Deep Blue Beats Kasparov, 1997

In May 1997, in a windowless studio in New York, IBM's Deep Blue defeated reigning world chess champion Garry Kasparov 3.5 to 2.5 in a six-game match. It was the first time a computer had beaten a world champion in a classical match under tournament conditions.

Kasparov was not gracious. He accused IBM of human interference, demanded a rematch IBM declined, and stormed out of the press conference. For many observers, the moment still felt like a turning point for machines.

Why it mattered beyond chess

• Chess had long been the drosophila of AI research, the favorite test organism
• The match made front pages worldwide and entered popular culture
• It showed that specialized systems plus enough compute could beat the best human in narrow domains
• It did not, as skeptics noted, bring anything like general intelligence

Deep Blue retired right after the match. IBM had won the PR battle and the research exploded in other directions. Two decades later, DeepMind's AlphaZero would play chess at a higher level than Deep Blue, having taught itself from nothing in hours.

“I lost my fighting spirit.”

The big idea: Deep Blue proved that AI could win at a prestige human skill. Chess fell. But fall is not the same as understand.