Deep Blue
/ deep BLOO /
Deep Blue was the IBM chess computer that, in 1997, became the first machine to beat a reigning world chess champion in a full match, defeating Garry Kasparov. It was a landmark moment — proof that a machine could outplay the very best human at a game long treated as a pinnacle of human intellect — and it captured the public imagination like few AI events before it.
But how it won is the instructive part. Deep Blue was not a learning system and had no general intelligence; it was a triumph of brute-force search married to expert chess knowledge. Special-purpose hardware let it examine on the order of hundreds of millions of board positions per second, looking many moves ahead, while a hand-crafted evaluation function — built with help from human chess masters — scored how good each resulting position was. In essence it out-calculated Kasparov rather than out-thought him in any human sense. It did one narrow thing, chess, with overwhelming power, and could do nothing else.
Deep Blue's significance is twofold and honest. Symbolically, it marked the moment machines surpassed humans at a flagship intellectual task and showed how far raw search plus domain knowledge could go. Conceptually, it underlined a lesson about AI: superhuman performance in one domain need not look anything like human thinking, and it tells us little about general intelligence. Chess fell to calculation; the much harder, more 'human' game of Go would hold out until a very different approach — learning — arrived two decades later with AlphaGo.
Where a human grandmaster relies on intuition and judges a handful of candidate moves deeply, Deep Blue did the opposite: it crunched hundreds of millions of positions per second and picked the move with the best calculated score. Two utterly different routes to strong chess — and only one of them resembles thinking.
Deep Blue won by brute-force calculation plus expert-tuned evaluation — not by anything resembling human intuition.
Deep Blue did not learn and had no general intelligence — it searched and evaluated chess positions faster than any human could. Its win is the classic illustration that beating people at one task can come from sheer calculation, not human-like thought.