I heard they based a book on it.

And which, in turn, was used to train LLMs without any guidance as to whether it's "good" or "bad" code.