DEEP BLUE FUNK

AS WAS TO BE EXPECTED, THE END OF civilization as we know it was announced on the back pages. On Feb. 10, 1996, in Philadelphia, while America was distracted by the rise of Pat Buchanan, the fall of Phil Gramm and other trifles, something large happened. German philosophers call such large events world-historical. This was larger. It was species-defining. The New York Times carried it on page A32.

On Feb. 10, Garry Kasparov, the best chess player in the world and quite possibly the best chess player who ever lived, sat down across a chessboard from a machine, an IBM computer called Deep Blue, and lost.

True, it was only one game. What kind of achievement is that? Well--as Henny Youngman used to say when asked, "How's your wife?"--compared to what? Compared to human challengers for the world championship? Just five months ago, the same Kasparov played a championship match against the next best player of the human species. The No. 2 human played Kasparov 18 games, and won one. Deep Blue played Kasparov and won its very first game. And it was no fluke. Over the first four games, the machine played Kasparov dead even--one win, one loss, two draws--before the champ rallied and came away with the final two games.

Kasparov won the match. That was expected. Game 1, however, was not supposed to happen. True, Kasparov had lost to machines in speed games and other lesser tests. And lesser grand masters have lost regulation games to machines. But never before in a real game played under championship conditions had a machine beaten the best living human.

Indeed Kasparov, who a few weeks ago single-handedly took on the entire national chess team of Brazil, was so confident of winning that he rejected the offer that the $500,000 purse be split 60-40 between winner and loser. Kasparov insisted on winner-take-all. They settled on 80-20. (What, by the way, does Deep Blue do with its 100-grand purse? New chips?)

Asked when he thought a computer would beat the best human, Kasparov had said 2010 or maybe never. A mutual friend tells me Garry would have gladly offered 1-10, perhaps even 1-100 odds on himself. That was all before Game 1. After Game 1, Kasparov was not offering any odds at all. "He was devastated," said his computer coach, Frederick Friedel. "It was a shattering experience. We didn't know what the game meant. There was the theoretical possibility that the computer would be invincible, and that he would lose all six games."

True, we have already created machines that can run faster, lift better, see farther than we can. But cars, cranes and telescopes shame only our limbs and our senses, not our essence. Thinking is our specialty, or so we think. How could a device capable of nothing more than calculation (of the possible moves) and scoring (of the relative strengths of the resulting positions) possibly beat a human with a lifetime of experience, instant pattern recognition, unfathomable intuition and a killer instinct?

How? With sheer brute force: calculation and evaluation at cosmic speeds. At the height of the game, Deep Blue was seeing about 200 million positions every second. You and I can see one; Kasparov, two. Maybe three. But 200 million? It was style vs. power, and power won. It was like watching Muhammad Ali, floating and stinging, try to box a steamroller in a very small ring. The results aren't pretty.

What is Deep Blue's secret? Grand master Yasser Seirawan put it most succinctly: "The machine has no fear." He did not just mean the obvious, that silicon cannot quake. He meant something deeper: because of its fantastic capacity to see all possible combinations some distance into the future, the machine, once it determines that its own position is safe, can take the kind of attacking chances no human would. The omniscient have no fear.

In Game 1, Blue took what grand master Robert Byrne called "crazy chances." On-site expert commentators labeled one move "insane." It wasn't. It was exactly right.

Here's what happened. Late in the game, Blue's king was under savage attack by Kasparov. Any human player under such assault by a world champion would be staring at his own king trying to figure out how to get away. Instead, Blue ignored the threat and quite nonchalantly went hunting for lowly pawns at the other end of the board. In fact, at the point of maximum peril, Blue expended two moves--many have died giving Kasparov even one--to snap one pawn. It was as if, at Gettysburg, General Meade had sent his soldiers out for a bit of apple picking moments before Pickett's charge because he had calculated that they could get back to their positions with a half-second to spare.

In humans, that is called sangfroid. And if you don't have any sang, you can be very froid. But then again if Meade had known absolutely--by calculating the precise trajectories of all the bullets and all the bayonets and all the cannons in Pickett's division--the time of arrival of the enemy, he could indeed, without fear, have ordered his men to pick apples.

Which is exactly what Deep Blue did. It had calculated every possible combination of Kasparov's available moves and determined with absolute certainty that it could return from its pawn-picking expedition and destroy Kasparov exactly one move before Kasparov could destroy it. Which it did.

It takes more than nerves of steel to do that. It takes a silicon brain. No human can achieve absolute certainty because no human can be sure to have seen everything. Deep Blue can.

Now, it cannot see everything forever--just everything within its horizon, which for Deep Blue means everything that can happen within the next 10 or 15 moves or so. The very best human player could still beat it (as Kasparov did subsequently) because he can intuit--God knows how--what the general shape of things will be 20 moves from now.

BUT IT IS ONLY A MATTER OF TIME before, having acquired yet more sheer computing power, Blue will see farther than Garry can feel. And then it's curtains. Then he gets shut out, 6-0. Then we don't even bother to play the brutes again. The world championship will consist of one box playing another. Men stopped foot racing against automobiles long ago.

Blue's omniscience will make it omnipotent. It can play--fight--with the abandon of an immortal. Wilhelm Steinitz, a world chess champion who was even more eccentric than most, once claimed to have played chess with God, given him an extra pawn and won. Fine. But how would he have done against Blue?

Kasparov himself said that with Deep Blue, quantity had become quality. When you can calculate so fast and so far, you rise to another level. At some point that happened in biology. There are neurons firing in lemmings and squid, but put them together in gigantic enough numbers and fantastic enough array, as in humans--and, behold, a thought, popping up like a cartoon bubble from the brain.

We call that consciousness. Deep Blue is not there yet, though one day we will be tempted to ask him. In the meantime he's done alchemy, turning quantity into quality. That in itself is scary.

At least to me. However, my son, age 10, who lives as comfortably with his computer as he does with his dog, was rooting for Deep Blue all along. "What are you worried about, Dad?" he says. "After all, we made the machine, didn't we? So we're just beating ourselves."

The next generation has gone over to Them. No need to wait for the rematch. It's over.