• zlatko@programming.dev
    link
    fedilink
    English
    arrow-up
    9
    ·
    1 day ago

    I would hardly call this dropping a hammer, though. I mean, the article itself says that OpenAI was lobbying for stuff like this.

    For one, I don’t think it’s addressing any of the most troublesome problems. No mention of environmental impact. No mention of societal impact, or economic one. Nobody is addressing the horrible misuse of LLMs in, say education or biases that are built into the LLMs.

    For two, what are the consequences? Sam Altman can go around saying whatever the hell he wants, mislead the government, the public, even their own users. No consequences. They can use bait and switch, no problems. The worst thing they can get is a few million dollars of fines. That’s for the long time been just the cost of doing business.

    It’s a fig leaf law that doesn’t look at any of the worst problems we have but the worst offenders can use it to say “no wait we’re regulated”.

  • CmdrShepard49@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    61
    arrow-down
    1
    ·
    edit-2
    1 day ago

    the AI Safety Measures Act creates a framework that will require the “developers of the largest advanced AI systems” (read: OpenAI, Anthropic, Google, etc.) to provide public disclosures as to their safety practices, report any significant safety incidents, and remain in line with a set of compliance processes. It also creates a channel for whistleblowers working inside AI labs to safely and securely relay any safety concerns they might have without fear of repercussions.

    Whoa this sounds pretty serious. It asks these companies to self report safety incidents and creates a “channel” for whistleblowers.

    That’s some serious hammer dropping. 🙄

    OpenAI and Anthropic both endorsed the bill as it moved through the legislature

    Of course they did because this is just the Democratic version of what Republicans do with regulatory capture. They just like to dress their version up with a facade of “consumer protections” when the law doesn’t actually protect anyone or improve anything.

    • buddascrayon@lemmy.world
      link
      fedilink
      English
      arrow-up
      12
      arrow-down
      5
      ·
      1 day ago

      That requirement will apply to any AI company reporting more than $500 million in revenue.

      This part right here makes it laughably simple for these companies to circumvent this law considering so many of these companies are operating at a loss. And it would only takes some creative accounting for them to just barely make just not enough to qualify.

      • cecilkorik@lemmy.ca
        link
        fedilink
        English
        arrow-up
        7
        ·
        1 day ago

        Revenue and profit/loss are not the same thing. Revenue only counts the money they bring in BEFORE any expenses or losses, and is used precisely to avoid the issue you raised. That said, it is still possible to get around it, but it’s not quite as trivial as just spending more.

      • MrEff@lemmy.world
        link
        fedilink
        English
        arrow-up
        9
        ·
        1 day ago

        If my company made a billion dollars while losing 10 billion dollars, I still had a billion dollars in revenue. I just also happened to piss away 9 more billion dollars than I made.

  • affiliate@lemmy.world
    link
    fedilink
    English
    arrow-up
    33
    ·
    2 days ago

    Those requirements are set to go into effect on January 1, 2028, now that the bill has been signed into law.

    Why is he waiting 1.5 years to drop the hammer?

    • WhoIzDisIz@lemmy.today
      link
      fedilink
      English
      arrow-up
      31
      arrow-down
      1
      ·
      2 days ago

      To give the companies plenty of time to build up a case that ties it up in courts for years. Meanwhile, he gets to use it for campaign propaganda.

      • EntheoNaut@lemmy.ml
        link
        fedilink
        English
        arrow-up
        7
        ·
        1 day ago

        Pritzker is a tool of AIPAC who dropped the proverbial nerf-hammer / rubber dildo

        Fuxk him and Newsome and all the DNC

        • cecilkorik@lemmy.ca
          link
          fedilink
          English
          arrow-up
          5
          ·
          edit-2
          1 day ago

          The solution to all the billionaires taking everyone’s money is not going to be solved by finding the right billionaire. Nobody with a billion dollars gives a shit about you. They aren’t going to help you. They don’t even have the personal human context to have any understanding of anything you need help with, and even if they can afford to have someone tell them, and have the empathy to bother, they’ll still end up trying to fix in ways that don’t actually help you. They don’t get it. They won’t get it. They can’t. They’re living in a totally different universe. Money doesn’t measure success, it measures privilege.

    • Tollana1234567@lemmy.today
      link
      fedilink
      English
      arrow-up
      3
      ·
      edit-2
      1 day ago

      llooks hes waiting out the election 2028, while wanting the current ones to be built. seems like alot of states are doing this, all these politicans likely got promised money or lucrative position.

      Mills, whitmer all made sure AI DC are built.

  • EmilieEasie@fedinsfw.app
    link
    fedilink
    English
    arrow-up
    9
    arrow-down
    1
    ·
    edit-2
    2 days ago

    I think this just means that people won’t build AI data centers there (for now), which is awesome for them tbh