• fartographer@lemmy.world
    link
    fedilink
    English
    arrow-up
    3
    ·
    3 hours ago

    I get your point, but I also argue “no it’s not.” The AI bros want to lie about how their super-special autocomplete is the most dangerous of all without any consequences. Good luck getting any employees to willingly substantiate your claims when 20 years in prison is on the line.

    Even better, if they try to hack another company and claim their AI did it on its own, they get treated like a private company who decided to perform an unsanctioned nuclear test.

    Make it dangerous for them to claim they’re dangerous.

    • boonhet@sopuli.xyz
      link
      fedilink
      English
      arrow-up
      2
      ·
      2 hours ago

      Okay, fair, it’s not the ONLY reason they claim it’s dangerous. Obviously marketing is a large part of it. But they sure as hell would both like a duopoly setup enforced by LAW for themselves. Well at this point, I guess quadropoly? I doubt Google or Musk would give up their seats at the table. But anyone making an AI better or cheaper than the established players, would be deemed dangerous. Those working with the DoD would of course not be considered dangerous. Foreign models? Absolutely dangerous, restrict access as much as possible, especially to open weight ones.