• kishkebab@lemmings.world
    link
    fedilink
    English
    arrow-up
    7
    ·
    10 hours ago

    LLM bots appear human enough on the surface, and can sometimes identify whats in an image correctly (context is usually lost on them).

    A big tell is the format, they all have a unique way of speech. If you see enough you start to recognize the patterns.

      • vateso5074@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        1 hour ago

        It’s a good Turing Test training ground, and if perfected, can be used to control the direction of discussion on the platform (and future platforms).