• fallaciousBasis@lemmy.world
    link
    fedilink
    English
    arrow-up
    10
    arrow-down
    5
    ·
    1 day ago

    His position was essentially:

    1. AI systems sometimes produce harmful or illegal content despite safeguards.
    2. Governments shouldn’t use that as a reason to pressure platform gatekeepers to remove an entire service.

    Important distinctions:

    1. He said he defends open platforms and free speech.
    2. He explicitly said he does not defend the harmful content itself.
    • exaybachae@startrek.website
      link
      fedilink
      English
      arrow-up
      2
      arrow-down
      3
      ·
      1 day ago

      Yeah, was pretty sure it was something like that…

      I used AI to make a Home Alone themed digital Christmas card last year. It was completely PG. But it was a nightmare to do as I kept tripping over gates on account of Kevin being a child and Home Alone being a copyrighted property.

      I went all the way to the dark end of the AI gen spectrum, to where kiddie porn abdolutely happens, but the product I produced there was unsatisfactory. So I went alI went back to the top players with all the gates and worked around them by reworking my prompts for what seemed like forever.

      1 year earlier I would’ve gotten the results I wanted with my first prompt, maybe tweaking fonts for the words, fixing spelling errors, or making minor visual features. 5min tops.

      It really turned me off from continuing to play with AI.

      Those gates have gotten really nasty.

      It’s as though users are treated as guilty in advance and then denied the opportunity to prove themselves innocent.

      Completely backwards attitude and practice.