On the other hand, I've contacted several of my heroes (not been able to meet as many of them in person) and that's always been an exhilerating, formative experience. I strongly recommend it if you can think of a good reason. (I have a list of heroes I have yet to reach out to because I haven't yet encountered an interesting enough problem to offer them. Several of them unfortunately have an actuarial deadline not too far into the future.)
His code did not match the hype, to say the least. His SDLC even less so.
There is probably an ego associated with being renowned that doesn't align with team-based work. He likened basic things like code reviews or PRs to being brought before The Hague and that the rest of the team was a bunch of bureaucrats.
some are just a bit better at guessing
You can legitly learn how to do things properly and people who learnt to do that do the polar opposite of guessing. It is just that the world of software development has yet to be made liable for their results in the same way as civil or electrical engineers. So in software development many are just guessing because guessing wrong won't ruin their life.