• Isthisreddit@lemmy.world
    link
    fedilink
    English
    arrow-up
    5
    arrow-down
    2
    ·
    6 months ago

    Good analysis, but I’d say your not taking a vital factor into account with “AI”, it’s only going to get better, and crooks, conmen, corporations, etc are going to find new ways to weaponize it. I use LLM’s often, and for my purposes, they are truly amazing (now that Google search is fucking useless)

    • ArmokGoB@lemmy.dbzer0.com
      link
      fedilink
      English
      arrow-up
      4
      ·
      6 months ago

      I wouldn’t be so sure it will get much better. It might just stagnate, barely getting any better as we reach the limits of what is possible with our current hardware.

      • Obi@sopuli.xyz
        link
        fedilink
        English
        arrow-up
        2
        ·
        6 months ago

        Also it’s starting to run out of good training material, once they’re ingesting a majority of stuff that’s already AI generated it might devolve like a kid whose dad is also his uncle.

    • drathvedro@lemm.ee
      link
      fedilink
      English
      arrow-up
      1
      ·
      6 months ago

      Well, I’d like that to happen, actually. As it stands, the only use of AI I am legit a fan of is music mixers like suno and udio - it makes me smile every time I hear a creative mix of lyrics and genre. Others mostly just litter the internet with unhelpful stuff of questionable legality. When it gets good enough that errors are so rare that I could trust the output without checking, while being built on free and licensed stuff, I’d be much more more inclined to use them.