Hacker News new | past | comments | ask | show | jobs | submit login

    and that the model surely used his films as data.
While I don't doubt it's true, this could be challenging to prove, because Studio Ponoc (ex-Ghibli) has produced work that uh, hews rather closely to Miyazaki's style. Were the models trained on Ghibli, Ponoc, both, something else, etc?

I mean, I have no doubts. But proving it seems tough!




Ponoc is made of former Ghibli employees who founded a new home when Ghibli's future was uncertain. I am sure they are on friendly terms, if not family, with Ghibli: they worked together for years. People like them can have a gentleman's agreement.

What is OpenAI in all this, if not a greedy, sloppy, soulless outsider stealing their Art and effort for financial gain without ever asking for permission?


I don't disagree with anything you said, but it doesn't seem to follow from what I wrote.

I pointed out a reason why litigation could be difficult. I'm quite sure nothing I wrote could have been seen as defending OpenAI. Just that I felt litigation would be tricky.


I gathered the Japanese government legalized using copyrighted works to train AI last year: https://www.privacyworld.blog/2024/03/japans-new-draft-guide...


Well they then stole from Ponoc, too, right?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: