Hacker News Reader

Comment by zugi | Hacker News Reader

I agree on this point!

And also, training an open weights AI on the outputs of an AI trained using "fair use", is itself "fair use" that benefits everyone.

View on HN

Unfortunately congress and the senate will only pass malicious compliance laws like that when it's to the benefit of megadonors.

Honestly if you are putting models behind your own terms of use it's only fair use if I get to do the same for my content, but it's too late for me I folks in US sue them for every dollar or force them to make their models "fair-use".

I don't think scooping all of the internet and saying rules for thee but not for me is the way to go.

> rules for thee but not for me

Exactly, the same copyright rules should and do apply to both parties.

I'm glad copyright law didn't prevent the creation of LLMs. I'm also glad it doesn't prevent others from creating LLMs. I think that benefits us all.

Without government intervention, I think distillation will prevent these giant AI companies from recouping their investment. But that's not our problem.

That pretty obviously fails the copyright test. Training on LLM weights isn't transformative because you are producing...another LLM.

That pretty obviously fails the copyright test. Training on text/pictures/audio isn't transformative because you are producing...text/pictures/audio.

The other LLM is obviously transformative because it's a different LLM and therefore is not even the same kind of thing, like how a white horse is not a horse, obviously.

Reply on news.ycombinator.com