- cross-posted to:
- [email protected]
- cross-posted to:
- [email protected]
Meta CEO Mark Zuckerberg gave Meta’s Llama team approval to train on copyrighted documents, according to a new court filing.
Case file: https://storage.courtlistener.com/recap/gov.uscourts.cand.415175/gov.uscourts.cand.415175.376.0.pdf
Case file: https://storage.courtlistener.com/recap/gov.uscourts.cand.415175/gov.uscourts.cand.415175.377.0_1.pdf
Llama is open source, so they should be allowed to train on public data if they’re going to release it to the public.
The problem is with them using non licensed data to train proprietary models, not models in general.
If it were open source, I would be able to freely build it from source files on my own machine.
Open source != freeware
Llama is NOT open source. Read the license https://github.com/meta-llama/llama3/blob/main/LICENSE.
Well, the licence tries to prohibit people doing various things with it, but the model is open weights. Anyone can physically run it on their hardware, not something they can do with ChatGPT or Claude for example. You’re right, I shouldn’t have implied it was fully open source, but at least it only tries to legally, rather than physically, prevent people running and modifying the model themselves.
That’s not what open source means
Friendly reminder that
So yeah, take that how you want.
Huh? Open source has a definition. It means the source is accessible and one can build the software themself. I think you might be mixing up open source and FOSS (which does have to do with licenses).