Comment by pertymcpert

I wonder if OpenAI have been less strict about not training on proprietary or legally questionable code sources.