NVIDIA: Copyrighted Books Are Just Statistical Correlations to Our AI Models * TorrentFreak

As part of a copyright lawsuit, NVIDIA has informed the court that copyrighted books are just statistical correlations to their AI models.

6 comments

Here's hoping the plaintiffs can bring the leaked internal docs from nvidia's management directing engineers to violate the TOS of several sites to show a pattern of malfeasance, and that the judge has at least a modicum of understanding regarding these topics. Throw the book(s) at 'em.
- No, use bricks, because they will take the books for their AI
  
  Iirc the leaked internal docs and reporting showed that nvidia was downloading 81 years worth of video per day. If the pattern holds, and we "throw" the books at them, it probably looks a lot like one of those building-sized front end loaders dropping the library of Alexandria on those executives.
  
  As an alternative, I would be fine if nvidia were no longer allowed to hold copyright or trademarks since the company believes that those don't matter, required to share all internal documents (including any and all planning), and required to open source all their code until the company implodes
  
  I just imagined what would happen to their GPUs if bricks were used :D
Fruit of the poisoned tree. Disney's "until author's death + 70 years" copyrights are BS, would be nice if nVidia and all the AI companies were to argue to change that.

You've viewed 6 comments.