• drkt@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    10 days ago

    I accidentally deleted my comment sorry here it is again

    Current models are utterly dependent on using others’ work without permission or compensation,

    They are not dependent on it, no. They simply do that because it’s the cheapest way to build a huge dataset to train on.

    the people behind AI companies are now advocating for the abolishing of IP law so they can exploit artists even more

    I advocate for the total abolishment of copyright, IP and any adjacent laws for the exact opposite reason; artists would not need copyright and innovators would not need IP to protect themselves if we lived in a society that nurtured a healthier culture of sharing. In its most extreme form, I want to get rid of money such that nobody, artists especially, need not money to justify their continued existence. Human beings were not meant to be enslaved to a monetary structure and it has become the driving force of misery all around the world.

    will not fail to succumb to the same enshittification

    It’s pretty clear to me that you haven’t participated in the open source AI race because we don’t need the corporate AIs. I don’t say that like a ‘lmao ur not as smart as me’ but open source AI development, especially stable diffusion and chat LLMs, has caught up to corporate AIs in every way but training data, because unlike the corporations, they walk a thin legal line. I’ve been following it closely since GPT2. It was open source that first came up with the idea of using smaller models to do specific things instead of trying to train one huge model to do everything.

    • Libra00@lemmy.world
      link
      fedilink
      English
      arrow-up
      2
      ·
      9 days ago

      I accidentally deleted my comment sorry here it is again

      You also replied to the wrong comment, but I got you. ;)

      They are not dependent on it, no. They simply do that because it’s the cheapest way to build a huge dataset to train on.

      Tell that to OpenAI. They’re not the only ones who have come out against IP law in the wake of the kerfuffle about how it relates to AI.

      I advocate for the total abolishment of copyright

      I am generally anti-IP law myself, so I only point it out because the likes of OpenAI and Elon Musk trying to undermine it shows just how desperately dependent they are upon unlicensed content to train their models, and how that makes an interesting contrast to the fact that they hire teams of lawyers to go after people who violate their IP.

      IP and any adjacent laws for the exact opposite reason; artists would not need copyright and innovators would not need IP to protect themselves if we lived in a society that nurtured a healthier culture of sharing. In its most extreme form, I want to get rid of money such that nobody, artists especially, need not money to justify their continued existence. Human beings were not meant to be enslaved to a monetary structure and it has become the driving force of misery all around the world.

      Yeah, I’m a pinko commie who also thinks human society should exist for the sole purpose of meeting the needs of all of its members too, but that’s a conversation for another time.

      It’s pretty clear to me that you haven’t participated in the open source AI race because we don’t need the corporate AIs.

      I honestly didn’t even know it was a thing until recently, so nope, I really have not. I have read about it some though. I use ChatGPT to help me structure writing, worldbuild, etc, and - like most people who use corporate AI - find it easier (for the moment) then installing a bunch of shit and fiddling with it for hours (so basically it has the same barriers to entry as widespread adoption of linux - if it doesn’t just work out of the box, most people don’t give a shit how much better it is) only to thrash my GPU into an early grave, and so do millions of other people. I am also skeptical about open source AI’s ability to compete long-term on practically any grounds (accessibility, volume of training data, responsiveness, advanced model features, integration with existing software, etc), but I would be quite happy to be proven wrong on that count.

      • Mystic Mushroom [Ze/Zir]@lemmy.dbzer0.com
        link
        fedilink
        arrow-up
        2
        arrow-down
        1
        ·
        9 days ago

        Tell that to OpenAI. They’re not the only ones who have come out against IP law in the wake of the kerfuffle about how it relates to AI.

        This is one of those cases where two groups seem like they are united on a subject but aren’t really. OpenAI claims they want to abolish copyright for the good of themselves and AI, but that isn’t really true. They just want immunity from it, complete death of copyright doesn’t benefit them or any of the companies, since they’d lose their moral high horse against civilians using their material without paying them, as well as their ability to legally retaliate.

        Really their claimed stance against copyright doesn’t undermine any of the anarchist and pirate arguments against IP law as it is today and as a concept, it doesn’t even align with it. It’s no different than right wingers saying there’s a war on science or that they want to protect women and children from groomers. It’s just obfuscating or lying about their intentions. OpenAI is just claiming their against copyright because it’s easier to swallow than saying they wish to be immune from copyright, maybe even be able to usurp individual people’s copyrights outright.

        • Libra00@lemmy.world
          link
          fedilink
          English
          arrow-up
          0
          arrow-down
          3
          ·
          9 days ago

          This was not an argument about how everyone who has talked about avoiding or abolishing IP law is the same or has the same intentions, this was an argument about how AI is utterly dependent upon unlicensed content, and as evidence goes, them saying ‘gosh we really need to be able to skirt/get rid of/whatever this to keep going’ is, regardless of the complexities of the situation, in itself evidence of that fact.

          The specifics of their particular flavor of opposition to being bound by IP law, while interesting, don’t particularly matter to that argument. But thanks for the added context.