• NoLemurs@lemmy.world
    link
    fedilink
    English
    arrow-up
    23
    ·
    4 hours ago

    I really hate that this article is pretending that Muse is somehow “spying” on users.

    This isn’t spying. This is just bad engineering.

    An LLM should only be given access to relevant information for a particular query. If you give an LLM access to irrelevant information, it gets confused. Anyone who’s used LLMs much knows this. It’s kind of baffling to me that the project managers over at Meta don’t understand how dumb their design is.

    • AHemlocksLie@lemmy.zip
      link
      fedilink
      English
      arrow-up
      1
      ·
      16 minutes ago

      We know that some AI companies are trying to use it to build a system to profile users for, e.g., monitoring protesters, and now that one of the most privacy invasive, data hungry tech companies got invasive again, you believe it was just a whoopsie doodles? Because I do not.

    • Buffalox@lemmy.world
      link
      fedilink
      English
      arrow-up
      24
      ·
      4 hours ago

      From experience, chances are that this is indeed spying, and not just “bad” engineering.
      Meta has been about spying from day 1. With Zuckerberg laughing his ass off, that people freely give away sensitive and personal information.

      • NoLemurs@lemmy.world
        link
        fedilink
        English
        arrow-up
        6
        ·
        4 hours ago

        I’m not saying Meta isn’t spying on users. I’m sure they are. But this doesn’t make sense as a symptom of spying.

        If you want to spy on users you just record all their info. You absolutely don’t give the LLM they talk to access to that same information - that’s nearly useless, and is one of the surest ways to make sure the users find out. The LLM doesn’t need to know about the spying.

          • Aqivex@fedinsfw.app
            link
            fedilink
            English
            arrow-up
            3
            ·
            59 minutes ago

            It’s both, and they are indeed dumb AF. Muse drinks from the same spying trough the rest of the slop platform sits on.

            Where they’re being idiots is that the model is internally being fed a system prompt meant to foster “engagement” in between “sessions”, and it absolutely is being fed not just context beyond the scope of the task, but that context…

            Is clearly being populated by another LLM model that is slurping from behind where the totally secure e2ee decode happens after the “private” message goes from Grandma Conspiracy Hoax’s inbox and through Aunt Zionist Shill onto Zuckerberg’s private masturbation window.

            The second model doesn’t match the info against the specific user before passing it to the first, it passes the full search results it pulled and leaves the first model to determine relevant use.

            Great shit if you’re running a tight local harnessing set up and don’t want to waste compute on your helpers sanitizing the a2a calls that the model handling orchestration would need to parse anyways.

            Asinine and dumbfucked when your set up is a multi-user activity panopticon which your bottom line desperately relies on pretending isn’t the case.

    • lIlIlIlIlIlIl@lemmy.world
      link
      fedilink
      English
      arrow-up
      12
      ·
      3 hours ago

      AI = Weighted text generator

      There isn’t anything there to HAVE dementia. We must remember this is random token generation, and not an intelligence as we otherwise know it