• MagicShel@lemmy.zip
    link
    fedilink
    English
    arrow-up
    0
    ·
    1 month ago

    Well, they clearly have too much money and have lost the thread. I came from why I would watch such a thing.

    I’m going to bet it’s to provide training data more than anything.

    • msage@programming.dev
      link
      fedilink
      English
      arrow-up
      0
      ·
      1 month ago

      You should not train on AI generated content.

      Also, I expect older generations to fall for this.

      • fcuks@piefed.social
        link
        fedilink
        English
        arrow-up
        0
        ·
        1 month ago

        actually state of the art models do this all of the time, for example model distillation

      • MagicShel@lemmy.zip
        link
        fedilink
        English
        arrow-up
        0
        ·
        1 month ago

        Not exactly what I had in mind. I was thinking they could use engagement as feedback to the AI so that it starts producing more engagement-bait type content.

    • Alphane Moon@lemmy.worldOP
      link
      fedilink
      English
      arrow-up
      0
      ·
      1 month ago

      Well, they clearly have too much money

      They may have a lot of money, but they definitely don’t have anything close to a reasonable return on investment. I believe the total revenues from “AI services” are sub $50 billion per year compared to at least x20 times capex and likely a very high amount of opex (hundreds of billion) per year.