Tag Archives: google

Do Not Cut Research Funding, Double It

Posted on by
Reply

nih-research$85 billion in federal spending cuts hit last week. In an attempt to gain the nation’s attention, a lot of the politicians and media outlets over-dramatized the short-term impact of these cuts. There will be some jobs lost, there will be some pay cuts, but in general the majority of the negative consequences will be in the long term.

I said it before many times, so I’ll just quote a recent Reuters article by Gabriel Debenedetti and Peter Rudegeair:

Federally funded, university research has long been a major engine of scientific advancement, spurring innovations from cancer treatments to the seeds of technology companies like Google.

Somehow investing into the future by starting a bunch of research projects that pursue some wild ideas doesn’t seem to be something that’s easy to sell to the general public. In fact, any kind of long term investment seems to be a hard sell, because it’s basically asking people to pay now for stuff they won’t get sooner than a decade from now.

We are $511 billion behind China in research investment. That might not mean much today, but in 10, 20 years, it will mean a huge shift in technological and economic power from the west to the east.

Avoid the Time Sink of Solving Small Problems

Posted on by
Reply

Linux (and the Computer in General) is Full of Time Sinks

time-sinkI just now went through the experience that took me just over an hour of concentrated “effort”. I regularly work on Windows, Linux, Android, and occasionally OS X. So here I am in Emacs, trying to do a regular expression forward search with the default Ctrl-Alt-S keyboard shortcut. You don’t need to know what any of that means. Basically I needed to do something simple that has worked in the past, and it didn’t work this time.

This time, pressing Ctrl-Alt-S did something else: it minimized the window. Okay, fine. I identified the problem, and went to try the first natural solution. I disabled the global keyboard shortcut for Ctrl-Alt-S. Of course I had to Google it to remind myself how to do it. That took about 5 minutes. Great. I thought I was done. Now, I went back and the Ctrl-Alt-S shortcut was still doing the same thing. “What the hell?” I thought, and… so on.

Long story short, I spent a total of 1 hour from start to finish finding the problem and the  solution that is nicely summarized in this Ask Ubuntu post.

Is It Worth It?

Looking back now, I’m struck with the question of why the hell I wasted 1 hour on something that does not add more than a small inconvenience to my work and my life. I encounter decisions like these sometimes several times a day. Here are some pros and cons in considering whether it’s worth it finding the fix for these problems.

Pros:

  • I fix the small problem the “right” way, so I don’t have to settle for a half-ass sweep-under-the-rug solution.
  • I get to spend an hour being open to the possibility of learning new things relevant to my work and exchanging information with the community of fellow Linux users. The computer world evolves on a daily basis, with new technologies, approaches, ideas constantly emerging. So, it’s important to stay connected to the latest developments.
  • I get to practice persevering in dealing with frustrating issues.

Cons:

  • It is not guaranteed that I find a solution.
  • It always takes longer than you thin. At first, it seems like the fix would not take more than 5 minutes, and when it does (50+% of the time), I become progressively more invested in it as time goes on. It’s the same principle that keeps you gambling in a Casino until all your money is gone.
  • Perfectionism is an addiction. We live in a world of inefficiencies that could be easily optimized if you just give it a few minutes. Sometimes it does take minutes, but sometimes it may take days. The more you feed this addiction, the harder it becomes to exist peacefully in an inefficient world.

Do One Thing, and Ignore Everything Else

To broaden this out a little bit. I know most of us have a to-do list of 10+ items every day. That list can grow to hundreds of items if you let it. On the other hand, a truly productive day is one where you usually focus on just one or two of those to-do items and do nothing else. That contradictions is at the core of my struggle as I think, read, program, etc.

Distractions take many forms, and sometimes they come at you disguised as urgent problems, when in reality they are nothing that can’t wait for a week, a month, or maybe an eternity.

Bridging the Gap Between Doctor and Patient

Posted on by
Reply

According Marty Makary‘s book Unaccountable, over 30% of medical procedures carried out in the United States are unnecessary. There are over 200,000 deaths from medical mistakes a year. Medical mistakes and preventable infections taken together is the #3 cause of death in the United States.

The problem is a lack of transparency in healthcare. I think the debate about healthcare in this country has been missing the main problem: in order for consumers to make a decision about their health, they need to have INFORMATION about the health services they’re “shopping” for. There is a culture of secrecy among doctors, as a result most consumers choose their doctor almost randomly based on factors that do not directly relate to their essential unmet needs.

Here’s my example of a good system… First, I Google the symptoms I’m having to get some basic information about the underlying conditions that could be causing the symptoms. As an imaginary example, say I suspect that I have an abdominal hernia. I look it up online, and it says that normally this requires a hernia repair surgery that costs an average of $5,000. I would like to be able to go online and find doctors near me that perform this surgery often and have gotten good reviews. Also, I would like to sort those options by price of the surgery. This data is out there! But unfortunately it’s not easily accessible.

Instead, as a consumer, we can’t know if the doctor we went to has done zero, one, or one hundred hernia repair surgeries. We don’t know up front what all the options are and what the price of each option is. What are the benefits, what are the costs?

According to Makary, doctors are pressured to hit quotas in the number of surgeries they perform, and maintain secrecy to avoid exposing themselves to lawsuits. I think doctors should be exempt from lawsuits. The only punishment for a screw-up should be whatever the consumer decides is appropriate when that information is made public.

Makary provides a lot of practical ideas for opening the channels of honest communication. For example, he suggests that all patients be provided with the video of their operation.

Check out a C-SPAN interview with Marty Makary where he preaches the message of openness in the medical profession.

Big Data Is In Need of a Big Brain

Posted on by
Reply

One of the big wide-open opportunities for innovation in the 21st century is around “big data”. Every day, approximately 2 billion gigabytes of data is created worldwide, and it’s growing exponentially (doubling per-capita every 40 months). In other words, we are flooded with information. The challenge is to process that information in a timely intelligent way so that we may live longer happier more productive lives (assuming all those things are not contradictory).

I think the task at hand is similar to the one the human brain does on a regular basis: it takes in several gigs of sensory data a second, and uses only a tiny fraction of that to arrive at facts about the past, present, and future state of its surroundings. As the universal stream of data grows (beyond comprehension), the collective machine brain has to grow with it as it scrambles desperately to pick out the useful bits.

Companies like Google are leading the way in that effort. But of course since big data is flooding every crevice of our lives, every company battling it out for our dollar, will have to invest in some kind of “big data analytics”. There are lots of opportunities for incremental improvements, but there are even more opportunities for futurists and dreamers to write books about the inevitable rise of intelligent machines, who will surely be based on some mixture of neural networks and genetic programming ;-)

The following is a good Google Tech Talk discussion on the current trends in big data:

YouTube Feature Suggestion: Paid Subscriptions to Channels

Posted on by
Reply

One of the things that makes YouTube popular is that most of the videos on there are free. I still refuse to believe that YouTube makes any profit in its current ad-supported operation. I think the plan is long term: to be the video hub of the future for ALL your video-viewing interests. Side note: I’m not sure how more eyes will equal more money (since more eyes equals higher costs too), but hey, I’m not going to speculate much further on this…

The feature that I would love to see YouTube implement is paid subscriptions to channels. For example, I’m a member of MGInAction.com and BJJWorldChampion.com which are sites that regularly publish instructional videos on grappling. I’m also a member of Lynda.com which publishes instructional videos on all kinds of software that I use.

It would be awesome if all of their content could move to YouTube, to a unified social network system. I could share some of the videos with my friends for free which could serve as excellent advertisement and would allow me share with the world how much I’ve learned from a particular video.

Anyway, I wanted to drop this little comment in the bottomless web bucket as a kind of Friday night prayer to the Google Gods so that they may continue innovating the heck out of our online experience.

 

The Cat and Mouse Game of General Intelligence

Posted on by
Reply

Google researchers published a paper that at once fills my heart with hopeful joy and eternal sadness. Joy because people care and are investing resources into developing intelligent systems. Sadness because of how poorly such systems perform 40+ years after the release of Kubrick’s 2001: A Space Odyssey.

The authors of the paper put together a large neural network that ran on a thousand 16-core machines for three days “learning” from a dataset of 10 million 200×200 pixel images.

The task is to train a face detector without ground truth (labeling images as containing a face or not). This task is absurdly difficult and I would even say just plain absurd. It’s like trying to teach a child algebra by giving him addition problems, but not ever telling him how to do addition or what the right answer is. It’s a fascinating and brave question to ask, because of how counter-intuitive it is.

Not surprisingly, the “breakthrough” that the paper touts is a 15.8% accuracy of classifying the objects in one of 20,000 categories. This is apparently a good improvement over the previous state of the art. My question is, in what universe is 15.8% deserving of a New York Times article? Granted it does exceed the approval rating of Congress, but that’s about it.

I don’t mean to be so dismissive. This is an excellent paper that scratches at the surface of an immense mystery: the gap between the most powerful supercomputer and the most primitive human brain? What’s even more exciting is that Google is funding this research and even more importantly putting its immense computational resources behind it.

Google Innovates, Microsoft Duplicates

Posted on by
2

I think implied in the title, and the observation that Microsoft seems to copy other company’s innovations, is that copying is a bad thing. As a researcher myself, I obviously respect and admire innovation. That’s why I love Google and other companies that fearlessly push the envelope of the possible. So from that perspective, copying ideas or making small incremental improvements on existing ideas is less than admirable.

However, I think that Microsoft is serving a critical role, and doing so better than almost any other technology company out there. They are the elephant in every room of every building, so risky innovation is perhaps the wrong strategy for them. The release of Internet Explorer 9, for example, is claimed by many to come “too late“. It’s essentially a good copy of the Google Chrome browser, just like Zune was a copy of iPod, that came out much later than the original.

I disagree with the claim that IE 9 came out “too late”. I think IE will regain market share, because it will be the default browser on all new Windows machines and now that the major problems with IE are fixed, non-technical people will be less inclined to switch away from it when they purchase a new computer.

That’s what seems to work for Microsoft, and is keeping me (arguably, a power user) happy switching back and forth from Linux to Windows (instead of just staying in Linux 100% of the time). The strategy is to let other people fight it out in the realm of risky innovations, and then copy the winners of that process as quickly as possible. I wouldn’t do it that way, but I’m glad someone is. The world runs on incremental improvement.