• deltreed@lemmy.world
    link
    fedilink
    English
    arrow-up
    0
    ·
    edit-2
    2 months ago

    I don’t mind him using copyrighted materials as long as it leads to OpenAI becoming truly open source. Humans can replicate anything found in the wild with minor variations, so AI should have the same access. This is how human creativity builds upon itself. Why limit AI? We already know all the jobs people have will be replaced anyway eventually.

    • JackbyDev@programming.dev
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      That’s simply not going to happen lol. They aren’t just going to release the secret sauce just because it would be a nice thing to do.

      • Starbuncle@lemmy.ca
        link
        fedilink
        English
        arrow-up
        0
        ·
        2 months ago

        They should have to make the weights public if they train on data scraped from the public internet. It’s be something like copyleft by default for AI training.

    • Zink@programming.dev
      link
      fedilink
      English
      arrow-up
      0
      ·
      2 months ago

      That’s a good point. AIs/LLMs will exist and will necessarily learn from copyrighted materials without traceability back to the copyright owners to compensate them.

      Sounds to me like AIs/LLMs can’t and shouldn’t be proprietary systems owned by private entities for profit, then.

      • emmy67@lemmy.world
        link
        fedilink
        English
        arrow-up
        0
        ·
        edit-2
        2 months ago

        Nor should what they produce be copyrightable in any form. Even if it’s the base upon which an artist builds.

        Also, it should all be free.