• Photuris@lemmy.ml
    link
    fedilink
    arrow-up
    17
    ·
    23 hours ago

    Wow, these dipshits don’t even take a few minutes to proofread their output and edit, do they‽

    • GreenKnight23@lemmy.world
      link
      fedilink
      arrow-up
      10
      ·
      21 hours ago

      I played around with using an local LLM to generate some short stories as a test to see what writing a novel would be like.

      it was easy to generate, but it only followed about 70% of the prompt and I had to do drastic rewrites along with serious proofreading for two reasons.

      1. word reuse was horrible
      2. death loops where it just repeats the slightly different versions of what it said already

      I’d like to use it as a creative scaffold, but honestly it’s not even worth the effort. it’s impossible to track characters, their arc, and all the plot devices. maybe in 10-20 years we’ll be there. until then it’s just slop.

    • Phoenixz@lemmy.ca
      link
      fedilink
      arrow-up
      12
      ·
      21 hours ago

      That right there is one of the biggest issues with the very many issues with Ai: people trust it blindly, even after countless examples of how stupid it can be