Based on a sarcastic comment I made a little bit ago but kept thinking about.

For a long time now (before the current AI bubble, even), the majority of stocks traded have been performed by algorithms using machine learning.

https://en.wikipedia.org/wiki/Algorithmic_trading

Tech companies keep developing AI-powered solutions because it’s a quick way to earn money through stocks. The majority of profit in the stock market right now is coming from AI slop.

https://fortune.com/2025/10/07/ai-bubble-cisco-moment-dotcom-crash-nvidia-jensen-huang-top-analyst/

So, when the majority of stocks being traded are decided by AI, and it just happens that the most profitable investments are likewise boosting AI development, is this essentially AI using the mechanics of capitalism to self-select its own advancement?

  • baggachipz@sh.itjust.works
    link
    fedilink
    arrow-up
    1
    ·
    4 days ago

    That’s a pretty good Black Mirror plot, it would be hilarious. Too bad the current “AI” isn’t even capable of math.

  • gkak.laₛ@lemmy.zip
    link
    fedilink
    English
    arrow-up
    1
    ·
    5 days ago

    No

    self-select its own advancement

    It’s a program that writes text, that’s why it’s called a “Large Language Model”; “AI” is just a hype term.

    It’s not self-aware, it can’t make arbitrary decisions, and it doesn’t have free will

    (PS I’m interpreting “AI” as "LLM"s, since other kind of Machine Learning models have been used for many years before we came up with the LLMs we have today)

  • makeshiftreaper@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    ·
    5 days ago

    Most algorithmic trading is not done by AI and won’t be at any time in the near future. Algo trading is primarily used for HFT (high frequency trading) and fundamental trading, both of which AI are terrible for. In the HFT world traders are fighting for fractions of microseconds, so AI is far too slow to be helpful. Trading on fundamentals also won’t work with AI because AI makes frequent (by trading standards) math errors

    • bjorney@lemmy.ca
      link
      fedilink
      arrow-up
      0
      ·
      5 days ago

      Things like gradient boosting, supervised + reinforcement learning, NLP (sentiment analysis), etc have been used in algorithmic trading for decades.

      Hell, much to the chagrin of statisticians, some people even lump linear regression under the AI umbrella

      Also I think you are confusing fundamentals and technicals

      • makeshiftreaper@lemmy.world
        link
        fedilink
        English
        arrow-up
        1
        ·
        4 days ago

        You’re correct on fundamentals and technicals, it’s been a while since I took a Financial Mathematics class and I don’t remember doing very well in it regardless

  • Bronzebeard@lemmy.zip
    link
    fedilink
    arrow-up
    1
    ·
    5 days ago

    The majority of profit is coming from the companies supplying the hardware, not the ai companies themselves.

    They’re the ones selling the shovels during a gold rush.

    • gravitas_deficiency@sh.itjust.works
      link
      fedilink
      English
      arrow-up
      1
      ·
      5 days ago

      This statement categorically ignores the rampant round-tripping and incestuous deals and that all the big tech players are making these days. They’re spending the same dollar like 20 times and calling it all profit. It’s financial engineering at its worst and most rampant. It’s the .com bubble look like a kindergarten scuffle. The crash is going to be economically cataclysmic.