4 ms·

Thank you for explaining this. I had the exact same misunderstanding as the parent poster, and was only corrected (and educated) by reading your comment. A hyp

by gspr 11d ago

Thank you for explaining this. I had the exact same misunderstanding as the parent poster, and was only corrected (and educated) by reading your comment. A hypothetical relating to the legal aspects: Since this was mostly done by an LLM, the legal considerations would become incredibly interesting in cases where the original code (source or binary) was illegally leaked. What happens if the LLM saw the leaked code in its training material? (My non-lawyer brain thinks it's completely obvious that in that case you have a derived work, but apparently the world has decided that it's completely fine to train LLMs on e.g. GPL code without the output being considered derived, so what the hell do I know…)

orthoxerox 11d ago
> the world has decided that it's completely fine to train LLMs on e.g. GPL code without the output being considered derived I think it's more of a "the constitution is not a suicide pact" approach to GPL: while the GPL is obviously violated by its inclusion in the training corpus, we have two options: - demand the destruction of LLMs that are trained on copyrighted or copylefted works and the training of new ones - accept that LLMs that are trained on copyrighted or copylefted works are more important to humanity than the rights of the authors
- gspr 11d ago
  Well if it's the second, then surely I can train my LLM using the frontier labs' models too? And have my LLM "interpret and reproduce" GTA 6? How is any of this going to work out if we just decide to throw out copyright like this? (I am well aware that it is not currently technically feasible to copyright-wash GTA 6. Replace that game with a best-selling novel or whatever.) And, if we do decide that copyright is just gone, can we please make laws to that effect? Instead of just shrugging our shoulders and going "oh well, this new thing is useful, so we'll let it slide, but don't you dare to pirate movies, because reasons!"
0points 11d ago
I think you are being misguided. The repository itself contains disassembly directly from the copyrighted ROM, and is thus not original work.

by gspr 11d ago

orthoxerox 11d ago

> the world has decided that it's completely fine to train LLMs on e.g. GPL code without the output being considered derived I think it's more of a "the constitution is not a suicide pact" approach to GPL: while the GPL is obviously violated by its inclusion in the training corpus, we have two options: - demand the destruction of LLMs that are trained on copyrighted or copylefted works and the training of new ones - accept that LLMs that are trained on copyrighted or copylefted works are more important to humanity than the rights of the authors

gspr 11d ago
Well if it's the second, then surely I can train my LLM using the frontier labs' models too? And have my LLM "interpret and reproduce" GTA 6? How is any of this going to work out if we just decide to throw out copyright like this? (I am well aware that it is not currently technically feasible to copyright-wash GTA 6. Replace that game with a best-selling novel or whatever.) And, if we do decide that copyright is just gone, can we please make laws to that effect? Instead of just shrugging our shoulders and going "oh well, this new thing is useful, so we'll let it slide, but don't you dare to pirate movies, because reasons!"

0points 11d ago

I think you are being misguided. The repository itself contains disassembly directly from the copyrighted ROM, and is thus not original work.