Humans Are On The Verge Of Losing One Of Their Last Big Advantages Over Computers
REUTERS/Kent SkibstadIt's been 16 years since Deep Blue first beat Gary Kasparov at chess representing a major breakthrough in terms of the ability of computers to surpass to outperform humans at certain tasks.
But now computers are beating humans in another crucial way according to a post by Tyler Cowen at his blog, Marginal Revolution.
See, it used to be that a human aided by a computer could still beat a computer by itself.
Increasingly that's no longer so.
Computers are getting so fast and good that humans still lose out despite the significant advantages a computer/human team ("centaur").
Basically as of 2008, a human plus a computer was generally better than a computer alone over a 90 minute chess game. That was because humans can do things like make sure the computer uses its time and processing power more efficiently by ruling out sets of moves, compare different programs and see where they disagree, and anticipate a computer program's opening moves and lay a trap.
That meant that a brilliant chess player paired with a computer could beat the best programs - until now.
According to a Rybka computer chess forum, computers may have recently gotten to the point where they're so good that those advantages don't matter. They work so rapidly that humans really can't add value. You have to seriously slow down the game for a centaur to compare programs at a deep enough level that they can add anything. Now a centaur might be better during a correspondence game over days, but not a regular ones.If it isn't already fair to say computers beat out freestyle chess players, it won't be long. Computer chess expert and University of Buffalo computer science professor Kenneth Regan found that many of the best chess games of all time have been played by centaurs, but they're being surpassed.
Beyond chess, that has huge implications. We've made an assumption that while computers will take over just about anything that can be automated, humans will be able to add value through talent, intuition, and the ability to correct a computer's mistakes.
What's left for even brilliant people when computers don't benefit much or at all from that help? The set of occupations where humans are worthwhile looks like it might be shrinking.
We need to reevaluate both how we think about chess and the economy of the future.
As Cowen puts it, "what are humans still good for?"