“You probably do. Newsweek called it ‘The Brain’s Last Stand.’ ” He laughed ruefully. Then he started to tell me about it.
The thing about the loss, he said, was this. If there had ever been a point to chess — and Aleksandr would be the first to admit that there might not be any point to chess — it was conclusively defeated by the revelation that all chess problems of the world could be unscrambled unconsciously by robot neurons firing into the void. Great chess was no longer the elegant accomplishment of the human mind; the true accomplishment was the ability to create something bigger and better than oneself and to then stand back, amazed. Humans should retire or else find more modest modes of occupation. Everybody knew this. Even the jokes afterward — at the bars, on the news, on the Internet — reflected this knowledge. “In a related story,” one of the talk-show hosts had said, “the New York Mets were beaten by a microwave oven.”
The worst part was the speed with which the program played — Aleksandr’s moves were instantaneously matched and outsmarted by the computer, without the hemming and sweating and doubting that made any brilliant human move feel as though it could have been otherwise. The computer moved with a clinical ruthlessness, and it made Aleksandr understand with a sickening certainty that there was nothing he could think of that the computer hadn’t thought of first. It worked with the efficiency of a guillotine.
The man who played for the computer was soft-looking, chubby-cheeked, his hands like chicken cutlets, his leporidian face innocent and wide. He made a little gesture with each move, a nearly imperceptible half-shrug (Aleksandr was never sure whether or not the cameras had recorded it) as if trying to disown it — not me, he seemed to say, not me who’s doing this to you, who’s humiliating you, who’s unraveling the human brain. I’m just the conduit here, the messenger, the mechanism. I am, humbly, just the pawn.
In the end, it took a paltry nineteen moves — the shortest loss of Aleksandr’s career. He’d opened with the Caro-Kann Defense — not his usual against human opponents, but for a little while, things were under control: he met the computer’s advancing duo of pawns with his own staggered pair, and a brief frenzy of exchange commenced. Next came the ritualistic introduction of the knights. He’d broken his own rule — don’t move the same piece twice in the opening — but the beginning was conventional and promising enough. The computer advanced its knight farther, and Aleksandr introduced the second of his. The three knights assembled in a crooked-elbow single file. The computer advanced its bishop. Aleksandr advanced his pawn to e6, bringing it to the flank of his farthermost knight. The computer roused its second knight in response. Then Aleksandr flicked his pawn forward to h6, and as soon as he lifted his finger, he knew. The avid watchers knew. He’d moved it too early in the sequence — he should have introduced his bishop, then awaited the grand entrance of the computer’s queen, and only then brought his pawn to h6 to menace the computer’s closest knight. That knight would have retreated to the center of the board only to be followed by Aleksandr’s. The h6 move in response to the knight was a mistake. It was a mistake, but it wasn’t a mistake of strategy — it wasn’t a misjudgment, an incorrect forecast into the future. It was a mistake of memory, of basic competence — like losing your car keys, like dropping a dish.
The computer’s knight took another pawn, at e6, and crouched breathing down the neck of Aleksandr’s king. Should he have taken the knight immediately? Maybe. Later, many, many people — mostly anonymous, mostly on the Internet, mostly people who’d had a decade in their pajamas to think about it — would say that he should have gone straight in then. But he hadn’t. He’d wanted to give his king another square to maneuver. He’d allowed the common knight sacrifice, which was not a reflection of the computer’s fantastic strategy, in particular; that sacrifice was a dull, almost juvenile move, well known to theory. He’d used it himself against Rusayev in one of their fifty-three games — back when he’d been the new astonishment, the brilliance at which everyone had marveled.
At this point, the newspapers said later, there’d been a look of “terror” on Aleksandr’s face that obtained for the rest of the game.
So he’d lost the ability to castle, which the computer then did — quietly, brutally, without comment, the soft man’s brow remaining smooth and dry.
Aleksandr took the knight, as he had to, and the bishop sailed into the space that his pawn had prematurely abandoned and put him in check. He jockeyed his king to the right; he had nowhere else to go. It was only the tenth move of the game.
The computer’s second bishop crept halfway down the board and sat there waiting. Aleksandr halfheartedly menaced it with his knight, and it temporarily retreated by one square.
There was another exchange of pawns, this one dirtier and more desperate. Aleksandr’s neck was soaking wet, and he instinctively looked around the room for an exit. Across the table, the fat man looked calm, his cheeks alternately swollen and slack with the movement of his self-satisfied breaths. This man — who was he? Had he helped to build the computer? Had he studied chess theory and computer code for years, learning how to translate the one into the other, hoping to create an entity that could extrapolate and infer? Probably not. Probably he was a nothing, a person who knew how to push a button or two. Aleksandr thought bitterly that he wasn’t only a traitor to chess, as some of the Internet critics had said. He was a traitor to people.
Aleksandr had closed his eyes and sacrificed his queen to take a bishop and a rook. He took the fat man’s bishop greedily, as a kind of petty, interim revenge. This was a frantic move: he could feel himself falling down a well; he could hear the scrape of fingernails against concrete. Everybody could. The fat man coughed. The crowd murmured, looked away.
And then he’d resigned. He might not have been smart enough to beat a computer, but he was smart enough to know when he was beaten by a computer. He wasn’t going to submit to a humiliating inevitability; he wasn’t going to let himself be chased into ever more hopeless cover as the entire world watched. He stood up. He walked out. He did not shake the fat man’s hand.
Afterward, people kept asking him about the pawn — the h6 move, a beat too early. He’d had to tell them he didn’t know, he didn’t know; it was a mistake, and he didn’t know where it had come from or why. On the Internet, conspiracy theorists wondered whether he’d thrown the match intentionally, so that he might one day demand a rematch, so that he might one day win more money. But that wasn’t true. Maybe it wasn’t true, either, that a computer couldn’t be beaten. Maybe it wasn’t true that a computer’s brilliance exceeded all human imagining. Maybe it was just that Aleksandr was forty. Maybe it was just that Aleksandr was tired.
Now the computer sat in the Smithsonian in Washington, D.C., and every day it played reenactments of that final game for public viewing, automatically and on repeat.
Aleksandr told me this, and we were silent. It was the kind of confession that makes you so uncomfortable that the only possible response is to offer one of your own.
“Well,” I said. “I have a disease that’s going to make me lose my mind.”
Aleksandr raised his eyebrows. “What?” Behind his voice, there was a faint hint of laughter. People’s response to outlandish information is often to laugh.
“It’s called Huntington’s,” I said. “It’s what my father died of. They can test you for it. It’s motor functioning, first, actually, then cognitive functioning. Cortex on down.”
Читать дальше