• Dr. Moose@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    edit-2
    9 hours ago

    You are in a cult

    nah you’re in a cult. I admit that LLMs are imperfect tools but to completely dismiss them is ludite, it’s crazy. It’s just a computer and computers are good. I understand the sentiment about billionaires and control and ecology but it is missing the forest for the trees. The tools need to be captured by us through open project that embrace LLMs rather than play this losing game with these fake-ass boycott that - realistically and statically speaking - will never yield to anything of value other than to produce toxic bullying this very thread’s topic is displaying.

    You all need to do more and whine less.

    • sakuraba@lemmy.ml
      link
      fedilink
      English
      arrow-up
      1
      ·
      8 hours ago

      I’m sorry and now how should we get the resources to run local LLMs that are actually useful and not some homelab chatbot when the same people that manufactures the components needed for them are a mafia that outpriced everyone?

      Like seriously, what game do you want “us” to play here?

      Who is the toxic one here? Hank Green won’t read any of the stuff we put here but we sure are reading this nonsense

    • webadict@lemmy.world
      link
      fedilink
      English
      arrow-up
      1
      arrow-down
      1
      ·
      7 hours ago

      nah you’re in a cult.

      Okay. Genius revelations from the LLM.

      Realistically, statistically, LLMs do not make you work better. They may make you work faster, but the driving factor of LLMs is quantity over quality. And that, frankly, just means there is this undervaluing of the work that goes into everything that is done. More is better than good, and I just don’t think most people agree.

      You also cannot, realistically, capture generative AI. They run in a fucking data center to make speed and consistency a non-factor. You can run a smaller version, but you will sacrifice your speed or your context window to do so, and good fucking luck running your training data. It’s sort of like saying you can make your own factory.

      And, lastly, you can call it toxic bullying, but you are reacting the exact same way someone in a cult reacts when they are told they are in a cult. Feel free to look it up! Find where I am wrong. Why can you not do that?