On the afternoon of May 11, 1997, in a television studio on the 34th floor of a building in midtown Manhattan, one of the strongest chess players in history sat down to play the sixth and final game of a match he had expected to win. Garry Kasparov was 34 and had held the world championship for 12 years, making his name a synonym for the human mind operating at the absolute edge of its power. Across the board from him sat an opponent with no mind at all.

The opponent was a black tower of processors built by IBM and named Deep Blue. It did not sit, because it could not sit — it was represented at the board by a quiet engineer who moved the pieces on its behalf. It calculated roughly 200 million positions per second and could not grow tired across a long session. It did not feel the heat of the lights, the weight of the audience, or the knowledge that a defeat represented something far larger than a single lost game. It felt nothing, because feeling was not among its capacities. It simply searched, evaluated, and chose.

The game lasted barely more than an hour. Kasparov, playing as black and opting for a traditional Caro-Kann defense, made an early error that strong players had warned against for decades. The machine answered with a sharp sacrifice of a knight, a move stored in its opening library, and the human champion found himself in a position no person enjoys defending, and no computer of that era knew how to lose. After just 19 moves, Kasparov resigned. It was the first time in his professional life that he had ever surrendered a game in such a way, and with it he became the first reigning world champion in history to lose a match to a machine under normal tournament conditions. He left the stage angry and shaken, while many in the room understood that they had watched a hinge turn in the long story of human intelligence and artificial intelligence.