• SickIcarus@sh.itjust.works
    link
    fedilink
    English
    arrow-up
    3
    ·
    1 year ago

    Huffman has argued the changes are a business decision to force AI companies training on Reddit’s data to pony up

    LMFAO that ship has fucking sailed. They already used Reddit data through 2021 and got everything they needed for free. Huffman is locking the barn after the horses are already out, and setting it on fire. What a fucking dumbass.

    • Moonrise2473@feddit.it
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 year ago

      Also no sane company would pay that ridiculous price for training a LLM on memes

      Why paying exorbitant api prices when you can scrape for free? Like all those image models clearly trained on scraped watermarked stock images

    • CrateDane@feddit.dk
      link
      fedilink
      English
      arrow-up
      2
      ·
      1 year ago

      LMFAO that ship has fucking sailed. They already used Reddit data through 2021 and got everything they needed for free. Huffman is locking the barn after the horses are already out, and setting it on fire. What a fucking dumbass.

      Plus they can just scrape the data without using the API. It’s a red herring, just a lie to cover up his desire to kill third party apps.