Chap claims Atari 2600 “absolutely wrecked” ChatGPT in chess

In 1977, the Atari 2600 console was released with a 1.19MhZ eight-bit processor and 128 bytes RAM. But that’s enough power to beat ChatGPT in chess.

Infrastructure architect Robert Caruso posted results of his experiment to “pit ChatGPT (via Stella emulator), against the Atari 2600 chess engine and see what happened.”

ChatGPT mistook rooks as bishops and lost track of pieces repeatedly

Caruso conducted the experiment after talking with ChatGPT about chess history. The bot offered to play against Atari at some point during the chat. This was a reasonable request, as “Video Chess”one of Atari’s console games, was a game that Atari had commissioned.

Online chess experts discussing the merits Video Chess suggests it may have played on a level that beginners might have found challenging and gave regular recreational players with intermediate skill a bit to worry about. Caruso thought that his experiment would be a “lighthearted stroll down retro memories lane.”

But instead, he watched the Atari humiliate ChatGPT. “ChatGPT was absolutely destroyed on the beginner’s level,” he wrote. Christmas 1984: the last hurrah of 8-bit home computer

  • Weeks with a BBC Micro
  • It’s good enough to fix a mainframe.
  • Building the ZX Spectrum 128 with the New ROM Antics
  • Caruso claimed that ChatGPT blamed Atari chess icons for being “too abstract to recognize”. Even after switching to standard chess, the chatbot still “made enough mistakes to get laughed out of an 3rd grade club.”

    I had to stop [ChatGPT] it from making awful moves, and correct its board knowledge multiple times per turn, he wrote. The chatbot “kept saying it would improve if we just started again.”

    Finally, the bot conceded.

    Caruso asked “Have You Played Atari Today?” invoking Atari’s advertising slogan.

    He concluded, “ChatGPT wished it hadn’t,”

    The Register is aware that many readers are retro-tech enthusiasts and use retro-tech devices. We challenge you to fight AI with them and let us know what happens. Send us the results here

    (r)

    www.aiobserver.co

    More from this stream

    Recomended