banner WaveTop WWWDirectory
May 14, 1997

Mind & Machine

If Deep Blue Wrote Hamlet
Would It Change the Endgame?

Garry Kasparov's defeat this weekend at the virtual hands of IBM's Deep Blue computer marked another milestone in the teetering relationship between humans and machines.

Related Articles
Flawed Strategy vs. Fierce Resistance
(May 13, 1997)

IBM Chess Machine Beats Humanity's Champ
(May 12, 1997)

Deep Blue and Kasparov Face Final Game
(May 11, 1997)

Saturday Is D-Day for Kasparov
(May 10, 1997)

Victory Eludes Champion After a Tense Fight With the Computer
(May 8, 1997)

Kasparov Draws Third Game Against Deep Blue
(May 7, 1997)

Computer Defeats Kasparov, Stunning the Chess Experts
(May 5, 1997)

Kasparov Beats Deep Blue in First Game of Rematch
(May 4, 1997)

Smarter Than Us? Who's Us?
(May 4, 1997)

For decades, computer scientists, chess players and other humans have debated whether a machine could ever beat a human at a game that has often been seen as a crystalline test of intelligence.

In matters of physical strength, we have long ago been surpassed by machines. But no one really cares if a human loses the Boston Marathon to a Porsche Carrera. It is only in matters of the mind that we are loathe to concede any evidence against our uniqueness.

Now that Deep Blue has defeated one of the best chess players of all time, have we become just stupid second-rate entities in the cosmos? Well, we can comfort ourselves with the knowledge that we can still beat monkeys and insects at the game.

I have to admit that after Kasparov's resignation on move 19 of the final game on Sunday, I did feel a twinge of IQ loss and an increase in hairness. But a quick shave and a recital of couple pages from Paradise Lost seemed to restore my humanity -- I think.

Actually, I have made a quick recovery. Part of the reason is that I have been losing to chess computers since the earliest versions of Chessmaster. I haven't even bothered to buy a new chess program since the mid-1980s. After all, there is no point in taking on the hottest, most powerful version when even the lamest versions whup me.

But after watching the match for a while, I began to feel that little about intelligence was really being settled with this grand confrontation. In many ways, we have way overestimated chess. Far more mundane activities, like navigating a crowded sidewalk or carrying on a meaningful conversation, are better benchmarks of intelligence. In these complex areas of judgment and constantly changing parameters, computers are not even in the same ballpark as 2 year olds. I can have a better conversation with a pet than I can with a computer.

The computer made up for its deficiencies with an opening book about the size of Rhode Island and enough processors to search 200 million moves a second. Deep Blue also surprised several masters with its insightful play, although it apparently committed some blunders as well.

In some ways, Deep Blue wasn't playing chess at all. What it was doing was analogous to trying to recreate Hamlet by typing, in random order, every word Shakespeare ever used. Admittedly, there was also something else at play, since Deep Blue had to exercise judgment in picking the best sounding words. But still, the game Deep Blue was playing was not really the same as the one Kasparov was playing.

Chess is challenging to humans precisely because our memories are poor, our abilities to calculate many moves ahead are low, and our minds are subject to all sorts of psychological distractions that can turn a Kasparov into a puddle of tears. For a computer, large chunks of the game -- the opening and the endgame -- are simply exercises in crunching and database searching.

It is only in matters of the mind that we are loathe to concede any evidence against our uniqueness.

While the match leaves the questions of machine and human intelligence unresolved, it does say something about how we have constructed the dynamic of modern anxiety.

After the sixth and final game, Kasparov said, "I was not in the mood of playing at all." He later explained: "I am a human being. When I see something that is well beyond my understanding, I'm afraid."

Kasparov believed that he faced a Zugzwang, the chess term for a forced move that is usually unpleasant. Out of his awe for the machine's calculating abilities and its awesome library of moves, Kasparov felt compelled to play a different strategy. Who knows how he would have played had he thought he was matched against a human like me -- an unranked lamer who still tries the exchange variation of the Ruy Lopez?

There are other examples of how humans have been forced to "play the machine" so to speak. Consider automated stock trading. Stock is a measure of a company's worth and the expectations of its future growth. But programmed trading can be triggered by all sort of unrelated factors. They can be triggered by other computers that, for whatever reasons, have decided to sell or buy, setting off a cascade of virtual panic that has no greater causation than a few lines of computer code in some program.

Who hasn't finessed a tax return so that the numbers will not trigger any computerized red flags, even going so far as to pay more just to avoid notice?

The machine has become a significant component of society, injecting speed, precision and coolness into ventures in which those attributes were not considered before. As Kasparov noted, the calculations are so complex now that they have become almost organically mysterious.

Kasparov is certain to be given another chance at Deep Blue -- if only so IBM can prove that its victory was no fluke. No doubt there will be many more human-computer matches, which could eventually lead to a different style of chess influenced by both people and machines. At least with more games humans might learn to relax a little better in the face of machinery, thus enhancing their chances of kicking some virtual behind.

Regardless of the possibilities, I doubt that even a dozen such chess matches will do much more to resolve the question of machine intelligence. But as for the machine becoming a player -- of that, there is certainly no question.

WaveTop WWWDirectory

MIND & MACHINE is published weekly, on Wednesdays. Click here for a list of links to other columns in the series.

Ashley Dunn at welcomes your comments and suggestions.

Home | Sections | Contents | Search | Forums | Help

Copyright 1997 The New York Times Company