• bagsy@lemmy.world
    link
    fedilink
    English
    arrow-up
    17
    ·
    4 hours ago

    What are they going to do in 5 years when the internet is 99% slop? How will then train the models then? Will we be forever stuck with 2025-2026 models because there isnt enough new human content to train on?

    • brucethemoose@lemmy.world
      link
      fedilink
      English
      arrow-up
      4
      ·
      edit-2
      2 hours ago

      The Chinese models are way ahead of this, training smartly and frugally instead of on a huge quantity. They don’t really need more internet to work fine as tools.

      …And they share data with each other, and ignore copyright. Seemingly. I wouldn’t be surprised if the Chinese govt is providing a lot of data to them.

      Also, it turns out multilingual training works very well. So even if the English internet turns to slop, other languages may fare better.

    • tigeruppercut@lemmy.zip
      link
      fedilink
      English
      arrow-up
      10
      arrow-down
      1
      ·
      3 hours ago

      Money line go up in the short term. Who cares if they’re creating a hallucinating slop ouroboros that will ruin the internet for the foreseeable future? That’s the next CEO’s problem!

    • altphoto@lemmy.today
      link
      fedilink
      English
      arrow-up
      1
      ·
      3 hours ago

      What about this coherent answer that I am composing for you? Clearly ten mice brains carefully placed over silicone chips can compute prime numbers much faster than 20 undergrads ever could. So who fo you want in your laptop? Ten rats, just more Pentium, 20 undergrads, or windows 25?