@xilatsad @netbsd a few of my projects have policies that basically say "no contributions from any currently existing LLM or anything trained using similar methods to one, but if you get a transformer model to spit out code without stealing any training data then let's see how good it is."