Thursday, December 08, 2011

A Great Time to be a Computer Scientist

Ask your friends if they'll be driving an electric car in ten years. The answer: No, cars will be driving us.

Today is the 105th anniversary of the birth of computing pioneer Grace Murray Hopper and Computer Science Education Week is being held this week in her honor. If only she could see what we have reaped from what she has sown.

The New York Times this week devoted Tuesday's Science Times to the Future of Computing. The section includes a series of essays from CS researchers including Daphne Koller, Stefan Savage, David Patterson, Kai-Fu Lee and fellow theory blogger Scott Aaronson. Scott also blogged about the experience. Though what would you do with a quantum computer on your desk? The factoring number thing could get boring pretty quick.

The Times also has a crowdsourced interactive timeline of future computing events. Try to guess which one I submitted. Almost all the advances listed are realistic and should keep CS very active for the long future. The most exciting future computer science advances will be the ones we haven't even thought of.

Finally the new class of ACM Fellows includes many theorists including Serge Abiteboul, Guy Blelloch, David Eppstein, Howard Karloff, Susan Landau, Joe Mitchell, Janos Pach and Diane Souvaine. A great year to be an ACM Fellow because the awards ceremony in June will follow a workshop celebrating the Turing Centenary featuring over 30 former Turing award winners.


  1. I think the most promising application of quantum computers remains the first one ever proposed: simulating quantum systems. Right now this task accounts for a large fraction of supercomputer time, and the approximations we get are still very imperfect.

  2. Google's self-driving car right now has something like $300-400K in sensors attached to it. For it to start to go mass-market, they need to get down to $10-20K. Ten years for these cost reductions is overly optimistic. I give it ten years for the technology to get more practical, and another ten years for cost reductions. Maybe truckers, military and luxury vehicles will start to get this technology incorporated, though. In another twenty years, maybe the technology will start to become mass market and in another twenty years maybe it will be ubiquitous. But that's 60 years away (and everyone knows how reliable even a 20 year projection is). It is not responsible for computer scientists to propagate overly optimistic predictions like this. You'll just disappoint people.

    I am not sure of that, even. From their presentations, Google's technology seems to have fundamental limits. It is based on memorizing every detail of the roads, with a human driver running each route before the computer can. It cannot sense and react to new roads. This is still a huge technology, but I do not know if it really is enough. For example, a trucker driving across the country will come across sporadic highway construction, where lanes have been shifted, on short notice and not always in a well-marked way.

    About having a quantum computer at your desk. In the future---thirty years, maybe---none of us will have computers at our desks, anyway. Computers and computer displays will be ubiquitous and highly networked. Quantum computers will be in super-computers as part of the cloud. Simulating quantum physics is obviously a big deal, but once we have quantum computers I'd expect them to be used even for problems without an exponential speedup. These algorithms have not been studied as much. It is hard to do good applied algorithms research without a computer. (Not impossible, hard.)

  3. In other news, an apple salesman predicts people will be eating more apples in the future...

  4. Speaking from a purely medical point-of-view, the recent literature amply documents that quantum apples are health-promoting and so (most likely) we are all going to end up eating quite a lot of these 21st century apples.

  5. Ten years for these cost reductions is overly optimistic.

    Let's do the math.

    A digital camera bought ten years ago for $900 can be bought today for $30, most of which pays for the lenses.

    The cost per megaflop in the last ten years went down by a factor of $350.

    There are lots of other similar examples for ladar/radar sensors, CCDs, etc.

    Considering this one can project the sensor cost that you quote to be somewhere in the $5-14K range in ten years just like Lance conjectured.

  6. Your cost predictions are nonsense. For example, consider the Canon 300D, available 8.5 years ago for US$900. The cheapest Canon DSLR today is $560. Yes, there are cheaper ones, the cheapest being ~$300. So a three-fold reduction in cost.

    Where has that cost reduction come from? Digital camera sensors are much cheaper today, as are any electronic chips, memory for example. Manufacturing efficiencies, and moving manufacturing to cheaper countries is the rest of it.

    Even if Google's technology follows the same cost curve, it will still be in the $100,000 range. And it won't. Computer processors and some related technologies have decreased in cost exponentially, but almost nothing else has. A car that cost $10,000 ten years ago will cost $15,000 today. Those big laser scanners that Google is using aren't going to get much cheaper.

  7. we are talking sensors so looking at a Canon dSLR isn't right

    2001 point and shoot 1.3mp camera = $300 to $350
    2011 point and shoot 1.3mp camera = $10

  8. A car that cost $10,000 ten years ago will cost $15,000 today.

    Price range for Toyota Camry in 2001, adjusted for inflation: $22,597 - $33,528

    Prince range for Toyota Camry in 2011: $21,995 - $24,725

    and that is before we adjust for the fact that the lowest range 2011 Camry (~$22,000) has about as many features as a medium to upper end 2001 Camry ($28,000+).

  9. Why is driver-less cars an exciting proposition? Do most car owners in the US hate driving their cars?

  10. The timeline description of "quantum computing" actually refers to when quantum computing will be used to introduce quantum mechanics in physics courses, not to when a quantum computer is available. Check it and see. I think the answer to when quantum computing will be used to introduce quantum mechanics is "never". I think if CS people took a few more physics courses then they would change their mind on this. There is a lot more to quantum mechanics than the abstract formalism.

  11. >Why is driver-less cars an exciting proposition? Do most car owners in the US hate driving their cars?

    Yes, we hate it--especially in Jersey.

    A fun application of driverless cars is less concern about parking: your car can just drop you off, and then pick you up.

  12. Do most car owners in the US hate driving their cars?

    You would too if it meant 70 minutes each way on bumper to bumper rush hour traffic.

    I'd rather read a book then. Also another big bonus of self-driven cars is less road fatalities.

  13. wouldn't a good metro solve these problems?

  14. Why spend $100K for a car that you won't drive?!!

  15. I expect my current car to still be running in ten years. Why should I buy another one just so it can drive me around?

  16. I suspect that by the time self-driving cars become economically viable, individual vehicles will not be economically viable. Not just fossil-fuel vehicles, but energy costs will be so high that Americans will finally suck it up and demand decent mass transit systems.

  17. Ideally, self-driving cars (that can communicate with each other in real time on the road) will provide several benefits: (1) lower accident rates; (2) faster travel (they will be able to drive in convoys, in effect simulating a train); (3) mobility for those with disabilities that currently prevent them from driving. Even those who can drive and enjoy driving will probably appreciate the option of switching to autopilot on long trips so that they can sleep or relax in other ways.