Kasparov vs. Deep Blue, the Rematch: And other ramblings on Artificial Intelligence Todd Murchison ski@sover.net Chess, Kasparov, Deep Blue, Kasparov vs. Deep Blue, Todd, todd, Todd Murchison, Less than a year ago in an article I discussed the state of artificial intelligence in the wake of World Chess Champion Garry Kasparov defeating the IBM computer Deep Blue in a formal chess tournament. Things change fast in the world of computers, and last year Deep Blue turned the tables on Kasparov and defeated him in a six-game rematch. Though world champions have lost individual games to computers before, this was the first time a current world champion has lost a match to a computer under tournament conditions. Garry Kasparov is also generally considered to be the best chess player in history. Although perhaps it's now going to be increasingly necessary to specify that Kasparov is the best human chess player in history. Kasparov and his dedicated fans have a litany of excuses for why he lost the match. I don't follow competitive chess closely enough to know if chess masters traditionally give excuses when loosing a match, but I suspect the excuses are not as ferocious when the match is human vs. human. Emotions ran very high surrounding the tournament blue competition. "Newsweek" ran an article last month which called the match "the brain's last stand". Seems like typical melodramatic journalism, the human brain is in no danger. In fact I believe that though Kasparov lost, it was the brain's finest moment. We are talking about a 1.8 ton machine that can calculate 200,000,000 positions per second! Kasparov has estimated that he can look at around 3 positions per second, but somehow his experience and intuition allows him to look at just the "right" moves. The very fact that Kasparov was able to give the computer such a challenge is remarkable, it is evidence of our intelligence. The computer that beat him is also a direct result of human intelligence and inspiration, and there are computers far more powerful than this one. Deep Blue is the most powerful chess specific computer in the world, however there are computers used in research that are a hundred-fold faster. Its probable that Kasparov and Deep Blue, or its successors, will still have a few good battles. These matches will become increasingly one sided though, and eventually as the technology trickles down - even a little handheld chess game will be able to best all the human masters. But so what? Even a little pocket calculator can out run the greatest mathematicians in the world at certain tasks, it's just a tool. Its funny that we feel so threatened that computers might be able to "outthink" us at some tasks. After all, do you feel threatened that cars can "outrun" you? That a megaphone can "out-yell" you? These are just other machines that we have invented to help us accomplish certain tasks. They are extensions of us, and amplifiers of our own abilities. They are neither good nor evil in themselves, we determine if machines are used to help or harm. Of course all bets are off if computers ever achieve self-awareness, and become truly independent thinkers. While this is a distinct possibility, it is also a very long way off. Though computers may soon be able to simulate intelligence so closely that few people can tell the difference. The benefit of this is that we will be able to communicate with our machines in natural language and not have to spend hours poring through manuals to learn to accomplish simple tasks. My wife expressed that her concern is not so much that computers will get "smarter" than us, but rather that they will take over all the functions that people perform now, putting more people out of work. This is a dilemma that has faced us for some time now, with machines having replaced millions of workers in the factories, and computers replacing whole floors of accountants in companies across the world. The original dream of technologists was that our machines would free us all from having to work at all. They hoped that we could all spend our lives in endless vacation as the machines toiled away at the tasks that keep the world running. This of course if far from what has really happened. You can build a robot that eliminates 500 jobs in various factories, but somewhere there are 500 people involved in designing, building, selling and maintaining that robot. Jobs have not been eliminated, they have simply been shifted around. There is a chance that computers will eventually get to the point where they are self-replicating, self-designing and self-maintaining, and if that happens most jobs could indeed be eliminated. Obviously the economic structures of the world would be forced to change radically at that point; neither capitalism, socialism or any of the models we humans have tried so far would be likely to work in their present form in such changed world. We can cross that bridge should we ever come to it. One thing is certain, we have a wild ride ahead of us. Deep Blues "intelligence" more than doubled from last year to this year and this trend shows no signs of subsiding. In fact the rate at which our computers double in speed is continuing to accelerate. We humans are darn smart creatures, but we are certainly not evolving at such a fast rate. In light of the recent international scholastic competitions and IQ tests it could indeed be argued that Americans at least are actually getting less intelligent even as our machines plunge forward! Todd Murchison is a professional skier, computer tech, and dreamer. His computer out spells him, out calculates him, beats him at chess, and is even better looking. "But just put that sucker on skis", he snarls, "and then we'll see who the boss is". Sad isn't it?