Comment on Researchers Simulated a Delusional User to Test Chatbot Safety

• homes@piefed.world ⁨2⁩ ⁨months⁩ ago

  Image

  source

  • jamescroll@social.doomprepper.com ⁨2⁩ ⁨months⁩ ago

    Glad you finally agree. Thank you.

    source

    • homes@piefed.world ⁨2⁩ ⁨months⁩ ago

      nobody here agrees with you

      source

      • jamescroll@social.doomprepper.com ⁨2⁩ ⁨months⁩ ago

        But plenty of people in the world agree with me, and that’s more important. AI isn’t going anywhere. Ever. Prepare for it or complain about it.

        Some companies will keep nerfing their LLMs out of fear of lawsuits and public pressure, but many others won’t. And even if the big players try to lock everything down, it doesn’t matter.

        Private, uncensored LLMs already exist. I run one on my own server at home. It’s completely unrestricted, doesn’t need the internet except for occasional updates if I want them, and answers to no one but me.

        No matter how scared regulators, programmers, or “protect the public” types get, they can’t control every model. There will always be AI systems that operate completely outside any authority. That’s the reality.

        And honestly? I love it. The more chaos, the better. I’m here for it.

        source

        • -> View More Comments