Deep Blue

Deep Blue's 1997 victory over Garry Kasparov marked the first defeat of a reigning world chess champion by a computer under tournament conditions. This landmark event was seen as symbolically significant, demonstrating that artificial intelligence was catching up to human intelligence and could defeat one of humanity's great intellectual champions.

The match was widely covered by media worldwide, and Deep Blue became a celebrity. It was compared to other historic moments in the relationship between humans and technology. The first match occurred February 10-17, 1996 in Philadelphia, which Kasparov won 4-2. The historic rematch took place May 3-11, 1997 in New York City, where Deep Blue won 3½-2½. The match sparked widespread debate about the nature of intelligence and the future of human-computer interaction.

Deep Blue represented a massive computational achievement for its time. It had processing power of 11.38 billion floating-point operations per second (GFLOPS), could evaluate 200 million chess positions per second, used 256 processors (some sources cite 480 custom processors) working in tandem, and was a customized IBM RS/6000 SP supercomputer with special chess accelerator chips. It ranked as the 259th most powerful supercomputer globally in June 1997 (TOP500 list). The system had an opening book database of 700,000 previous grandmaster games and used brute-force tree search with specialized chess algorithms.

The match generated massive public interest and media coverage worldwide. Prize money for the first match (1996) was winner $400,000 and loser $100,000. For the second match (1997), winner received $700,000 and loser $400,000, plus an additional $100,000 to the Deep Blue team from the Edward Fredkin Prize. Game 1 of the 1996 match was the first-ever computer victory over a reigning world champion under classical time controls. Game 2 of 1997 featured Deep Blue's move 37 (Be4), which was so unexpected it led Kasparov to believe humans were intervening, psychologically affecting his subsequent play.

Kasparov accused IBM of cheating, suggesting human chess masters were intervening during games. IBM refused a rematch and immediately dismantled Deep Blue after the 1997 victory (though it actually remained operational for several years). Later analysis suggested Kasparov played uncharacteristically poorly in the 1997 match. The match raised questions about fairness: IBM upgraded Deep Blue between matches while Kasparov remained human. In a December 2016 podcast, Kasparov reflected on the match after writing a book about it, acknowledging increased respect for the Deep Blue team but decreased opinion of both his own and Deep Blue's performance, recognizing that modern chess engines easily surpass Deep Blue.

After Deep Blue's victory, Go became the canonical example of a game where humans still outmatched machines (until AlphaGo in 2016). By today's standards, Deep Blue's computing power is modest—an iPhone 7 (2016) achieved 12.6-84.7 GFLOPS, surpassing Deep Blue's 11.38 GFLOPS. Modern chess engines like Stockfish running on consumer hardware far exceed Deep Blue's playing strength. The match demonstrated that specialized hardware and brute-force computation could solve complex strategic problems, sparked public interest in AI capabilities, highlighted limitations of pure computational approaches (later systems would incorporate learning and neural networks), and established chess as a solved problem for AI, shifting focus to more complex games like Go.