@FatCrab - eviltoast
  • 0 Posts
  • 20 Comments
Joined 18 天前
cake
Cake day: 2025年6月26日

help-circle

  • Diffusion models iteratively convert noise across a space into forms and that’s what they are trained to do. In contrast to, say, a GPT that basically performs a recursive token prediction in sequence. They’re just totally different models, both in structure and mode of operation. Diffusion models are actually pretty incredible imo and I think we’re just beginning to scratch the surface of their power. A very fundamental part of most modes of cognition is converting the noise of unstructured multimodal signal data into something with form and intention, so being able to do this with a model, even if only in very very narrow domains right now, is a pretty massive leap forward.









  • FatCrab@slrpnk.nettoMemes@lemmy.mlJust baffling
    link
    fedilink
    arrow-up
    3
    arrow-down
    7
    ·
    13 天前

    We have seen authoritarian communism that regressed into autocratic nationalist imperialist communism. We have also seen authoritarian communism that has shifted into state capitalism. Just like there are many variations of capitalism that we have and have not seen, there are likewise many of socialism and even communism that we have and have not seen.









  • It isn’t quite what you’ve italicized, but make no mistake that’s it’s still a terrible precedent to set. Preliminary injunctions can still be granted by district courts, but now they need to be brought in the form of a class action and all the tediousness that entails. ACLU thankfully had one ready to go it seems just in case of this and it’s been filed already.

    Nevertheless, this also enables insane infringement of the first amendment. There is nothing stopping him from declaring membership of a particular political party is illegal, including state parties, and then requiring each state to independently challenge under a restricted class action. It’s ludicrous.



  • You are agreeing with the post you responded to. This ruling is only about training a model on legally obtained training data. It does not say it is ok to pirate works–if you pirate a work, no matter what you do with the infringing copy you’ve made, you’ve committed copyright infringement. It does not talk about model outputs, which is a very nuanced issue and likely to fall along similar analyses as music copyright imo. It only talks about whether training a model is intrinsically an infringement of copyright. And it isn’t because anything else is insane and be functionally impossible to differentiate from learning a writing technique by reading a book you bought from an author. Even a model that has overfit training data, it is in no way recognizable to any particular training datum. It’s hyperdimensioned matrix of numbers defining relationships between features and relationships between relationships.