• wizardbeard@lemmy.dbzer0.com
    link
    fedilink
    English
    arrow-up
    15
    ·
    28 days ago

    I would love if a group of hackers could reverse engineer the network connections used to send the data back to Meta and distribute software people could choose to run to flood Meta’s “incoming pipe” with convincing nonsense, like how the Ad Nauseum ad blocker fakes random ad clicks to help destroy any attempts at ad profiling.

    Like SETI@home but for spamming false info at the corporate surveillance assholes.

  • FlashMobOfOne@lemmy.world
    link
    fedilink
    arrow-up
    3
    ·
    28 days ago

    They’re really not useless, though at this point maybe a little half-baked, since they’re “ai-powered” and LLM’s are very much half-baked.

    They could, for an example, improve things greatly for people who can’t see or hear; problem is they are so closely associated with creeps, and Meta is so incentivized to market to creepy people, that the positive possibilities for this kind of tech are an afterthought.

    • BallyM@lemmy.worldOP
      link
      fedilink
      arrow-up
      1
      ·
      28 days ago

      No, Meta AI (like OpenAI and others) relies on sweatshop labour mostly in Africa and Latin America to review and classify images mostly