I am so tired of the breathless LinkedIn posts and the corporate media manufacturing consent for this technology. We need to stop calling it “AI.” That implies there is some sort of sentient thought process happening behind the curtain. There isn’t.

It cites court cases that don’t exist. It writes code filled with security vulnerabilities. It tells you to put glue on your pizza. It can’t even do basic math without being hand-held by a prompt engineer.

The only reason the ruling class is pushing this Actually Incompetent slush so hard isn’t because it’s better than human labor. It’s because they are desperate to break labor power. They would rather replace you with a broken, hallucinating chatbot that costs pennies to run than pay a living wage to a human being who actually knows what they are doing.

As long as it destroys the bargaining power of the working class, the VCs don’t care if the product works or not.

  • Flying_Lynx@lemmy.ml
    link
    fedilink
    arrow-up
    7
    ·
    edit-2
    11 days ago

    It’s used as an AG. An Acknowledgment Generator. A production slave, for ease of mind or the likes.

    Which is (indeed) hiding incompetence while the LLM-slave is doing the actual word-sorting stuff. Use it often enough and your brain can’t perform any decision making process without acknowledgement from this LLM. From no-time to incapable of summarizing a large text. From filtering values to expecting personal advice. Outsource your brain functionality too much and a dependency shift happens where you’re the one becoming the slave while your brain shrinks. And that while some kid at the operational side puts up all kinds of “truth-filters”.