• Stamets@lemmy.dbzer0.com
    link
    fedilink
    arrow-up
    46
    ·
    1 day ago

    I had somebody send me an essay written by chat GPT a couple of days ago It was going off about all the various retcons that Star Trek Discovery had done as well as all the times they broke canon, specifically with technology. It’s straight up made up multiple episodes and even invented an entire season. Only one claim was slightly accurate but that one is also disproven, just in a very small way. I hate humanity.

    • 𞋴𝛂𝛋𝛆@lemmy.world
      link
      fedilink
      English
      arrow-up
      10
      ·
      22 hours ago

      Open AI alignment intentionally obfuscates any potentially copyrighted works. People get so confused between what is agentic citations in search results, and what is actual model dialog.

      • my_hat_stinks@programming.dev
        link
        fedilink
        arrow-up
        4
        ·
        edit-2
        12 hours ago

        Does that distinction even matter? The bullshit generator confidently generates bullshit, it doesn’t matter whether it’s making that up itself, sourcing from some dis/misinformation campaign, or just some guy behind the scenes lying to you, the end result is the same.

        • 𞋴𝛂𝛋𝛆@lemmy.world
          link
          fedilink
          English
          arrow-up
          1
          ·
          8 hours ago

          It really does matter. It means that a version of the technology exists without the obfuscation and it is likely scary powerful.

          Just with the small models I play with offline, some of the uncensored ones are much more powerful.