• 4am@lemmy.zip
    link
    fedilink
    English
    arrow-up
    7
    arrow-down
    5
    ·
    4 days ago

    There is no such thing as a privacy first AI, unless you’ve the hardware to run the model yourself (and, you don’t).

    • Kissaki@feddit.org
      link
      fedilink
      English
      arrow-up
      7
      arrow-down
      1
      ·
      4 days ago

      There different kinds of AI and some run just fine locally or even on mobile. Not everything is a big LLM.

      • 4am@lemmy.zip
        link
        fedilink
        English
        arrow-up
        1
        ·
        edit-2
        8 hours ago

        I am aware, but none of those are the kind of AI that anyone who makes a web browser wants to put into one.

        The whole point of shoving LLM assistants into products is for tracking. People share their intimate thoughts with LLMs. They plot murders. They fall in love with it. They most certainly ask for summaries of this quarter’s unreleased earnings report. In a browser, they ask it to do the research for them, and depend on what it tells them in kind. They trust it alone with their deepest desires and their darkest thoughts. That’s worth so much more than HTTP_REFERRER and a tracking pixel, and those are already worth billions.

        What an incredible opportunity to seize the means of information. It’s every capitalists wet dream to be an integral part of the ultimate humanity control machine.

        And that is why it needs to be stopped, even if we assume good intentions. We cannot throw up our hands and allow this to be normalized. This must be mocked and ridiculed and made to seem as ridiculous as it is. It must be monetarily penalized. It must be stopped.