Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I have successfully reproduced a few projects with LLM assistance via strict cleanroom rules and only working off public specifications.


Once you use a LLM, the room is no longer clean.


Does that hold up if it’s in an entirely different language?

Because I don’t think so


[flagged]


Human training and AI training are legally distinct processes.


No, it isn't. A human wasn't trained on the material they're trying to reproduce.




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

Search: