I haven’t seen this discussed and it confuses me: companies tend to avoid tricky licenses like (A)GPLv3 for legal reasons. Why is there no hesitation to using LLMs to generate code, when we still don’t know whether that is fine under copyright law?

Training an LLM has been ruled fair use in the US. I am not aware of a ruling whether that applies to using it. And if the output of the model isn’t a copyright infringement of the input, then the output cannot be copyrighted because it was not created by a person. So AI-generated software cannot be copyrighted.

en.wikipedia.org/wiki/Monkey_s

0

If you have a fediverse account, you can quote this note from your own instance. Search https://hachyderm.io/users/mxey/statuses/115878527876364329 on your instance and quote it. (Note that quoting is not supported in Mastodon.)