• mangaskahn@lemmy.world
      link
      fedilink
      English
      arrow-up
      16
      arrow-down
      3
      ·
      2 days ago

      “Generative” is a misnomer. It will never generate anything new, it can only regurgitate existing ideas based on patterns that already exist. It’s very good at pattern recognition and summarizing, but lacks the ability to form a distinct new idea.

      • jj4211@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        15 hours ago

        lacks the ability to form a distinct new idea.

        Yeah, but it’s got that in common with a frighteningly large number of people…

        See management, marketing, streaming, social media, etc…

      • dondelelcaro@lemmy.world
        link
        fedilink
        English
        arrow-up
        8
        arrow-down
        1
        ·
        2 days ago

        It’s only good at summarizing things which have coherence to its training set. Any ability to summarize input outside of its training is accidental.

      • Ajen@sh.itjust.works
        link
        fedilink
        English
        arrow-up
        8
        arrow-down
        2
        ·
        2 days ago

        It will never generate something novel. Whether it will generate something “new” depends on your definition of “new,” which is a little more ambiguous than “novel.”

        Sorry if I’m being too pedantic.

        • mangaskahn@lemmy.world
          link
          fedilink
          English
          arrow-up
          3
          ·
          2 days ago

          That’s an interesting thought. Genetics is largely a mixing and copying process with occasional “hallucinations” in the form of transcription errors. Most of these errors result in the termination of the hallucinated code. Hallucinations that damage the termination process result in cancer. In the larger sense of evolution, there’s a robust external “review” process. Environmental pressures, predation, and resource availability weed out most of the mistakes and selects the results most likely to succeed.