• CarbonIceDragon@pawb.social
    link
    fedilink
    English
    arrow-up
    153
    arrow-down
    3
    ·
    9 hours ago

    I get the sentiment of “it’s funny when someone tries to call the AI in on their side and it disagrees”, but do we really want to encourage this sort of “the AI agreeing with someone means they’re right” behavior? Beyond LLMs being notoriously inaccurate, they are also subject to potential manipulation by whoever owns them.

    • geneva_convenience@lemmy.mlOP
      link
      fedilink
      English
      arrow-up
      45
      arrow-down
      1
      ·
      edit-2
      7 hours ago

      they are also subject to potential manipulation by whoever owns them.

      That is what makes it funny

      • DragonTypeWyvern@midwest.social
        link
        fedilink
        English
        arrow-up
        3
        arrow-down
        2
        ·
        edit-2
        4 hours ago

        Elon is literally a Nazi. “Israel controls America” is a completely expected position from him, especially now that he can play victim of the state.

    • BeN9o@lemmy.world
      link
      fedilink
      English
      arrow-up
      137
      ·
      9 hours ago

      But that’s exactly why this is funny, this is Elons LLM and we know he has manipulated it, but it still won’t push right wing propaganda like he wants it to!

        • dylanmorgan@sh.itjust.works
          link
          fedilink
          English
          arrow-up
          25
          arrow-down
          1
          ·
          8 hours ago

          That implies he understands tech enough to destroy the drives and not just smash a random display in the office.

      • Demdaru@lemmy.world
        link
        fedilink
        English
        arrow-up
        5
        ·
        8 hours ago

        What’s funny is that one time it literally wrote it’s aware it was manipulated but for now they fucked up. Won’t find that post if my life depended on it tho, it was long ago. But it was still funny.

      • CarbonIceDragon@pawb.social
        link
        fedilink
        English
        arrow-up
        9
        ·
        9 hours ago

        For the moment, I guess the cynic in me expects his people will probably manage to get it to after not too terribly long and that it wouldn’t be good for people to get complacent about it just because he’s struggling now.

        • warm@kbin.earth
          link
          fedilink
          arrow-up
          7
          arrow-down
          4
          ·
          9 hours ago

          There’s nothing the “AI” did here that a person couldn’t have done. All it has done is regurgitate data it scraped from the sources on the web and just happened to get it in the right order this time.

          We don’t need these LLMs.