bk99.de entertain the web since 1997

bk99.de
IBM · AICurated

Deep Blue beats Kasparov: computing power against a world champion

Summary

On 11 May 1997 world chess champion Garry Kasparov lost a six-game match in New York to IBM’s computer Deep Blue, which evaluated 200 million positions per second with 32 processors. It was the first time a reigning world champion lost to a computer under regular tournament conditions. IBM tells the story as a breakthrough for supercomputers and artificial intelligence; technically, Deep Blue was above all a triumph of specialised hardware and raw computing power.

From the first chess program to the world champion

IBM mathematician Alex Bernstein wrote the first complete chess program in 1957, on an IBM 704 running 42,000 instructions per second. From 1985 Feng-hsiung Hsu developed the chess machine ChipTest at Carnegie Mellon University; in 1989 IBM brought Hsu and Murray Campbell into its research division, and the project was renamed Deep Blue, a nod to the nickname Big Blue.

In February 1996 Deep Blue won the first game against Kasparov in Philadelphia but lost the match. For the rematch in May 1997 the team improved the endgame databases and the evaluation function and hired grandmasters as advisers. Kasparov won the first game, Deep Blue the second, three were drawn, and Deep Blue won the sixth game.

Ideas

  • Raw computing power can solve tasks once considered the domain of human intuition.
  • Specialised hardware beats general-purpose computers on narrowly defined tasks.
  • A public match makes technology understandable for everyone.
  • Endgame databases and the evaluation function mattered more than new theory.

Insights

  • Deep Blue did not learn but searched: the victory shows how far targeted search with lots of computing power goes, not how machines think. Learning systems such as AlphaZero took a different path two decades later.
  • The match was also marketing: IBM retired Deep Blue afterwards, and the technology flowed into supercomputers such as Blue Gene.

Facts

  • Deep Blue used 32 processors and evaluated 200 million positions per second.
  • Its computing power was 11.38 billion floating-point operations per second.
  • The 1997 match ended 3.5 to 2.5 in Deep Blue’s favour.
  • After the victory Deep Blue went to the Smithsonian Institution in Washington.

References

Critique

  • The IBM page is a success story written by IBM itself: it does not mention that Kasparov suspected human intervention after the match and that IBM declined a further rematch.

Remarks

  • According to IBM, Kasparov said after the match: “The computer is far stronger than anybody expected.”

Recommendations

  • To compare with today’s game AI, read about AlphaZero, which learned chess without human knowledge.
  • With computing power, always ask what it is optimised for: Deep Blue could only play chess.

Read the original

Search the Web Archive

Own project · 2025

Honest Bad Guy: My system prompt against people-pleasing AI

At the end of 2025 I gave my AI assistants a fixed system prompt because their constant agreement without real criticism annoyed me. It demands checking instead of guessing, contradiction instead of sugarcoating and no claimed tests that never happened. Here it is in full to copy, together with the reasons behind the rules and the places where it gets in its own way.

Black Forest Labs · 2026

FLUX 3 Image Klein: what lies behind “5.6B parameters” and “8 GB”

With FLUX 3 Image Klein, Black Forest Labs has released an open image model with 5.6 billion parameters that generates images, edits them with up to ten reference images and places image elements via bounding boxes. According to the vendor, it runs with INT8 on graphics cards with as little as 8 GB and beats larger open models in internal comparisons. This post explains the technology and shows what the mandatory separate prompt upsampler, the memory figures and the first community experiences mean for running it on your own hardware.

Own project · 2026

An AI book for my birthday: cordless drill instead of screwdriver

For my birthday I was given a book designed by an AI, with my photo on the cover and the title “Die Matrix ruft – Peter antwortet später” (“The Matrix is calling – Peter will answer later”). You might think an AI gift is impersonal, but the personal part lies in the idea, the photo and the punchline, not in setting the letters. This post explains why I see AI as a cordless drill and where the comparison still falls short.