In the article I wrote for TIME last year after my victorious match against IBM's Deep Blue supercomputer in Philadelphia, I expressed my surprise and amazement at seeing a new kind of intelligence. I referred to Game 1, in which the computer's decision to sacrifice a pawn, based strictly on the machine's calculations, coincided with what a human would have done using human logic. Thus I stepped into a discussion of whether artificial intelligence has to be an exact copy of human thinking procedures or whether we should judge intelligence by the end result. I viewed the match with an improved version of Deep Blue as an opportunity to study this further--and of course to win a competitive event.
Unfortunately, I based my preparation for this match, played two weeks ago in New York City, on the conventional wisdom of what would constitute good anticomputer strategy. Conventional wisdom is--or was until the end of this match--to avoid early confrontations, play a slow game, try to out-maneuver the machine, force positional mistakes, and then, when the climax comes, not lose your concentration and not make any tactical mistakes.
It was my bad luck that this strategy worked perfectly in Game 1--but never again for the rest of the match. By the middle of the match, I found myself unprepared for what turned out to be a totally new kind of intellectual challenge.
The decisive game of the match was Game 2, which left a scar in my memory and prevented me from achieving my usual total concentration in the following games. In Deep Blue's Game 2 we saw something that went well beyond our wildest expectations of how well a computer would be able to foresee the long-term positional consequences of its decisions. The machine refused to move to a position that had a decisive short-term advantage--showing a very human sense of danger. I think this moment could mark a revolution in computer science that could earn IBM and the Deep Blue team a Nobel Prize. Even today, weeks later, no other chess-playing program in the world has been able to evaluate correctly the consequences of Deep Blue's position.
Also, Game 2 had a very unfortunate finish. Deep Blue held a strategically winning position, but it made a tactical blunder that, if I had sacrificed a piece, could have given me a miraculous escape. But I trusted the machine's calculations, thinking it would not miss such a continuation, and resigned instead.
Game 2 created an enigma for me that I never solved and from which I never recovered. I would like the IBM team to start disclosing the secrets of how they achieved this unthinkable success in chess programming. They claim they developed software that enabled them to change the style of the program in mid-match and the evaluation ability of the machine from game to game. This also is revolutionary, because any change, any tweak in the computer normally needs weeks of testing to avoid potential bugs.
I discovered that I was playing a very flexible, quickly changing opponent with an ability to avoid any mistakes in long-term calculations. My opponent was psychologically stable, undisturbed and unconcerned about anything going on around it, and it made almost none of the typical computer-chess errors.