• piecat@lemmy.world
    link
    fedilink
    English
    arrow-up
    15
    arrow-down
    10
    ·
    11 months ago

    See, I would argue the exact opposite. It sounds like you don’t understand how it works.

    Because it’s not “replication” or “copying”.

    • BURN@lemmy.world
      link
      fedilink
      English
      arrow-up
      12
      arrow-down
      8
      ·
      11 months ago

      Most LLMs can be made to spit out training data. That’s pretty much replication in my book.

      Statistical models don’t create anything. They replicate variations of their training data.

      • Dkarma@lemmy.world
        link
        fedilink
        English
        arrow-up
        7
        arrow-down
        6
        ·
        11 months ago

        Painters replicate variations of their training pieces too. You’re pretending there’s a difference between human inspired and training inspired and that you should get paid for that inspiration in one case just cuz “big corp”

        • BURN@lemmy.world
          link
          fedilink
          English
          arrow-up
          7
          arrow-down
          4
          ·
          11 months ago

          Because there is a difference. A computer does not learn or understand anything. Human beings can transform a concept. A LLM or other generative AI does not transform a concept at all.

          • stephen01king@lemmy.zip
            link
            fedilink
            English
            arrow-up
            2
            arrow-down
            4
            ·
            edit-2
            11 months ago

            So if I ask it to create a story about a cow juggling bowling balls, it was not creating an original story? Just spitting out stories it has heard of before?

            Edit: missed a ‘not’.

            • BURN@lemmy.world
              link
              fedilink
              English
              arrow-up
              4
              arrow-down
              2
              ·
              edit-2
              11 months ago

              It’s spitting out variations of the statistical results based on your input parameter. It reorganizes ideas and reorganizes the stories it has seen into something else. That’s not transforming the data by adding something new, rather just retrofitting existing data to sound like it’s creating something new

              • stephen01king@lemmy.zip
                link
                fedilink
                English
                arrow-up
                2
                arrow-down
                1
                ·
                11 months ago

                retrofitting existing data to sound like it’s creating something new.

                What the difference? That is basically how new human ideas are formed. Did you think you add completely new ideas everytime you transform your previous knowledge?

                But since you’re so confident in your claims, I’m sure it should be easy to prove the following ChatGPT output is not new and can be easily traced back to its training data:

                Prompt: Create a short poem about a cow juggling bowling balls on a boat

                Output: In a boat on gentle waves it sways, A cow, not grazing in greenish bays. Hooves deftly juggle, balls in flight, Bowling orbs, a whimsical sight.

                Bovine artist, on the sea’s embrace, Balancing spheres with tranquil grace. Ocean breeze, a playful gale, A cow’s performance, a quirky tale.

      • zwaetschgeraeuber@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        arrow-down
        1
        ·
        11 months ago

        when you read something and recite it, what do you do? exactly, spitting out the training data, if you trained long enough