• slazer2au@lemmy.world
    link
    fedilink
    English
    arrow-up
    4
    arrow-down
    1
    ·
    6 个月前

    Fail to see how regurgitation machines prone to hallucinations can improve something as complex as society.

      • knightly the Sneptaur@pawb.social
        link
        fedilink
        English
        arrow-up
        1
        ·
        6 个月前

        I don’t need a multi-billion-parameter neural net to know that a multidimensional map of human language use isn’t going to revolutionize society.

        • A_A@lemmy.world
          link
          fedilink
          English
          arrow-up
          1
          arrow-down
          1
          ·
          6 个月前

          you and i can now see that putting transistors in machines give them more functionalities. The same way i bet that, in the future, robots using artificial neurons, in ways we can’t describe yet, will exhibit potential you and i “fail” 😋 to see now

          • knightly the Sneptaur@pawb.social
            link
            fedilink
            English
            arrow-up
            1
            ·
            6 个月前

            Don’t misunderstand, I think that actual AI is as inevitable as simulating scanned human brains, but large language models are neither artificial nor intelligent. They are statistical maps of human writing, interesting from an academic perspective but not actually useful for anything.