a paper where we just calculated, this was super fun. We calculated how long would the lunar lander exist on the Moon before micrometeorites just chewed it down? How long would you be able to land on the Moon and go, “Oh, look, somebody was here and left some debris.”

(01:43:34)
So there’s this process called gardening, which is just the micrometeorite, constant rain of micrometeorites, and that’s where you get the lunar regolith. That fine powder on the Moon is because of this gardening. And it turns out it is literally hundreds of millions to billions of years-
Lex Fridman
(01:43:50)
Oh, nice.
Adam Frank
(01:43:50)
That the lunar lander will be visible.
Lex Fridman
(01:43:54)
Oh, so we should be able to find artifacts.
Adam Frank
(01:43:58)
If there are artifacts on there, and people have proposed doing this with artificial intelligence. The Moon has been mapped down to a couple of meters with various probes and all that data is sitting there. So why not use machine learning to look through all those things and look for anything that looks not like the lunar surface? And they did a test program where they gave the computer, I don’t know, 50 miles around the Apollo 11 or maybe it was Apollo 17 site, and it instantly was able to pull out the lander.
Lex Fridman
(01:44:27)
The whole task of looking for anomaly, something that looks not like the lunar surface. You make it sound obvious, but it’s not exactly obvious. Detect something that doesn’t look right about this room?
Adam Frank
(01:44:42)
Yeah.
Lex Fridman
(01:44:43)
It’s actually really difficult.
Adam Frank
(01:44:44)
Really difficult. It’s really difficult. And what’s cool, it’s a really information theoretic kind of proposal. You really have to use information theory to say, “What’s the background?” How do I define something that I can say, “That looks weird?”
Lex Fridman
(01:44:58)
Yeah, maybe when you’re looking at a spectrograph or something, it’s still like…
Lex Fridman
(01:45:00)
[inaudible 01:45:00] or something, it’s going to look really weird potentially. We’re hypothesizing all the things that humans would build and how do we detect that.
Adam Frank
(01:45:12)
Right.
Lex Fridman
(01:45:13)
But that could be really weird stuff.
Adam Frank
(01:45:15)
That’s why there’s this emphasis now on these agnostic signatures. So, actually disequilibrium is a nice one. One way to define life is it is a system that is far from equilibrium, it’s alive, because as soon as it dies, it goes back to equilibrium. And so, you can look at all chemicals in an atmosphere, even if you don’t know whether these could be chemicals that you have no idea whether or not they have anything to do with life. But the degree of disequilibrium, the degree to which they show that that atmosphere has not, the chemicals have not all just gone down to, they’ve all reacted away to an equilibrium state. You can actually tell that in very general ways using what’s called the Gibbs free energy, and that’s a signature.

(01:45:56)
If you see an atmosphere that is wildly out of equilibrium that indicates that there’s something happening on that planet biosphere or technosphere that is pumping gases into the atmosphere, that is keeping the whole system from relaxing.
Lex Fridman
(01:46:13)
So, is it possible we can detect anomalies in spacetime?
Adam Frank
(01:46:17)
Well, you could detect, and there’s been some work on this with the Alcubierre drive, these proposals for warp drives, and we can talk about that later, I’m skeptical of those. Because it may really be possible, you just can’t go faster than the speed of light. But people have done work on what would be the signature of an Alcubierre drive? What would be the signature? Could you detect if you’re using a drive like that, then you certainly are distorting spacetime, which means any light that’s passing by, its trajectory has gotten altered because it had to pass through the distorted spacetime.

(01:46:51)
So yeah, there are possibilities along with that. One of the funny things, I don’t know if they’ve gotten past this, but somebody calculated the problem with the Alcubierre drive or this warp drive was that if you dropped out of warp, there would be this spray of gamma rays that would sterilize any planet in front of you. So, it’s like, “Well yeah, you probably don’t want to do that,” but that would be a great bios or techno signature, another planet obliterated.

Warp drives

Lex Fridman
(01:47:15)
So, you think it’s not possible to travel fast than the speed of light?
Adam Frank
(01:47:17)
I wouldn’t say that. I wouldn’t say that, but what I think, if you look at the physics, we understand, every possibility for faster than light travel really relies on something that doesn’t exist. So, the cool thing is Einstein’s field equations, you can actually play with them, the equations are right there. You can add things to the right or left-hand side that allow you to get something like the Alcubierre drive. That was a metric that showed you like, “Oh, it’s a warped bubble.” It’s a warping of spacetime that moves through spacetime faster than the speed of light.

(01:47:52)
Because nothing can move across space faster than the speed of light, but spacetime itself can move faster than the speed of light. But here’s the problem with all of those proposals is they all need something. The thing you added, the little fictional term you added into the equations is something called exotic matter and it doesn’t exist. It’s really just something we dreamed up to make the equation to do what we wanted them to do. So, it’s a nice fiction but really right now, we live in this weird moment in history of the great acceleration where the technology we used now is completely different from the technology we used 10 years ago is remarkably different from the technology from 100 years ago.

(01:48:38)
But I remember playing Assassin’s Creed where everybody’s like, “What is it, it’s 1200?” And everybody’s like, “Stab, stab, stab.” And I was like, “Yeah, it’s a great game.” And then I got Assassin’s Creed II and it was 300 years later and everybody’s like, “Stab, stab, stab.” And it was like 300 years and the technology hadn’t changed and that was actually true for most of human history. You used your great-grandfather’s tools because there was no need to have any other new tools and you probably did his job. So, we could be fooled into thinking like, “Oh, technology’s going to go on forever, we’re always going to find new advances.”

(01:49:14)
As opposed to sometimes things just flatten out for a long time. So, you have to be careful about that bias that we have living in this time of great acceleration.
Lex Fridman
(01:49:23)
Yeah. But also, it is a great acceleration and we also are not good at predicting what that entails if it does keep accelerating. So, for example, somebody like Eric Weinstein often talks about we underinvest in theoretical physics research. Basically, we’re trying too hard for traditional chemical propulsion on rockets versus trying to hack physics, warp drives and so on.
Adam Frank
(01:49:53)
Yeah.
Lex Fridman
(01:49:54)
Because it’s really hard to do space travel, and it seems like in the long arc of human history, if we survive the way to really travel across long distances is going to be some new totally new thing.
Adam Frank
(01:50:07)
Right.
Lex Fridman
(01:50:07)
So, it’s not going to be an engineering problem, it’s going to be a physics problem-
Adam Frank
(01:50:12)
A fundamental physics problem.
Lex Fridman
(01:50:14)
Fundamental physics problem.
Adam Frank
(01:50:15)
Yeah. I agree with that in principle, but I think there’s a lot of ideas out there. String theory, people have been playing with string theory now for 40 years, it’s not like there hasn’t been a lot of effort. And again, I’m not going to predict, I think it’s entirely possible that there’s incredible boundaries of physics that have yet to be poked through, in which case then all bets are off. Once you get fast interstellar travel, whoa, who knows what can happen? But I tend to be drawn to science fiction stories that take the speed of light seriously. What kind of civilization can you build where it takes 50 years to get to where you’re going and a 50 years back?

(01:50:59)
So, I don’t know. Yeah, there’s no way I’m going to say that we won’t get warp drives. But as of right now, it’s all fictional. It’s barely even a coherent concept.
Lex Fridman
(01:51:08)
Well, it’s also a really exciting possibility of hacking this whole thing by extending human lifespan or extending our notion of time and maybe as dark as to say, but the value of an individual human life versus the value of life from the perspective of generations.
Adam Frank
(01:51:27)
Yeah.
Lex Fridman
(01:51:27)
So, you can have something like a generational ship that travels for hundreds of thousands of years and you’re not sad that you’ll never see the destination because you have the value for the prolonged survival of humanity versus your own individual life.
Adam Frank
(01:51:45)
Yeah. It’s a wild ethical question, isn’t it? That book I told you about Aurora, I love the book because it was such a inversion of the usual. Because I love science fiction, I’ve read so many generation ship stories. And they get to that planet, the planet turns out to be uninhabitable. It’s inhabited, but it’s uninhabitable for Earth because again, he has this idea of life is particular to their planets. So, they turn around and they come back, and then when they land, the main character goes, there’s still people who are arguing for more generation ships, and she goes, and she punches the guy out because she spent her whole life in a tube with this.

(01:52:20)
I thought that was a really interesting inversion. The interesting thing about, we were talking about these space habitats.
Lex Fridman
(01:52:26)
Yes.
Adam Frank
(01:52:26)
But if you really had a space habitat, not some super cramped, crappy, usual version of a century ship. But if you had these space habitats that were really like the O’Neill cylinders, they’re actually pretty nice places to live, put a thruster on those. Why keep them in the solar system? Maybe space is full of these traveling space habitats that are in some sense, they’re worlds in and of themselves.
Lex Fridman
(01:52:49)
There’s the show Silo, which raises the question of basically, if you are putting on a generational ship, what do you tell the inhabitants of that ship? You might want to lie to them.
Adam Frank
(01:53:00)
Yeah.
Lex Fridman
(01:53:01)
You might want to tell them a story that they believe.
Adam Frank
(01:53:04)
Right.
Lex Fridman
(01:53:04)
Because there is a society, there’s human nature. It’s like how do you maintain a homeostasis of that little society? That’s a fascinating technical question, the social question, the psychology question

Cryogenics

Adam Frank
(01:53:17)
The generation ship too, which I talked about in the book, the idea of also you talked about the extending human lifetimes or the stasis, the cryostasis, which is a mainstay of science fiction that you can basically put in suspended animation and such. None of these things we know are possible. But what’s so interesting, and this is why I love science fiction, the way it seeds ideas, all these ideas we’re going to talk about because they’ve been staples of science fiction for 50 years.
Lex Fridman
(01:53:44)
The whole field of cryogenics.
Adam Frank
(01:53:45)
Yeah. Where are we at with that?
Lex Fridman
(01:53:47)
Yeah. I wonder what the state of the art is for complex organism. Can you freeze? How long can you freeze? And then unfreeze maybe with bacteria you could do freeze.
Adam Frank
(01:53:56)
Oh, bacteria can last. This is the thing about panspermia, how long can a bacteria survive in a rock that’s been blasted? If there’s a comet impact across interstellar distances, that does seem to actually be possible. People have done those kinds of calculations, it’s not out of the realm of possibility. But a complex organism or multi-systems, with organs and such.
Lex Fridman
(01:54:20)
Also, what makes an organism? Which part do you want to preserve? Because maybe for humans, it seems like what makes a personality? It feels like you want to preserve a set of memories. If I woke up in a different body with the same memories, I pretty much, I would feel like I would be the same person.
Adam Frank
(01:54:43)
Altered Carbon, that’s a great series. I think it’s on Netflix, that’s a really great series where that’s exactly the idea of sleeves. Everybody’s able to, you can re-sleeve in another body, and it raises exactly this question. It’s not the greatest cyberpunk, but it’s pretty good, it’s got some great action sequences too.
Lex Fridman
(01:55:01)
As we get better and better advancements in large language models that are able to be fine-tuned on you, it raises a question because to me, they’ve already passed the Turing test as we traditionally have defined it. So, if there’s going to be an LLM that’s able to copy you in terms of language extremely well, it’s going to raise ethical and I don’t know, philosophical questions about what makes you, you. If there’s a thing that can talk exactly like you, what is the thing that makes you? It’s going to speak about your memories very effectively.
Adam Frank
(01:55:41)
This leads us to, if we’re going to get to the blind spot. I am of the opinion, heretical in some camps that the brain is not the minimal structure for consciousness, it’s the whole body. It’s embodied and may actually, in some sense, it’s communities actually. So yeah, I could be wrong, but this is what this whole work that I did with Marcelo Gleiser and Evan Thompson, the philosophy of science. Which is interesting, because it leads to this question about, “Oh, maybe we should just download ourselves into computers.” That’s another story that one tells. I’m super skeptical about those, but that’s one of the narratives about interstellar travel.

(01:56:20)
And that anybody we meet is going to be a machine anyway, whether it’s downloaded bodies or it’s just going to be artificial intelligence. There’s the whole idea of how long does biological evolution last? Maybe it’s a very short period before everybody goes to, or the machines take over and kill you, or it’s some hybrid.

What aliens look like

Lex Fridman
(01:56:39)
What do you think aliens look like? So, we talked about all the different kinds of bio signatures that might leave or techno signatures, but what would they look like when we show up? Are they going to have arms and legs? Are they going to be recognizable at all? Are they going to be carbon-based?
Adam Frank
(01:56:57)
Yeah. So, great question, and this question gets to the heart of thinking about life, about what life is. And this is the physical part of that, there’s also the informational part of it. But let’s just talk about the physical part of it, which is anything that we’re going to call life is probably going to work on Darwinian evolution. That’s the nice thing about Darwinian evolution, just like we know the laws of physics are general, the laws of Darwinian evolution are this logic, this basic logic that anything we’d reasonably call life probably has to operate under these kinds of principles.

(01:57:32)
And so, evolution’s about solving problems to survive that the environment presents. And the environment’s always going to present these problems in physical and chemical terms, so that you’d expect a balance between what we call convergence, evolutionary convergence and evolutionary contingency. So, if you’ve got to move along a surface, a hard surface and air, then the idea of some kind of jointed stick legs makes sense that you’re probably going to trigger that. If you look at Earth’s history multiple times, multiple lineages that had nothing to do with each other are going to solve the problem of getting towards energy sources using some kind of stick-like apparatus.
Lex Fridman
(01:58:18)
So, that’s about movement?
Adam Frank
(01:58:19)
Yeah. So, that’s one problem that has to be solved. The one problem that has to be solved is I got to get to food, right?
Lex Fridman
(01:58:22)
Yeah.
Adam Frank
(01:58:22)
Another problem is they got to get away from predators. You’ve seen wings, we’ve seen wings, the line that went through dinosaurs to birds involved wings, insects evolved wings, mammals evolved wings. If the gas is dense enough that a curved surface, if you move through the curved surface, it’s going to produce lift. Yeah, there you go, evolutionary trip on that. So, I think you can expect certain classes of solutions to the basic problems that life is going to be presented with stay alive, reproduce. But one of the weird things about with the UFO things is that you always see like, “Oh, they all look like humans, they’re just basically humans with triangular heads.” And that’s where we get to contingency.

(01:59:06)
So, what we’ve been talking about is convergence. You expect that evolution will converge on wings multiple times when presented with the problems that wings can solve. But contingency is accidents that you’ve got something that’s evolving a certain kind of wing, a leathery wing. And then the climate changes and they all die out, end of story or an asteroid, total accident, asteroid hits. And so, contingency accidents play also a huge role in evolution. And one of the things that lots of evolutionary biologists have talked about is the idea that if you ran the tape of Earth’s history over again, would you get the same creatures? Now, Stephen Jay Gould was of the opinion that no way, you wouldn’t find anything on Earth that resembled any species today.

(01:59:52)
They’ve done experiments actually on this with E. coli. You take a bunch of E. coli, you let them evolve for a while, you take a bunch of them out, freeze them, let one, let that population continue to evolve, the other one’s frozen. Now, started over again with the frozen. And it seems to be that contingency tends to win. At least from what we can tell, that’s not a hard result, but in those experiments, what you find is that accidents really do matter. And this is important, so yes, you should expect legs or jointed sticks, how many joints they’re going to be? Anybody’s guess.

(02:00:27)
Do you expect humanoids, things with a sensing apparatus on top of a shoulder with two arms and two legs? That’s probably a pretty random set of occurrences that led to that.
Lex Fridman
(02:00:39)
I guess what is a brain versus the nervous system? Where is most of the cognition competition going on?
Adam Frank
(02:00:46)
Yeah.
Lex Fridman
(02:00:47)
You could see that in organisms. Actually, I don’t know how the brain evolved. Why does it have to be in one place?
Adam Frank
(02:00:56)
It doesn’t have to be. So, my favorite word, word of the day is liquid brains. This idea of distributed cognition, which fascinating idea, and we’ve come to understand how much distributed cognition there is. Obviously, you social animals like termites, and ants, that’s an example of distributed cognition, the organism is the whole colony. This is one thing that’s been really interesting in the state of the study for aliens, is that when we’ve come to recognize that human intelligence, the kinds of things that go into intelligence are distributed all across the biosphere. Lots of different examples of things show various pieces of what we have. Jason Wright described it as a deck of cards. The cards are all there, we got the hand that actually led to the technological progress that we see. But the basic idea of using tools, the basic idea of recognizing each other eye to eye, all the things that we define as intelligence. You can find many places in many other places across many other lineages across the earth. So, they could be very, very different with something like, yeah, maybe the hive mind idea or bacterial colonies that actually managed to come to their own version of high cognition.
Lex Fridman
(02:02:10)
Well, I wonder if we stretch out time across 10s, 20 billion years, whether there’s an Darwinian evolution stops working at some point in terms of the biology or the chemistry of the organisms, and it switches to ideas for example. It’s much more rapidly you’re operating maybe, I guess it’s a kind of Darwinian evolution on the space of memes or whatever, as [inaudible 02:02:36]-
Adam Frank
(02:02:35)
Technology seems to operate, but certainly markets can operate in ways that look very Darwinian.
Lex Fridman
(02:02:43)
So, basically a planet is working hard to get to the first kind of organism that’s able to be a nice platform for ideas to compete.
Adam Frank
(02:02:53)
Yeah.
Lex Fridman
(02:02:53)
And then it stops evolving there, and then these ideas that take off.
Adam Frank
(02:02:57)
Right. Because yeah, cultural Lex it’s true. It’s amazing that cultural evolution totally disconnects from the Darwinian process. But I’d be careful to say that a planet is working hard to do this. Because really looking at us, what we think of as ideas and culture, and it’s quite possible we’re going to make it another 200 years, and this is gone because it actually wasn’t a very good idea long-term, we just don’t know.
Lex Fridman
(02:03:22)
So, maybe the idea generation organism is actually the thing that destroys.
Adam Frank
(02:03:26)
Not the biosphere, because again, but it destroys itself. It may not be very long- term, it may be very potent for a short period of time but that it’s not sustainable. It doesn’t become, like we were talking about before, mature. It’s very hard to make it into integrated into a mature bio/technosphere. And of course, evolution that is not working for anything. Well, here’s the actually interesting thing, so people are very much evolutionary biologists will get their hair will stand on end if you start talking about evolution, having a purpose or anything.

(02:03:53)
But the very interesting thing about purpose is that once you do get to a idea generating species or collective organism, yeah, then all bets are off and there is goals, there is teleology. Now suddenly, absolutely, there’s a direction implied. So that’s a cool interesting thing that once you get to that, evolution stops being goalless and directionless and suddenly, yeah, we’re the ones who supply or any kind of creature like us has an absolute direction that way they decide on.
Lex Fridman
(02:04:26)
Although you could argue that from a perspective of the entire human civilization, we’re also directionless. We have a sense that there’s a direction in this cluster of humans.
Adam Frank
(02:04:37)
Yeah.
Lex Fridman
(02:04:37)
And then there’s another cluster has a different sense of direction, there’s all kinds of religions that are competing. There’s different ideologies that are competing.
Adam Frank
(02:04:45)
Yeah.
Lex Fridman
(02:04:45)
And when you just zoom out across, if we survive across thousands of years, it will seem directionless. It will seem like a pinball.
Adam Frank
(02:04:55)
It’s an unholy mess. But at some point, the expansion into the solar system say, that would be both direction. Depending on how you look at it, it was directional. There was a decision that the collective of human beings made to like anti-accrete, to start spreading out into the solar system. So, that was definitely a goal there that may have been reached in some crazy nonlinear way, but it still a goal was set and it was achieved.

Alien contact

Lex Fridman
(02:05:25)
If there’s advanced civilizations out there, what do you think is the proper protocol for interacting with them? Do you think they would be peaceful? Do you think they would be warlike? What do we do next? We detect the civilizations through all the technosignatures we’ve been talking about, maybe direct imaging, maybe there’s really strong signal. We come up with a strategy of how to actually get there.
Adam Frank
(02:05:49)
Yeah.
Lex Fridman
(02:05:50)
But then the general says, they always do, the military industrial complex-
Adam Frank
(02:05:56)
We’ve watched that movie.
Lex Fridman
(02:05:58)
What kind of rockets and do we bring rockets?
Adam Frank
(02:06:02)
Right. Well, this general question also leads to many messaging, extraterrestrial intelligence, and I’m definitely of the opinion of you should be very careful. I don’t think it’s necessarily a bad idea to have your head below the grass. The people who advocate like, “Oh yeah, we should be sending powerful messages that are easily detectable into interstellar space.” I’m like, “Why would you?” Because we just don’t know, I’m not going to say they are warlike. I’m not going to say they’re not warlike, I have no idea. But we sure as hell, well, first of all, who gets to decide that? The idea that a bunch of astronomers who happen to have a radio telescope, Who Speaks for Earth, which I think was a great book somebody wrote.

(02:06:44)
So, definitely we should be cautious, I would say, because we just have zero information. And the idea, you used to have this idea of, well, if they’re advanced, they’ve managed to survive. So of course, they’re going to be wearing togas and be singing kumbaya, but I just wouldn’t assume that. It’s also possible though that their cognitive structure is so different that we’re not even living in the same universe in a certain way. I think we have to be prepared for that. We may not even be able to recognize each other in some way as cognizing beings. One of my favorite movies is Arrival, I don’t know if you’ve ever seen that one.

(02:07:18)
I really love that one because they literally, they have a different language. They have a different cognitive structure in terms of their language, and they’re literally living in a different physics.
Lex Fridman
(02:07:25)
Different physics, different language, different everything. But in the case of Arrival, it can at least recognize that they’re there.
Adam Frank
(02:07:34)
And they managed to cross the language barrier. Yeah.
Lex Fridman
(02:07:38)
But that’s, both sides have an interest in communicating, which you suppose that an advanced civilization would have a curiosity. Because how do you become advanced without curiosity about the mysteries about the other.
Adam Frank
(02:07:54)
But also, if they’re long-lived, they may just be like, “We’re not even interested. Say 10 million years ago, we were really interested in this, in communicating with you youngins, but now we’re not at all.” And that’s just one of the beauties of this again, is how to think about this systematically because you’re so far past the hairy edge of our experience of what we know that you want to think about it. You don’t want to be like, “Don’t know, can’t say anything,” because that’s not fun. But you also have to systematically go after your own biases. So, one of the things I loved about Arrival too was Carl Sagan always had this idea, “We’ll teach them math, we’ll teach them our math, then they’ll teach us their math, and then we’ll be telling each other, knock-knock jokes and swapping cures for cancer.”

(02:08:42)
And in the movie, they send a Carl Sagan guy in and a linguist, and the Carl Sagan guy fails immediately. And it’s the linguist who understands that language is actually embodied. Language is not just something that happens in your head, it’s actually the whole experience and she’s the one who breaks through. And it just points to the idea that how utterly different the cognitive structures of a different species should be. So somehow, we have to figure out how to think about it, but be so careful of our biases or figure out a systematic way to break through our biases and not just make science fiction movies. You know what I mean?
Lex Fridman
(02:09:17)
Yeah. Speaking of biases, do you think aliens have visited Earth? You’ve mentioned that they could have visited and started civilizations and we wouldn’t even know about it if it was 100 million years ago. How can we even begin to answer this question, whether-
Adam Frank
(02:09:32)
Got to look, got to figure out ways to look. So, it’s not high on my list of things that I think are probable, but it certainly, it needs to be explored. And unless you look, you never know. So, looking on the moon, where would we find if aliens had passed through the solar system anytime in the last 3 billion years, where might we find artifacts? Where might artifacts still be around? Earth? Probably not because of weathering and resurfacing. The moon’s a good place. Certain kinds of orbits, maybe they parked a probe in an orbit that was stable. So, you got to figure out which orbits actually you could put something there and it’ll last for a billion years.

(02:10:10)
So, those are the kind of questions. Like I said, it’s not high on my list of thinking this could happen, but it could happen. Unless you look, you don’t know.
Lex Fridman
(02:10:20)
Speaking of biases, what about if aliens visiting Earth is the elephant in the room? Meaning the potential of aliens, say seeding life on earth?
Adam Frank
(02:10:30)
You mean in that directed panspermia, [inaudible 02:10:33]-
Lex Fridman
(02:10:32)
Directed panspermia.
Adam Frank
(02:10:33)
Yeah.
Lex Fridman
(02:10:34)
Or seeding some aspect of the evolution.
Adam Frank
(02:10:39)
Like 2001.
Lex Fridman
(02:10:40)
Yeah.
Adam Frank
(02:10:41)
Yeah. It’s a great story, but always with Occam’s razor or whatever with science. If I can answer that question without that extra very detailed hypothesis, then I should. And the idea that evolution is a natural process, that’s what I would go for first. That just seems it’s so much easier to do it that way than adding, because it’s kind of a duo sex machina thing of like, “Oh, then the aliens came down and they solved that problem that you’re trying to solve by just coming down and putting their finger on the scales.”
Lex Fridman
(02:11:13)
So, to you, the origin of life is a pretty simple thing that doesn’t require an alien?
Adam Frank
(02:11:19)
I wouldn’t say that, it’s not a simple thing. Because all you’re doing is kicking the can down the road. The aliens formed, right? So, you’re just saying like, ” All right, I’m just kicking the can down the road to the aliens. What was their abiogenesis event?
Lex Fridman
(02:11:35)
Well, so from a different perspective, I’m just saying, it seems to me that there’s obviously advanced civilizations everywhere throughout the galaxy and through the universe from the Drake equation perspective. And then if I was an alien, what would I do? I’ve gotten a chance to learn about the uncontacted tribes in the Amazon. I recently went to the Amazon, and you get to understand how they function and how the humans in the Amazon, they’re in contact with the civilized world, how they interact with the uncontacted tribes. First of all, the uncontacted tribes are very violent towards the outside world, but everybody else tried to stay away from them. They try to protect them, don’t talk about them, don’t talk about their location and all this kind of stuff.

(02:12:19)
And I’ve begun to internalize and understand that perspective of why you’re doing that. And if I was an alien civilization, I probably would be doing a similar kind of thing. And of course, there’s always the teenager or the troll who’s going to start messing with this stuff or the scientists.
Adam Frank
(02:12:34)
Yeah, right.
Lex Fridman
(02:12:35)
And so, from our perspective, yes. And if you’re in the Truman Show like Occam’s razor, but also the Occam’s razor from the perspective of the alien civilization, we have to have the humility to understand that that interaction will be extremely difficult to detect, that it would not be obvious.
Adam Frank
(02:12:57)
Right. I understand the logic of what you’re saying, but the problem for me with that is that first you have to assume that alien civilizations are common, which I’m not sure about it, that most of them may be dead or they’re not. While I think that life is common, and again, this is just my biases. So now, the problem is how do we sort out the biases we’re bringing or the assumptions we’re bringing in from the causal chain that comes out of that? I would first want to try and do this without, if we’re looking at the origin of life or the evolution of life on Earth. I’d want to do it just on its own without asking for this other layer because it requires a bunch of these other assumptions which also have their own breaking of causal chains.

(02:13:44)
Because the idea that when you ask, what would you do if you were an alien? But again, alien minds could be so unbelievably different that they wouldn’t even recognize the question you just posed.
Lex Fridman
(02:13:56)
Right.
Adam Frank
(02:13:56)
Because it’s just like we have a very particular cognitive structure or cognitive, and we’re very governed by, even if you went and talked to, this is an interesting thing to think about. If I could suddenly magically appear 100,000 years ago and talked to a hunter-gatherer about their worldview and their motivations, I might find something that, or no resemblance to things that I think are sort of, “Oh, that’s what naturally humans do.”
Lex Fridman
(02:14:20)
Well, let me ask you this question. Let’s together do the thought experience.
Adam Frank
(02:14:23)
Yeah.
Lex Fridman
(02:14:23)
If we either create a time machine that allows us to travel back and to talk to them.
Adam Frank
(02:14:28)
Yeah.
Lex Fridman
(02:14:28)
Or we discover maybe a primitive alien civilization on a nearby star system, what would we do?
Adam Frank
(02:14:37)
Yeah. I think that’s a great question. It’s interesting how that even brings up the ethical questions. Let’s say that we’d have to first sort out what are the consequences for them and what do we feel our ethical responsibilities are to them?
Lex Fridman
(02:14:51)
And also, sorry, from a capitalist perspective, what are we to gain from this interaction?
Adam Frank
(02:14:56)
Right. You look at the way the missionaries, missionaries had these interactions because they thought converting them to whatever religion they were was the most important, that’s what the gain was. So, from our perspective, we’d have to sort that out. I think given if we’re doing this thought experiment, we are curious, and I think eventually we’d want to reach out to them.
Lex Fridman
(02:15:19)
I think when you say we, let’s start with the people in this room, right?
Adam Frank
(02:15:23)
Yeah.
Lex Fridman
(02:15:25)
I wonder who the dominant forces are in the world, because I think there’s a lot of people, the military they’ll probably move first so they can steal whatever advantage they can from this new discovery so they can hurt China or China hurt America. That’s one perspective. Then there’s the capitalist school will see how the benefits and the costs here, and how can I make money off of this? There’s opportunity here, there’s gold in them hills. And I wonder, and I think the scientist is just not going to, unlike the movies-
Adam Frank
(02:16:00)
We’re not going to get much say.
Lex Fridman
(02:16:00)
They’re going to put them-
Adam Frank
(02:16:01)
“Hey guys, wait a minute.”
Lex Fridman
(02:16:03)
They would engage probably. Just as a human society as we are now, we would engage and we would be detectable, I think.
Adam Frank
(02:16:12)
In our engagement.
Lex Fridman
(02:16:13)
In our engagement.
Adam Frank
(02:16:14)
Yeah, probably.
Lex Fridman
(02:16:15)
So, using that trivial bias logic, it just feels like aliens would need to be engaging in a very obvious way. Just brings up that old direct for me paradox for me. What do you make of all the UFO sightings?

UFO sightings

Adam Frank
(02:16:32)
I am all in favor of an open, agnostic, transparent, scientific investigation of UFOs and UAPs. But the idea that there’s any data that we have that links UFOs and UAPs to non-human technology, I just think the standards, none of what is claimed to be the data lives up to the standards of evidence. So, let’s just take a moment on that idea of standards of evidence, because I made a big deal about this both in the book and elsewhere whenever I talk about this. So, what people have to understand about science is we are really, our scientists, we are really mean to each other, we are brutal to each other.

(02:17:10)
Because we have this thing that we call standards of evidence, and it’s the idea of you have a piece of evidence that you want to link to a claim. And under what conditions can you say, “Oh, look, I’ve got evidence of this claim X, Y, and Z.” And in science, we are so mean to each other about whether or not that piece of evidence lives up to the standards that we have. And we spent 400 years determining what those standards are, and that is why cell phones work. If you didn’t have super rigorous standards about what you think that’s, “Oh, this little antenna, I’ve invented a new kind of antenna that I can slip into the cell phone and I can show you that it works.”

(02:17:50)
If you didn’t have these standards, every cell phone would be a brick. And when it comes to UFOs and UAPs, the evidence you have and the claim that though this shows that we are being visited by non-human, advanced civilization just doesn’t even come close to the same standards. I’m going to have to obey or whatever live under. If my team, the group I work with is one of them says, “Look, we’ve discovered and he wants to announce that, oh, we’ve discovered a technosignature on an alien planet.” We’re going to get shredded as we expect to be, we expect to be beaten up. And the UAP, UFO community should expect the same thing. You don’t get a pass because it’s a really cool topic.

(02:18:32)
So, that’s where I am right now. I just don’t think any of the evidence is even close to anything that could support that claim.
Lex Fridman
(02:18:39)
Well, I generally assign a lot of value to anecdotal evidence from pilots. Not scientific value, but just like it’s always nice to get anecdotal evidence as a first step. Because I was like, “I wonder if there’s something there.” But unfortunately, with this topic, there’s so much excitement around that there’s a lot of people that are basically trying to make money off of it. There’s hoaxes all this kind of stuff. So, even if there’s some signal, there’s just so much noise it’s very difficult to operate with. So, how do we get better signal? So, you’ve talked about if we wanted to really search for UFOs on Earth and maybe detect things like weird physics, what kind of instruments would we be using?
Adam Frank
(02:19:23)
Yeah, so in the book, I talked about the idea that this is really stupid, but you want to look up, you want to look down and you want to look all around.
Lex Fridman
(02:19:31)
I think that’s brilliant. It’s simple, not stupid. It’s like literally.
Adam Frank
(02:19:35)
Yeah, right. So, you want to do ground-based detectors, upward-looking, ground-based detectors of the kind we’re already building for meteors, for tracking meteors. You want to have space-based detectors, put them on satellites, this is what the NASA UAP panel was thinking about. And then probably on, we have lots of people in the sky there should be detectors on the planes, or at least some kind of alert system that if a pilot says, “Oh, look, I’m seeing something I don’t understand.” Boop presses the red button, and that triggers the ground.
Adam Frank
(02:20:03)
I’m seeing something I don’t understand. Boop. Presses the red button and that triggers the ground-based and space-based data collectors. And then the data collectors themselves, this is something that people really don’t understand and it’s so important. In order to actually do science with anything, the data you have, you have to understand where it came from down to the nth degree. You have to know how that camera behaves in a bunch of different wavelengths. You have to have characterized that. You have to know what the software does, what the limits of the software are possible. You have to know what happened to the camera. Was it refurbished recently? In every spectral wavelength in all of its data collection and processing, you have to know all of those steps and have them all characterized because especially if you want to claim like, “Oh my God, I saw something, take a right-hand turn at Mach-500.” Right?

(02:20:51)
You better have all of that nailed down before you make that kind of claim. So we have to have characterized detectors looking up, down, and maybe on planes themselves, we need a rational search strategy. So let’s say you want to lay out these ground-based detectors. Where do you put them? Right? There’s only so much money in the world, so do you want to put them near places where you’ve seen a lot of things beforehand or do you want to have them try and do sparse coverage of the entire country?

(02:21:17)
And then you need the data analysts analysis, right? You’re going to have so much data, so many false positives or false triggering that you need a way of sorting through enormous amounts of data and figuring out what you’re going to throw out and what you’re going to keep, and all of these things we’re used to doing in other scientific enterprises. And without that, if we don’t do that, we’re going to be having the same damn argument about these things for the next 100 years.
Lex Fridman
(02:21:40)
But if I asked you, I give you $1 trillion and asked you to allocate to one place looking out, SETI or looking at Earth, should you allocate it?
Adam Frank
(02:21:52)
Oh God, looking out. Looking out. Because that’s the, as I always like to say, here’s my codification of this. If you said, “Hey, Adam, I’d like to find some Nebraskans.” And I said, “Oh, good, let’s go to the Himalayas.” You’d be like, “Why am I going there?” I’m like, “Well, maybe there’s some Himalayas, some Nebraskans in Himalayas.” You’d say, “No, no. Let’s go to Nebraska.” If we’re looking for aliens, why don’t we look on alien planets where they live? We have that technology now as opposed to the bucket of assumptions that you have to come up with in order to say like, “Oh, they’re here right now. They just happen to be here right now.” And also the very important thing, I called this the high beam argument to deal with the UFO stuff, you have to answer these weird, irrational things that are happening.

(02:22:36)
Like, okay, there’s an advanced civilization that is visiting Earth regularly. They don’t want to be detected. They’ve got super powerful technology, but they really suck at using it because we keep seeing them, we keep seeing them, but then they disappear. I mean, explain to me what rational world that works under. So there’s that whole sort of argument. You’ve got to explain why if they want to stay hidden, are they so bad at it? So that’s why I take that level of difficulty and then I put it on top of where should I look? I should look at where they’re from. That makes me want to look at do the telescopic stuff.
Lex Fridman
(02:23:17)
Yeah, I think the more likely explanation is either the sensors are not working correctly or it’s secret military technology being tested.
Adam Frank
(02:23:27)
Absolutely. I mean, listen, that’s why again, I think UAP, absolutely UAP should be studied scientifically, but if I had to make a bet and it’s just a bet, I would say this is pure state adversary stuff. When I did, I did a New York Times op-ed for this in 2021, which blew up, and so I had a lot of people talking to me. While I was doing that. I sort of looked at the signals intelligence people, the SIGINT and EINT, electronic intelligence communities, and what they were saying about the New York Times articles and the various videos, and really none of them were talking about UFOs. They were all talking about pure state. That’s why I learned the word pure state adversaries, how even simple drone technologies and you purposely want to do this. You want to fake signals into the electronics of their adversary, so they crank it up so then you can just soak up all the electromagnetic radiation and know exactly what those advanced radars can do.
Lex Fridman
(02:24:25)
That said, I’m not saying that’s what this is. If I was the head of an alien civilization and I chose not to minimize the amount of contact I’m doing, I would try to figure out what would these humans, what would these aliens like to see? That’s why the big heads in the humanoid form, I mean, that’s kind of how I would approach communication. If I was much more intelligent, I would observe them enough. It’s like, all right, if I wanted to communicate with an ant colony, I would observe it long enough to see what are the basic elements of communication. And maybe I would do a trivial thing, do a fake ant in there.
Adam Frank
(02:25:07)
Right. A robot ant.
Lex Fridman
(02:25:08)
A robot ant, but then it’s not enough to just do a robot ant. You have to do a robot ant that moves in the way they do, and maybe aliens are just shitty at doing the robot ants. But no, I just wanted to make the case for that,
Adam Frank
(02:25:21)
This is the plot actually of a great science fiction book called Eon by Greg Baer, and the idea was these sort of, this is actually where my first, I became sort of more than agnostic, anti-medy, because the idea is that yes, our aliens come, they sort of make their arrival and really their point is to get rid of us. It’s the dark forest hypothesis. And what they do is they literally, the way they present themselves is in this sort of classic UFO thing, and they do it and they arrive at, this was during the Soviet Union. They arrive at the USSR, they arrive in China, and they’re kind of faking us out so that we never can organize ourselves against… So it was really, they did exactly what you’re talking about, but for nefarious purposes.
Lex Fridman
(02:26:03)
Okay, let me ask the pothead question. Yet another pothead-
Adam Frank
(02:26:07)
Another pothead. The whole conversation-
Lex Fridman
(02:26:09)
I’m sorry.
Adam Frank
(02:26:09)
Boggs before breakfast.
Lex Fridman
(02:26:11)
It’s signs and pothead questions back and forth. Okay, what if aliens take a form that’s unlike what we kind of traditionally envision in analyzing physical objects? What if they take the form of say ideas? What if real pothead, if it’s consciousness itself, like the subjective experience as an alien being, maybe ideas and is an easier one to visualize? Because we can think of ideas as entities traveling from human to human.
Adam Frank
(02:26:44)
I made the claim that the most important, that finding life any kind of life would be the most important discovery in human history. And one of the reasons is, again, as I said, that life, if we’re not an accident and there’s other life, then there’s probably lots of other life. And because the most significant thing about life is it can innovate, right? If I give you a star and tell you the mass and the composition, you can basically pretty much use the laws of physics, tell exactly what’s going to happen to that star over its entire lifetime. Maybe not the little tiny details, but overall it’s going to be a white dwarf, if it’s going to be a black hole end of story. If I gave you a single cell and said, “What’s going to happen in a few billion years?” You’d never be able to predict a giant rabbit that can punch you in the face, right?
Lex Fridman
(02:27:24)
Yeah.
Adam Frank
(02:27:25)
A kangaroo.

(02:27:26)
So life has this possibility of innovating, of being creative. So what it means is, and that’s kind of a fundamental definition of what it means to be alive. It goes past itself. So give life enough time and what are the end result? That’s why I love science fiction so much. At some point, does life reach a point where it climbs into the laws of physics itself. It becomes the laws of physics or these sort of lie at the extreme limits of thinking about what we mean by reality, what we mean by experience. But I’m not sure there was much we can do with them scientifically, but they’re open-ended question about the open-ended nature of what it means to be alive and what life can do.

Physics of life

Lex Fridman
(02:28:14)
Since you said it’s the biggest question, which is an interesting thought experiment, what is the biggest scientific question we can possibly answer? Some people might say about what happened before the Big Bang, some big physics questions about the universe. I could see the argument for how many alien civilizations or if there’s other life out there? You want to speak to that a little bit? Why is it the biggest question in… Why is it number one in your top five?
Adam Frank
(02:28:43)
I’ve evolved in this, right? I started off as a theoretical physicist. I went into computational astrophysics, magnetohydrodynamics of star formation, but I always was a philosophy minor. I always had these sort of bigger questions sort of floating around the back of my mind. And what I’ve come to now is the most important question for physics is, what is life? What the hell is the difference between a rock and a cell, fundamentally? And what I really mean by this, this is where I’m going to go non-traditional, is that really the fundamental question that is agency. What does it mean to be an autonomous agent? How the hell does that happen? I’m not a reductionist. I’m not somebody who’s just like, well, you just put together enough chemicals and bing, bang, boom, and it suddenly appears there’s something that really is going to demand a reconception of what nature itself is.

(02:29:30)
And so yeah, black holes are super cool. Cosmology is super cool. But really this question of what is life? Especially, from by viewing it from the inside, because it’s really about the verb to be. Really what is the most impressing philosophical question beyond science? Is the verb to be, what is being? This is what Stephen Hawking said when he talked about, “What puts the fire in the equations? The fire.” The fire is this presence and this is where it touches things like whatever you want to say it, the sacred, spirituality, whatever you want to talk about. My first book was about science and human spirituality. So this question of life, what makes life as a physical system so different is to me much more because that is where being appears. Being doesn’t appear out there. The only place that ever appears to any of us is us. I can do this kind of projection into this third person thing, but nobody ever has that, that God’s eye view. That’s a story we tell. This is where, this between us is where the verb to be, appears.
Lex Fridman
(02:30:36)
So this is something that you write about in The Blind Spot, why science cannot ignore human experience, sort of trying to pull the fire into the process of science. And it’s a kind of critique of materialism. Can you explain the main thesis of this book?
Adam Frank
(02:30:56)
Yeah. So the idea of The Blind Spot is that there is this thing that is central to science. So we’re using the blind spot as a metaphor. So the eye has an optic nerve, and the optic nerve is what allows vision to happen. So you can’t have vision without the optic nerve, but actually you’re blind to the optic nerve. There’s a little hole in your vision where the optic nerve is. And what we’re saying is that science has something like this. There’s something that without which science would not be possible, but that science, the way it’s been configured, and actually, when we mean the blind spot, I’ll get into exactly what I mean what it is, but it’s not really science. It is a set of ideas that got glued onto science. It’s a metaphysics that got glued on science. And so what is that thing? What is the blind spot? It’s experience. It is presence. And if I experience, people have to be very careful. I’m not talking about being an observer. There’s lots of words for it. There’s direct experience. There is presence. Being. The life world. Within the philosophy called phenomenology. There’s the life world.

(02:32:00)
It’s this sort of raw presence that you can’t get away from until you die. And then who the hell knows that as long as you’re around, it’s there. And what we’re saying is that, that is the way to say this, that is the precondition for the possibility of science and the whole nature of science, the way it has evolved is that it purposely pushed that out. It pushed that out. So it could make progress, and that’s fine for a certain class of problems. But when we try to answer, when we try and go deeper, there’s a whole other class of problems. The nature of consciousness, the nature of time, quantum mechanics, that comes back to bite us. And that if we don’t learn how to take, understand that, that is always the background, that experience is always the background. Then we just end up with these paradoxes and that require this intellectual yoga to get out of.
Lex Fridman
(02:32:54)
I think you give a bunch of examples of that. Looking at temperature as a number is a very objective, scientific way of looking at that. And then there’s the experience of the temperature.
Adam Frank
(02:33:02)
And how you build the parable of temperature that we call it. So what is the blind spot? We use the term it’s a constellation. It’s not just materialism. It’s a constellation of ideas that are all really sort of philosophical views. They’re not what science says, but because of the evolution of the history of science and culture got like pin the tail on the donkey, they were sort of pinned on and to tell us that this is what science says.

(02:33:25)
So what is it? One is reductionism that you are nothing but your nerve cells, which are nothing but the chemistry, which is nothing but all the way down to quarks. That’s it. So that’s reductionism.

(02:33:36)
The objective frame that science gives us this god’s eye view, this third-person view of the world to view the world from the outside. That’s what science bequeaths to us, that view.

(02:33:46)
Physicalism, that everything in the world is basically made of stuff. There’s nothing else to talk about that, that’s all there is. And everything can be reduced to that.

(02:33:55)
And then also the reification of mathematics, that mathematics is somehow more real than this.

(02:34:01)
And there’s a bunch of other things. But all these together, what they all do is they end up pushing experience out and saying experience is an epiphenomena. Consciousness. I tend not to use the word consciousness. I think it leads us in the wrong direction. We should focus on experience because it is a verb kind of in a way. It is verb-like and by being blind to that, we end up with these paradoxes and problems that really not only block science, but also have been detrimental to society as a whole, especially where we’re at right now.
Lex Fridman
(02:34:33)
So you actually say that, that from a perspective of detrimental society, that there’s a crisis of meaning, and then we respond to that in a way that’s counterproductive to these bigger questions, scientific questions. So the three ways responses you mentioned is scientific triumphalism, and then on the other side is rejecting science completely, both on the left and the right. I think the postmodernist on the left and the anti-establishment people on the right, and then just pseudoscience that kind of does this in-between thing. Can you just speak to those responses and to the crisis of meaning?
Adam Frank
(02:35:08)
Right, right. So the crisis of meaning is that on the one hand, science wants to tell us that we’re insignificant. We’re not important. We’re just biological machines. And so we’re basically an insignificant part of the universe. On the other hand, we also find ourselves being completely significant. In cosmology, we have to figure out how to look from the inside. At cosmology, we’re always the observers. We’re at the center of this collapsing wavefront of light. Quantum mechanics, it really comes in, it comes the measurement problem just puts us front and center. And we’ve spent 100… Some people spent 100 years trying to ignore the measurement part of the measurement problem. So on the one hand, we’re insignificant, and on the other hand, we’re central. So which one is it? And so this all comes from not understanding actually the foundational role of experience, this inability, we can’t do science without already being present in the world.

(02:36:03)
We can’t reduce what happens in science to some sort of formal… A lot of it is about we love our formal systems, our mathematics, and we’re substituting. That’s one of the things that, there’s two philosophers we really like for our heroes. One is Husserl, who is a mathematician, who invented phenomenology. And the other is Whitehead, who’s one of the greatest mathematicians of the 20th century. And Husserl came up with this idea of the surreptitious substitution. Part of The Blind Spot is substituting a formal system, a calculus of data for actual experience that that’s more important.

(02:36:39)
And so let me just do, before I go to those three responses, let’s just do the parable of temperature because I think people can… It’ll help them understand what we mean. So think about degree Celsius. We have in the modern scientific culture we live in, we think like, oh yeah, degree Celsius. They’re out there. The universe, the molecular cloud in space is 10 degrees Kelvin. The way we got there is we’ve forgotten how that idea is rooted in experience. We started off with science by, we had the subjective experience of hot and cold. I feel hot, I feel cold, you feel hot, you feel cold. Science was this process of trying to extract from those experiences what Michel Bitbol philosopher calls, “The structural invariance.” The things that we could both kind of agree on. So we figured out like, oh, we could make a gradiated little cylinder that’s got mercury in it and that hot things will be higher in on that gradiated cylinder, cold things will be lower, and we can both kind of figure out what we’re going to agree on are our standards for that. And then we have thermometry, yay. We have a way of having a structural invariant of this sort of very personal experience of hot or cold.

(02:37:53)
And then from that, we can come up with thermodynamics, etc. And then we end up at the bottom of that with this idea of everyday I wake up and I check my phone and I’m like, oh, it’s going to be 60 degrees out. Great. And we start thinking that 60 degrees is more real than hot and cold. That thermodynamics, the whole formal structure of thermodynamics is more real than the basic experience of hot and cold that it came from. It required that bodily experience that also, not just me, I have to tell you, it’s part of my communication with you, cold today, isn’t it? That from that basic irreducible experience of being in the world with everything that it involves, I developed degree Celsius, but then I forgot about it. I forgot the experience. So that’s called the amnesia of experience.

(02:38:41)
So that’s what we mean by how the blind spot emerges, how we end up, how science purposely pushes experience out of the way so it can make progress, but then it forgets that experience was important. So where does this show up? Why is this? What are the responses to trying to get this back in and where this crisis of meaning emerge? So scientific triumphalism is the idea that the only thing that’s true for us are scientific truths. Unless it can be codified in a formal system and represented as data, captured in some kind of scientific causal network, it doesn’t even exist. And anything else that’s not part of it that can be formalized in that way is an epiphenomenon. It’s not real.

(02:39:25)
So scientific triumphalism is this response to the weirdness of, I could call it the mystery, the weirdness of experience by just ignoring it completely. So there’s no other truth. Art, music, human spirituality, it’s all actually reducible it neural correlates. So that’s one way that it’s been dealt with.

(02:39:47)
The other way is this sort of, right, you’ve got on the postmodern, the left academic left, you get this thing, science is just a game. It’s just a game from the powerful come up with, which is also not true. Science is totally potent and requires an account for what is happening. So that’s another way to push science away or respond to it. The denial, science denial that happens. That’s also another way of not understanding the balance that science is trying, that we need to establish with experience.

(02:40:18)
And then there’s just pseudoscience, which wants to sort of say, oh, the new age movement or whatever, which wants to deal with experience by kind of elevating it in this weird pseudo spiritual way or that doesn’t have the rigor of science.

(02:40:33)
So all of these ways, all of these responses, we have this difficulty about experience. We need to understand how experience fits into the web of meaning, and we don’t really have a good way of doing it yet. And the point of the book was to identify very clearly how the problem manifests, what the problem is, and what its effects are in the various sciences.
Lex Fridman
(02:40:55)
And by the way, we should mention that at least the first two responses, they kind of feed each other just to observe the scientific community, those who gravitate a little bit towards the scientific triumphalism, is an arrogance that builds in the human soul. I mean, it has to do with PhDs, it has to do with sitting on an academic throne, all those things. And the human nature with the egos and so on, it builds. And of course, that nobody likes arrogance. And so those that reject science, the arrogance is fuel for the people that reject science.
Adam Frank
(02:41:33)
I absolutely agree.
Lex Fridman
(02:41:34)
It just goes back and is this divide that builds.
Adam Frank
(02:41:37)
Yeah, no, and that was a problem when you saw, so like I said, my first book was about science and human spirituality. So I was trying to say that science is actually, if we look at what happens in human spirituality, not religion. Religion is about politics, but about for the entire history of the species, we’ve had this experience of, for lack of a better word, the sacredness. I’m not connecting this God or anything. I’m just saying this experience of the more, and then with the new atheist movement, you got people saying that, “Anybody who feels that is an idiot.” They just can’t handle the hardcore science. When in fact their views of the world are so denuded of, they can’t even see the role that experience plays in how they came up with their formal systems. And experience fundamentally is weird, mysterious, it’s, it kind of goes down forever in some sense. There is always more. So yeah, that arrogance then just if you’re telling everybody who’s not hardcore enough to do the standard model of cosmology, that they’re idiots, that’s not going to bode well for the advance of your project.
Lex Fridman
(02:42:37)
So you’re proposing at least to consider the idea that experience is fundamental, experience is not just an illusion that emerges from the set of quirks, that there could be something about the conscious experience of the world that is at the core of reality?
Adam Frank
(02:42:56)
But I wouldn’t do it. I wouldn’t because there is panpsychism, right? Which wants to say-
Lex Fridman
(02:43:02)
Right. So that’s all the way there. So panpsychism is, that’s literally one of the laws of physics is consciousness.
Adam Frank
(02:43:04)
Right. But see what all those do is just the idea of say, physicalism versus idealism, which are kind of the two philosophical schools you can go with. Physicalism says, “All that exists as physical.” Idealism says, “All that exists is mind.” We’re actually saying, “Look, both of these to take either of those positions is already to project out into that third-person view. And that third-person view we want to really emphasize is a fiction.” It’s a useful fiction when you’re doing science. If I want to do the Newtonian physics of billiard balls on a pool table, great. I don’t want to have to think about experience at all, right? But if I’m asking deeper questions, I can’t ignore the fact that there really is no person view and that any story I tell about the world is coming from, it’s not just first person, but it’s literally because I’m going to argue that experience always involves all of us. Experience always originates out of a community.

(02:43:58)
That you are always telling those stories from the perspective of already existing, of already being in experience. So whatever account we want to give of the world is going to have to take that as experience as being irreducible and the irreducible starting point. So ultimately, we don’t have an answer. That’s when people are like, “Well, what are you suggesting is the alternative?” It’s like, look, that’s the good work of the next science to come. Well, our job was to point out the problem with this, but what we would argue with is, and we’re thinking about the next book, is this is really going to require a new conception of nature. That doesn’t sort of jump to that third-person… That fictional third-person view and somehow figures out how to do science. Recognizing that it always starts from experience. It always starts from this field of experience. Or in phenomenology, the word is the life world that you’re embedded in. You can’t un-embed yourself from it.

(02:44:52)
So how do you do… So one of the things that Whitehead said was, “We have to avoid the bifurcation of nature.” And what he meant by that is the bifurcation into scientific concepts, wavelength. Think about seeing a sunset. You can say like, “Oh look, it’s just wavelengths and scattering particles.” And your experience of the redness, the actual experience of the redness and all the other things. It’s not just red. There’s no qualia, there’s no pure redness. Everything that’s happening in the experiential part is just an epiphenomena. It’s just brain states, whatever. He said, “You can’t do that. They’re both real. They’re both accounts. They both need to be integrated.” And so that required, I think, really a different of what we mean by nature.
Lex Fridman
(02:45:34)
Is it something like incorporating in the physics, in the study of nature, the observer, the experiencing observer, or is that still also looking from a third-person?
Adam Frank
(02:45:45)
I think that that’s what we have to figure out. And so actually a great place to think about this is quantum mechanics, because one of the things we’re arguing is look.. In the chapter that I wrote on, because I wrote on, because I wrote this with Evan Thompson, who’s a wonderful philosopher, and Marcelo Gleiser, who’s a theoretical physicist. When I was writing the chapter on the origin of The Blind Spot, sort of how this emerged out of history, the subheader was like, “Well, it made sense at the time.” Because it did. It really, there was a reason why people adopted this third person, God’s eye deterministic view. This view of sort of like, yeah, the perfect clockwork of the universe. Yeah, totally made sense. But by the time you got to the beginning of the 20th century, science itself was telling you, “Eh-eh.” And no place does this appear more than in quantum mechanics, right?

(02:46:29)
Quantum mechanics slams you with the idea of the measurement problem. And most important thing about quantum mechanics is you have a dynamical equation, the Schrodinger equation, which you put in, like we talked about before, you have initial conditions and now you’ve got a differential equation and you crank out the differential equation and it makes predictions for the future, right? Exactly like Newtonian physics or its higher versions of the Lagrange or Hamiltonians. But then this other thing happens where it’s like, oh, by the way, as soon as you look at it, as soon as the measurement is made, I have a whole nother set of rules for you. That’s what we call the born rule. And it was telling you right from the beginning that measurement matters, right? So when you’re asking, how will we do this? Quantum mechanics is actually pointing to how to do it.

(02:47:17)
So there’s been all these different interpretations of the quantum mechanics. Many of them try to pretend the measurement problem isn’t there. Go to enormous lengths like the many-worlds interpretation, literally inventing an infinite number of unobservable parallel universes to avoid the thing that quantum mechanics is telling them, which is that measurements matter. And then you get something like QBism, which is I’m going to advocate for, is a new interpretation of quantum mechanics, which puts the Born rule at the center. Instead of focusing on the Schrodinger equation and the weird things that come out of it, like Schrodinger’s cat and all that other stuff. It says, “No, no, actually the real mystery is the Born rule. Let’s think about the Born rule.” And like you said, that puts the agent, the agent and information at the center of the whole thing.
Lex Fridman
(02:48:01)
So that’s not a thing you’re trying to get rid of? That’s the thing you’re trying to integrate at the center of the thing in quantum mechanics, it becomes super obvious, but maybe the same kind of thing should be incorporated in every layer of study of nature.
Adam Frank
(02:48:19)
Absolutely. That’s exactly it. So one of the things that’s really interesting to me, so we have a project, I’m part of a big project that Chris Fuchs and Jacques Pienaar on QBism. So I’ve been part of that. And what I’ve been amazed by is the language they use. So what’s cool about QBism is it comes from quantum information theory. It’s a pretty modern version of thinking about quantum mechanics. And it’s always about do you have an agent who makes an action on the world? And then the information they get from that action through the experiment, that’s the action on the world. Updates, their priors updates, their Bayesian, that’s why it’s called QBism. Quantum Bayesianism updates how the information they’ve gotten from the world. Now, this turns out to be, it’s kind of the same language that we’re using in a project that’s about the physics of life, where we have a grant from the Templeton Foundation to look at semantic information and the role of semantic information in living systems like cells.

(02:49:16)
So we have Shannon information, which is a probability distribution that tells you basically how much surprise there is in a message. Semantic information focuses on meaning, right? Focuses on in a very simple way, just how much of the information that the agent, the critter is getting from the world actually helps it survive. That’s the most basic idea of meaning. We can get all philosophical about meaning, but this is it. Does it help me stay alive or not? And the whole question of agency and autonomy that occurs in this setting of just asking about how do cells move up a chemical gradient to get more food? Kind of has the feel the same sort of architecture as what’s going on in quantum mechanics. So I think what you said is exactly it, how do we bring this sort of recognition? That there’s always us, the agent or life the agent interacting with the world and drawing both giving information and passing information back as a way of doing science, doing hardcore science with experiments, but never forgetting that agency, which also means experience in some sense, is at the center of the whole thing.
Lex Fridman
(02:50:27)
So you think there could be something like QBism, Quantum Bayesianism that creates a theory, like a Nobel Prize winning theory, sort of hardcore real theories that put the agent at the center.
Adam Frank
(02:50:42)
Yes. That’s what we’re looking for. I think that is really, that’s the exciting part. And it’s a move, the scientific triumphalist thing says, you understand why people love this? I have these equations. And these equations represent, there’s this platonic ideal that they are, they exist eternally on their own. It’s kind of quasi-religious, right? It’s sort of somehow, look, these equations are the, you’re reading the mind of God, but this other approach to me is just as exciting because what you’re saying is there’s us and the world, they’re inseparable. It’s always us and the world. And what we’re now finding about is this co-creation, this interaction between the agent and the world such that these powerful laws of physics that need an account. In no way am I saying these laws aren’t important. These laws are amazing, but they need an account, but not an account that strips, that turns the experience, turns the agent into just an epiphenomena, that it pushes the agent out and makes it seem as if the agent’s not the most important part of the story.
Lex Fridman
(02:51:45)
So if you pull on this thread and say, there’s a whole discipline born of this, putting the agent as the primary thing in a theory, in a physics theory, is it possible it just breaks the whole thing open? So there’s this whole effort of unifying general relativity and quantum mechanics of coming up with a theory of everything. What if these are the tip of the iceberg? What if the agent thing is really important?
Adam Frank
(02:52:18)
So listen, that would be kind of my dream. I’m not going to be the one to do it because I’m not smart enough to do it. Marcelo and I have for a while have been sort of critical of where foundational physics has been for a while. With string theory, I’ve spent my whole life listening to talks about, “String theory, real soon.” And it’s gotten ever more disconnected from data, observations. There were people talking for a while that it is post-empirical. I always wanted to write a paper or an article that was like, physicists have been smoking their own stash. There’s this way we’ve gotten used to, you have to out-weird the other person, my theory has 38 dimensions. My theory is 22 dimensions, but it’s got psychedelic squirrels in it. And so there’s a problem. I don’t need to tell you there’s a crisis in physics or there’s a crisis in cosmology. Other people have used that. That’s been the headline on scientific American stories.

(02:53:18)
So clearly another direction has to be found, and maybe it has nothing to do with this, but I suspect that because so many times the agent or having to deal with the view from the inside or the role of agency. When it comes to time thinking that you can replace the block universe with the actual experience of time. Clocks don’t tell time. We use clocks to tell time. So maybe that even the fundamental nature of time can’t be viewed from the outside, that there’s a new physics theory that is going to come from, that comes from this agential, informational, computational view. I don’t know. But that’s kind of what I think it would be fertile ground to explore.

Nature of time

Lex Fridman
(02:54:05)
Yeah, time is really interesting one. Time is really important to us humans. What is time?
Adam Frank
(02:54:12)
Yeah, right. What is time? So the way we have tended to view it is we’ve taken, this is what, when Husserl talks about the surreptitious substitution, we’ve taken Einstein’s beautiful, powerful, formal system for viewing time, and we substituted that for the actual experience of time. So the block universe, where next Tuesday is already written down in the block universe, the four dimensional universe, all events are already there. Which is very potent for making certain kinds of predictions within the scientific framework. But it is not lived time. And this was pointed out to Einstein, and he eventually recognized it. Very famous meeting between Henri Bergson, who was the most famous philosopher of the early 20th century and Einstein, where Einstein was giving a talk on relativity-
Adam Frank
(02:55:03)
The 20th century and Einstein, where Einstein was giving a talk on relativity and Berkson, whose whole thing was about time and it was about duration. He wanted to separate the scientific image of time, the map of time from the actual terrain, which he used the word duration like we humans where duration for us is full. It’s stretched out. It’s got a little bit of the past, a little bit of the future, a little bit of the present. Music is the best example, right? You’re hearing music, you’re both already anticipating what’s going to happen and you are remembering what’s going on.

(02:55:34)
There’s a kind of phenomenal structure there, which is different from the representation of time that you have with the formal mathematics. And the way we would look at this is that the problem with the surreptitious substitution, the problem with the blind spot is it says, “Oh, no, no, the formal system is time,” but really the only place time appears is with us, where we’re so having a theory that actually could start with us and then stretch out into the universe rather than imposing this imaginary third-person view back on us. Could, that’s a route towards a different way of approaching the whole problem.
Lex Fridman
(02:56:13)
I just wonder who is the observer? I mean, define what the agent is in any kind of frame is difficult.
Adam Frank
(02:56:20)
Is difficult, but that’s the good work of the science ahead of us. So what happened with this idea of the structural invariance I was talking about? So we start with experience, which is irreducible. There’s no atoms of experience. It’s a whole, and we go through the whole process, which is a communal process, by the way. There’s a philosopher, Robert Crease, who talks about the workshop that started in the 1700s, 1600s, we developed this communal space to work in, sometimes it was literally a physical space, a laboratory where these ideas would be pulled apart, refined, argued over, and then validated. And we want to the next step.

(02:56:54)
So this idea of pulling out from experience, these thinner, abstract, structural invariance, the things that we could actually do science with, and it’s kind of like, we call it an ascending spiral of abstraction. So the problem with the way we do things now is we take those abstractions, which came from experience, and then with something like a computational model of consciousness or experience, we think we can put it back in. You literally pulled out these super thin things, these abstractions neglecting experience because that’s the only way to do science. And then you think somehow, oh, I’m going to jam experience back in and have an explanation for experience.
Lex Fridman
(02:57:36)
So do you think it’s possible to show that something like free will is quote, unquote real if you integrate experience back into into the physics model of the world?
Adam Frank
(02:57:46)
What I would say is that free will is a given. And that’s the thing about experience. So one of the things that Whitehead said, I really love this quote. It says, “It’s not the job of either science or philosophy to account for the concrete. It’s the job to account for the abstract.” The concrete, what’s happening between us right now, is just given. It’s presented to us every day. It’s presented to me. If you want an explanation, fine, but the explanation actually doesn’t add anything to it. So that free will in some sense is the nature of being an agent. To be an agent agency and autonomy are sort of the two things that are, they’re equivalent. And so in some sense, to be an agent is to be autonomous. And so then the question really to ask is, can you have an account for agency and autonomy that captures aspects of it’s arising in the world or the way it and the world sort of co-arise.

(02:58:41)
But the idea why we argue about free will often is because we already have this blind spot view that the world is deterministic because of our equations, which themselves, we treat the equations as if they’re more real than experience. And the equations are a paler… They don’t corral experience. They are a thinner representation. As we like to say, “Don’t confuse the map for the terrain.” What’s happening between us right now and all the weirdness of it. That’s the terrain. The map is what I can write down on equations. And then in the workshop, do experiments on. Super powerful, needs an account, but experience overflows that.
Lex Fridman
(02:59:17)
What if the experience is an illusion? How do we know what if the agency that we experience is an illusion?
Adam Frank
(02:59:26)
An illusion looking from where? Because that already requires to take that stance is you’ve already pushed yourself into that third person view. And so what we’re saying is that third person view, which now you’re going to say like, “Oh, I’ve got a whole other set of entities, of ontological entities,” meaning things that I think exist in God’s living room in spite that are independent of me and the community of living things I’m part of.
Lex Fridman
(02:59:51)
So you’re pushing it elsewhere just like there’s a stack of turtles is probably, if this experience, the human experience is an illusion, maybe there’s an observer for whom it’s not an illusion. So you always have to find an observer somewhere.
Adam Frank
(03:00:06)
And that’s why fundamentally the blind spot, especially the scientific triumphalist part is following a religious impulse. It’s wanting the god’s eye view. And what’s really interesting, and when we think about this and the way this gets talked about, especially publicly, there’s a line of philosophical inquiry that this language gets couched in and it is actually a pretty, it’s only one version of philosophy. So it is pretty much what we call the analytic tradition. But there’s even in Europe or in the Western tradition for Western, what we’ll call western philosophy, there’s phenomenology. And Heidegger and Merleau-Ponty, which took an entirely different track. They were really interested in the structure of experience. They spent all their time trying to understand, trying to develop a language that could kind of climb into the circle. That is experience, right experience. You’re not going to be able to start with axioms and work your way to it.

(03:01:00)
It’s given. So you have to kind of jump in and then try and find a language to account for its structure. But then, so that has not been part of this discussion about you’ll never, good luck finding a YouTube video where someone, a famous scientist is talking about science from a phenomenological point of view, even though it’s a huge branch of philosophy. And then you get the philosophies that occurred from other cores of civilization. So there’s the western core out of which comes the Greeks and the Judeo- Christian Islamic tradition. But then you get India and you get Asia and they developed their own. They were highly complex societies that developed their own responses to these questions. And they, for reasons they had contemplative practice. They were very focused on direct, trying to directly probe attention and experience. They asked questions in ways that the West never really did.

(03:01:52)
Phenomenology kind of started it, but there’s philosophers like Nagarjuna and Vasubandhu. They’re like the Plato and the Aristotle of those philosophies. And they were really focused on experience in the West. I think maybe because we had the Judeo-Christian tradition where we already had this kind of God who was going to be the frame on which you could always point to that frame in the traditions that came from the classical philosophies of India and Asia. They started always with this. They wanted to know about experience. Their whole philosophies and their logic and their argumentation was based on, I’ve got this experience, I can’t get out of this experience. How do I reason from it? So I think there’s a lot of other philosophical traditions that we could draw from. Not slavishly, we don’t all have to become Buddhists to do it, but there are traditions that really tried to work this out in a way that the Western traditions just didn’t.
Lex Fridman
(03:02:47)
But there’s also the practical fact that it’s difficult to build a logical system on top of experience. It’s difficult to have the rigor of science on top of experience. And so as science advances, we might get better and better. The same is it’s very difficult to have any kind of mathematical or kind of scientific rigor to why complexity emerges from simple rules and simple objects, sort of the Santa Fe questions.
Adam Frank
(03:03:16)
But I think we can do it. I think there’s aspects of it. I mean, as long as you’re never trying to, “This is what experience is,” I think that’s kind of where you’re never going to have a causal account of experience just given. But you can do lots about, and that’s what the good work is to how do I approach this? How do I approach this in a way that’s rigorous that I can do experiments with also? But so for example, I was just reading this beautiful paper that was talking about in this is what we’re counting with our semantic information too. Causal closure. Love this idea. The idea that… So we talked about autopoiesis a while back, the idea that living systems, they are self creating and self maintaining. And so the membrane, cell membrane is a great example of this, right? The cell membrane, you can’t have a cell without a cell membrane.

(03:04:02)
The cell membrane lets stuff through, keeps other stuff out. But the cell membrane is part of the processes and it’s a product of the processes that the cell membrane needs, right? In some sense, the cell membrane creates itself. So there’s this strange, it’s always with life, there’s always this strange loop. And so somehow figuring out how to jump into that strange loop is the science that’s ahead of us. And so this idea of causal closure accounting for how the, we talk about downward causation. So reductionism says everything only depends on the microstate. Everything just depends on the atoms. That’s it. If you know the Lagrangian for the standard model, you are done. Of course, in principle, you need God’s computer, but fine. In principle, it could be done. Causal closure, and I was just reading this great paper that sort of argues for this.

(03:04:57)
There’s ways in which using Epsilon machines and all this machinery from information theory, that you can see ways in which the system can organize itself so that it decouples from the microstates. Now, the macrostate fundamentally no longer needs the microstate for its own description, its own account of the laws, whether that paper is true or not. It’s an example of heading down that road. There’s also Robert Rosen’s work. He was a theoretical biologist who he talked about closure to efficient cause that living systems are organizationally closed, are causally closed so that they don’t depend anymore on the microstate. And he had a proof, which is very contentious. Nobody knows if it’s some argue it’s true, some argue it’s not. But he said that because of this, living systems are not church-turing complete, they cannot be represented as formal systems. So in that way, they’re not axioms, they’re not living systems will not be axioms.

(03:05:55)
They can only be partially captured by algorithms. Now again, people fight back and forth about whether or not his proof is valid or not. But I’m saying them giving you examples of when you see the blind spot, when you acknowledge the blind spot, it opens up a whole other class of kinds of scientific investigations. The book we thought was going to be really heretical. Obviously most public facing scientists are very sort of in that, especially scientific triumphant. So we were just waiting for the fight. Then the review from science came out and it was totally pro… It was very positive. We’re like, “Oh my God.” Then a review came out in Nature Physics and it was totally positive.

(03:06:38)
Then a review came out in the Wall Street Journal, kind of criticized, not capitalism, but we criticized all industrial economies for that they had been touched by the blind spots, socialism, communism. It doesn’t matter. These extractive sort of had that sort of view that the is just reducible to resources. The Wall Street Journal gave us a great review. So it feels like there’s actually out there, there is some, among working scientists in particular, there is some dissatisfaction with this triumphalist view and a recognition that we need to shift something in order to jump past these hurdles that we’ve been arguing about forever and we’re sort of stuck in a vortex.
Lex Fridman
(03:07:18)
Well, it is. I mean, I think there is a hunger to acknowledge that there’s an elephant in the room, that we’re just removing the agent. Everyone is doing it and it’s like, yeah, yeah, there’s the experience and then there’s the third-person perspective on the world. And so, man, science from, applying scientific rigor from a first-person perspective is very difficult. I mean, it’s fascinating.
Adam Frank
(03:07:44)
I think we can do it. Also, the thing, what’s really interesting is I think it’s not just first-person, first and second, because one idea is that, the idea that, oh, science gives us this objective third-person view. That’s one way of talking about objectivity. There’s a whole other way is that I do the experiment. You do the experiment, we talk to each other, we agree on methods, and we both get the same result. That is a very different way of thinking about objectivity, and it acknowledges that when we talk about agents, agency and individuality are flexible.

(03:08:18)
So there’s a great paper, speaking of Santa Fe by David Krakauer, where they looked at sort of information, theoretic measures of individuality. And what you find is it’s actually pretty fluid. My liver cell is an individual, but really it’s part of the liver. And my liver is a separate system, but really it’s part of me. So I’m an individual, yay. But actually I’m part of a society and I couldn’t be me without the entire community of say, language users. I wouldn’t even be able to frame any questions. And my community of language users is part of ecosystems that are alive, that I am a part of, a lineage of. This is like Sarah Walker stuff, and then those ecosystems are part of the biosphere. We’re never separable as opposed to this very atomizing, the triumphal, this science view is wants like Boltzmann brains, you’re just a brain floating in the space.
Lex Fridman
(03:09:07)
There is a fascinating degree to which agency is fluid. You are an individual, but you and I talking is the kind of individual, and then the person listening to this right now is also an individual. I mean, that’s a weird thing too.
Adam Frank
(03:09:24)
That’s a weird thing, right?
Lex Fridman
(03:09:26)
Because there’s a broadcast nature too.
Adam Frank
(03:09:29)
This is why information theoretic. So the idea that we’re pursuing now, which I get really excited about, is this idea of information architecture or organization. Organizational organization. Because physicalism is like everything’s atoms, but Kant recognized, Kant is apparently the one who came up with the word organism. He recognized that life has a weird organization that would see specifically different from machines. And so this idea that how do we engage with the idea that organization, which is often I can be cast in information theoretic terms or computational terms even. It’s not really quite physical. It’s embodied in physical, in the physical. It has to instantiate in the physical, but it also has this other realm of design, not design like intelligent design, but there’s a… The organization itself is a relationship of constraints and information flow. And I think again, that’s an entirely new interesting way that we might get a very different kind of science that would flow out of that.

Cognition

Lex Fridman
(03:10:29)
So going back to Kant and organism versus machine. So I showed you a couple of legged robots.
Adam Frank
(03:10:40)
Very cool.
Lex Fridman
(03:10:41)
Is it possible for machines to have agency?
Adam Frank
(03:10:44)
I would not discount that possibility. I think there’s no reason I would say that it’s impossible that machines could, whatever it manifests that strange loop that we’re talking about that autopoiesis I don’t think there’s a reason to say it can’t happen in silicon. I think whatever, it would be very different from us, the idea that it would be like, oh, it would be just like us. But now it’s instantiated and I think it might have very different kind of experiential nature. I don’t think what we have now, the LLMs are really there, but yeah, I’m not going to say that it’s not possible.
Lex Fridman
(03:11:26)
I wonder how far you can get with imitation, which is essentially what LLMs are doing. So imitating humans, and I wouldn’t discount either the possibility that through imitation you can achieve what you would call consciousness or agency or the ability to have experience. I think for most of us humans to think, oh, that’s just fake. That’s copying. But there’s some degree to which us, we humans are just copying each other. We just are really good imitation machines coming from babies. We were born in this world and we’re just learning to imitate each other. And through the imitation and the tension in the disagreements in the imitations. We gain personality, perspective, all that kind of stuff.
Adam Frank
(03:12:08)
Yeah, it’s possible, right? It’s possible. But I think probably the view I’m advocating would say that one of the most important parts of agency is there’s something called, E-four. The E-four theory of cognition, embodiment, enaction, embedding, and there’s another one, extension. But so the idea is that you actually have to be in a body which is itself part of an environment that is the physical nature of it and of the extension with other living systems as well is essential.

(03:12:46)
So that’s why I think the LLMs are not going to, it’s not imitation. It’s going to require, this goes to the brain in the VAT thing. I did an article about the brain in the vat, which was really Evans, I was reporting on Evans. Where they did the brain in the VAT argument. But they said, “Look, in the end, actually the only way to actually get a real brain in the VAT is actually to have a brain in a body.” And it could be a robot body, but you still need a brain in the body. So I don’t think LLMs will get there because they can’t. You really need to be embedded in a world, at least that’s the E-four idea.
Lex Fridman
(03:13:13)
The E-four, the 4E approach to cognition argues that cognition does not occur solely in the head, but is also embodied, embedded, enacted, and extended. And by way of extra cranial processes and structures, they’re very much in vogue. 4E cognition has received relatively few critical evaluations. This is a paper, but reflecting on two recent collections, this article reviews the four E paradigm with a view to assessing the strengths and weaknesses. It’s fascinating. I mean, yeah, the branches of what is cognition extends far, and it could go real far.
Adam Frank
(03:13:49)
Right. There’s a great story about an interaction between Jonas Salk, who was very much a reductionist, the great biologist, and Gregory Bateson, who was a cyberneticist, and Bateson always loved to poke people. And he said to Salk, he said, “Where’s your mind?” And Salk went, “Up here,” and Bateson said, “No, no, no, out here.” And what he really meant was this extended idea. It’s not just within your cranium to have experience. Experience in some sense is not a thing you have. It is a thing you do. Almost perform it in a way, which is why both actually having a body, but having the body itself be in a world with other bodies is, from this perspective, is really important. And it’s very attractive to me. And again, if we’re really going to do science with them, we’re going to have to have these ideas crash up against data, crash up against, we can’t just armchair it or couch quarterbacking it, but I think there’s a lot of possibility here. It’s a very radically different way of looking at what we mean by nature.

Mortality

Lex Fridman
(03:14:53)
What do you make of the fact that this individual observer, you as an individual observer only get a finite amount of time to exist in this world? Does it make you sad?
Adam Frank
(03:15:04)
No, actually it doesn’t make me sad. Okay, so full reveal, I have been doing contemplative practice in the zen tradition for 30 years. I’ve been staring at a wall for 30 years, and it’s taught me a lot. I really value what that practice has given me about the nature of experience. And one of the things it’s taught me is I don’t really matter that very much. This thing I call Adam Frank is really, it’s kind of a construct. There’s this process going on of which I am actually fundamentally, and that’s super cool, but it’s going to go. I don’t know where it came from. It’s going to go, I don’t really need it to, and then who the hell knows? I’m not an advocate for an afterlife, but just that what I love, zen, has this idea of beyond birth and death, and they don’t mean reincarnation. What they mean is, “Dude, you don’t even really understand what life is.” You know what I mean? I’m like this core level of your own experience. So your ideas about what death is are equally ill-formed.

(03:16:07)
The contemplative practice really tries to focus on experience itself. Spend five days at a zen session doing contemplative practice from 7:00 AM. until 9:00 PM, obviously with breaks. And you’ll really get a much deeper understanding of what my own experience is. What is it really like? It forces you to learn how to stabilize your attention because attention is kind of like this thing. It’s usually just like, “Oh, over there. Oh, my foot hurts. Oh, I got to do my taxes. Oh, what’s that guy over there? Why is he wearing those stupid shoes?” And with a contemplative practice, you learn how to stabilize it.

(03:16:39)
And once you stabilize it, you can now begin to sort of explore the phenomenal nature of it. So what I think I’ve learned from that is kind of whatever, I’m not really kind of real to begin with. The Adam Frank part, the identity, the thing, and the part of me that is real is everything’s coming and going. It’s all coming and going. Well, how could I ever not come and go? And the entire world is just, Buddhism has this idea of co-dependent arising. Nothing exists, nothing has self-nature. Nothing exists by itself. It’s an endless, infinitely connected web.
Lex Fridman
(03:17:15)
But still, there’s a deliciousness to the individual experience. You get attached to it and it ends and it’s good while it last, and it sucks that it ends. You can just be like, “Ah, well, everything comes and goes,” but I was eating ice cream yesterday. I found this awesome low-carb ice cream called, Delights here in Austin, and it ends. And I was staring at the empty container, and it was-
Adam Frank
(03:17:42)
That’s beautiful, man. I love that.
Lex Fridman
(03:17:44)
You could say, “Yeah, well, that’s how it all is, but…”
Adam Frank
(03:17:47)
Can I say that what I’ve learned from, because I love your idea of the deliciousness of it. But what I think happens with contemplative practice when it deepens is that you’re not just saying, this is I do koan practice. So this is a tradition in zen that it was established, it was a teaching method that was established a thousand years ago. They’re these book of koans. And every koan, if you’ve ever read Godel, he’s got a whole chapter on koans. They’re kind of non-logical problems that you have to work on. One of my favorite one was, “Stop the sound of the distant temple bell.”

(03:18:23)
You’re like, “What?” Every time my teacher gives it to him, I’m like, “What are you talking about?” This is the whole zen thing of up is down, but down is up. You must understand this. So your job with these koans is to sit with them, is to sit with them until you realize what the thing is trying to teach you what aspect of experience it’s trying to teach you. So there’s no answer. No. And in fact, actually, you don’t give an answer. You actually usually have to demonstrate. The first time when I did a call on and the guy was like, “Don’t tell me the answer, show me the answer.” I was like, what are you talking about? But after doing these for years now, I’ve kind of learned the language of them. So I could never tell you. If I told you the answer, I could give you a call and tell you the answer. You’d be like, “What?”

(03:19:05)
It’s not the words, it’s the So your experience of like, yeah, the cup is empty. With a contemplative practice as it deepens over years, it really does take years. Just like anything in math, it took me years to understand the Lagrangians. You kind of come to a deeper understanding with yeah, the words of, it’s not just like, oh, everything changes. You actually feel that movement. You feel it with breath to breath, and it really becomes, sometimes I have this feeling, this is messed up, but of just joy and it’s not connected to anything. That’s what I’ve kind of gotten from practice. It’s just like, yeah, that passage, that infinite passage of moment to moment that is truly the way things are. And it’s okay. It’s not okay because I have a feeling about it. Okay. I want it to be okay. It just is okay. And so really, it’s a pretty awesome thing.
Lex Fridman
(03:19:51)
Yeah, that’s beautiful. Maybe it’s the genetics, maybe it’s the biochemistry in my brain, but I generally have that joy about experience, amorphous joy. But it seems like, again, maybe it’s my Eastern European roots, but there’s always a melancholy that’s also sitting next to the joy, and I think it always feels like they’re intricately linked. So the melancholy is about, maybe about the finiteness of experience, and the joy is just about the beauty of experience, and they’re just kind of sitting there.
Adam Frank
(03:20:22)
Which is cool actually, because I’m also, I come from Eastern, my roots are Eastern European as well, going back, and I get it right, but that’s also the cool thing. I think one of the things is, well, that is what it is. That is what it is. You don’t have to do anything. You don’t have to manipulate it or move it around or yeah, this is the experience.
Lex Fridman
(03:20:41)
Can you speak to just the practical nature of sitting there from 7:00 AM to 9:00 PM?
Adam Frank
(03:20:45)
I’m like, what the hell are you doing, bro?
Lex Fridman
(03:20:46)
What’s powerful? What’s fascinating to you? What have you learned from just the experience of staring at a wall?
Adam Frank
(03:20:51)
Yeah. Yeah. So not really. I mean, you’re staring. You’re facing a wall, and what you’re doing is you’re just sitting with, there’s different meditative practices, there’s counting breaths. So that’s usually what I do. I sit down and I start counting breaths, and for the first half hour it’s just like, blah, blah, blah. Like I said, I’m thinking about my taxes. I’m thinking about what I got to do later on, yada, yada, yada. First time I ever did a full session, a two-day session, I swear to God, I had Bruce Springsteen’s, Born To Run album track through from the beginning to the end with the pauses. This was back in when there were LPs with the fricking pauses.
Lex Fridman
(03:21:22)
Nice.
Adam Frank
(03:21:24)
My mind was just like, I need to do something. So it literally played the whole album in order.
Lex Fridman
(03:21:28)
That’s pretty cool, actually.
Adam Frank
(03:21:29)
Yeah, it was pretty amazing to see because you really do, you see the dynamics of your mind. But what happens is, and this took me a while, I used to hate sitting. I do it, but after a while the mind gets exhausted. That part of the mind, the upper level, the roof brain chatter is just like, there’s nothing else to do. And then you get bored. And now I realize that’s when something interesting is going to happen. You drop down and now it’s a very physical practice. People think you’re just sitting there not thinking or thinking about not thinking. Actually, it becomes a very physical process where you’re really just following the breath, you’re kind of riding the breath and it gets very quiet. And within that quietness, there’s a path. Because obviously there’s been, Buddhism is always not about thinking, but there’s a huge literature.

(03:22:18)
So these guys are always about, don’t think. I’ve written all this stuff, but they’re guideposts. They’re like the finger pointing at the moon. And there’s the idea of first, your mind is usually scattered. Right now, when I walk out, I’m going to go get the Uber and everything. My mind’s going to be all over the place, but with sitting, first, you concentrate the mind so that there’s no more scatter anymore. The thoughts are still happening, but you’re just not there happening up there. You’re not even paying attention to them. And then as time goes on, you unify the mind, which is this very powerful thing where kind of the self drops away and there’s just this presence.

(03:22:49)
It’s kind of like a raw presence, and that’s often where the joy up wells from, but you sit with whatever, maybe you’re going to sit and maybe you’re going to go through an hour of being bummed out about your mom who died or something. You’re just going to sit with whatever comes up you’re going to make. That’s why the sitting part, you’re making the commitment. I’m going to sit here with whatever comes up, I will not be moved. And then what You come away with it actually over time, it actually changes kind of who you are. I’m still the asshole I was from New Jersey growing up, but I just have more space now for things.
Lex Fridman
(03:23:24)
Once Jersey, always Jersey.
Adam Frank
(03:23:26)
Always Jersey.
Lex Fridman
(03:23:26)
But I love the Bruce Springsteen is just blasting in your head.
Adam Frank
(03:23:29)
Yeah, that was amazing.
Lex Fridman
(03:23:30)
Why are we here? What do you think is the purpose, the meaning of human existence?
Adam Frank
(03:23:35)
It’s good that we just had the last conversation because I’m going to give this answer, which is so corny. It’s love, and I’m not messing around. Because really actually, what happens, so within Buddhism, there’s the idea of the Bodhisattva principle. You’re here to help. You’re just here to help, right? Compassion. That’s a really essential part of this path, of the Dharma path. And when I first started, I was like, “I don’t care about compassion. I’m here for knowledge.” I started contemplative practice because of the usual thing I was suffering. The reason everybody comes to things like this, life was hard. I was going through stuff, but I also wanted knowledge. I wanted to understand the foundational nature of reality. So it was like compassion or whatever. But then I found out that you can’t get that. You can’t get those. You can’t go to this level without compassion.

(03:24:18)
Somehow in this process, you realize that it really is about helping all sentient beings. That’s the way they frame, just being here to help. So I know that sounds cornball, but especially for a guy from Jersey, which is the main thing is to get over. Your job is to get over. But that’s really what I found. It is actually kind… And so that joy, the joy, some of that joy is just, it’s like this. One of the things I have when I have really, there’s a kind of experience I’ll have in contemplative practice, which will carry out into the world, which is just this gratitude for the fact that the world gives you everything, and there’s a certain way, just the blue sky and the breath, the world is just giving you itself completely unhindered. It holds nothing back. And yeah, that’s kind of the experience. And then you kind of like, “Oh, I need to be helpful, because who’s not having this experience.”
Lex Fridman
(03:25:09)
So just love for the world as it is?
Adam Frank
(03:25:10)
Love for the, and all the beings who are suffering. Everybody’s suffering, everybody’s suffering. Your worst political opponent, they’re suffering. And our job is just to try and drop our biases and our stories and see this fundamental level at which life is occurring.
Lex Fridman
(03:25:26)
And hopefully there’s many alien civilizations out there going through the same journey, out of suffering, towards love.
Adam Frank
(03:25:33)
That may be a universal thing about what it means to be alive.
Lex Fridman
(03:25:36)
I hope so.
Adam Frank
(03:25:37)
I hope so too. Either that or they’re coming to eat us.
Lex Fridman
(03:25:39)
Especially if they’re a type three civilization.
Adam Frank
(03:25:41)
That’s right. And they got really big guns.
Lex Fridman
(03:25:45)
Well, this was truly mind-blowing. Fascinating. Just awesome conversation. Adam, thank you for everything you do, and thank you for talking today.
Adam Frank
(03:25:52)
Oh, thank you. This was a lot of fun.
Lex Fridman
(03:25:54)
Thanks for listening to this conversation with Adam Frank. To support this podcast, please check out our sponsors in the description. And now, let me leave you with some words from Carl Sagan. “The cosmos is all that is or ever was, or ever will be. Our feeblest, contemplations of the cosmos stir us. There’s a tingling in the spine, a catch in the voice, a faint sensation as if a distant memory or falling from a height. We know we are approaching the greatest of mysteries.” Thank you for listening and hope to see you next time.