The late Hungarian mathematician Paul Erdős thought he had the last word on a geometry problem. Now an OpenAI chatbot has proved him wrong.
  • SailorFuzz@lemmy.worldEnglish
    351·
    16 days ago

    See, this is the kind of thing that I thought AI would be used for.

    But now, after all the generative bullshit, and theft and hallucinations and antisocial reinforcement and environmental havok and techbro douchebaggery… I just dont care.

    • chocrates@piefed.worldEnglish
      141·
      16 days ago

      Right!? Ai should be good at iterating a ton on existing data that researchers can pull novel ideas out of. Not do everything badly

    • givesomefucks@lemmy.worldEnglish
      3·
      16 days ago

      Because slime molds are better at this than humans…

      AI can beat humans at some super specific stuff, but that doesn’t mean that in general it’s better than humans.

      Like, our “working memory” compared to any of primate is absolutely dogshit, we’ve got too much other shit going on to remember a 27 step pattern we were flashed on a screen in five seconds. But a chimp has that down the first time they see it, and so does AI.

      If it was used to supplement a human’s natural defiency it would have been world changing already at a tiny fraction of the cost. But they want to completely replace humans which just flat out isn’t gonna happen in our lifetimes.

  • technocrit@lemmy.dbzer0.comEnglish
    15·
    16 days ago

    Paul Erdos made like 10000 conjectures that nobody cares about.

    It’s completely unsurprising that a bot would solve a few if not many.

    That’s what mathematicians have used computers for since they invented them.

    Of course a computer can arrange dots better than a mathematicians from 1946. jfc.

    This reasoning is contained in a 125-page document, which the company has not fully released.

    I’m shocked.