Fifty Years of P vs. NP and the Possibility of the Impossible

Here is Scott Aaronson's paper on it[0].

[0]: https://www.scottaaronson.com/papers/pnp.pdf

That was a terrible interview. "P vs NP has seen no progress, but now we have machine learning, so let's use that!"

This has to be one of the worst takes on P vs NP.