• Arthur Besse@lemmy.ml
    link
    fedilink
    English
    arrow-up
    7
    ·
    5 days ago

    Fuck this site; it lost me when it got to listing “What kinds of things might they be good at?”

    Just say no.

    Using an LLM to summarize something is still a bad idea. The chances of it emphasizing the wrong thing, omitting the most important thing, or just outright making up “facts” remains high. LLMs Will Always Hallucinate.

    • HugeNerd@lemmy.ca
      link
      fedilink
      arrow-up
      5
      ·
      5 days ago

      We’re too late. People are copy/pasting code from AI into low-level functions that run hardware. I’m glad I can retire soon, this is absolute lunacy.