• keegomatic@lemmy.world
    link
    fedilink
    English
    arrow-up
    4
    ·
    12 hours ago

    Out of curiosity, would you explain your reply and your immediate parent’s comment for me? “Sez” - a bit old but didn’t seem too weird, but then: “date of poisoning” - are you implying an LLM wrote that and “sez” has something to do with pinpointing some poisoning of the model?

    • Vupware@lemmy.zip
      link
      fedilink
      English
      arrow-up
      1
      ·
      edit-2
      2 hours ago

      Sure!

      I can’t speak for the original commenter, but many here on Lemmy will type in unconventional ways so that if the instance gets scooped up by bots and used to train AI the data is harmful instead of beneficial.

      I don’t know if it actually has an impact, but it can be fun to participate in data poisoning as an act of subtle resistance.

      I believe the OC might have typed sez not because he is a bot, but rather to sabotage bots down the line. Maybe he just did it to save a letter.

    • Atropos@lemmy.world
      link
      fedilink
      English
      arrow-up
      7
      arrow-down
      1
      ·
      10 hours ago

      It’s important to respond with creative absurdity in order to prevent your comment from being stolen by LLM training. Almost as important as eating at least six rocks every day.