I lean more towards the "proofs from the book" mindset - if there's a really beautiful proof of some theorem out there I would be happy to see it no matter whether it was discovered by a human or a computer, and conversely if a computer has churned through a search space and generated a clunky proof in 100000 lines of lean code that just means that people trying to find a nice proof can do so with the assurance that the theorem is true. note that mathematicians didn't give up on trying to find a better proof of the four colour theorem once a clunky computer assisted proof showed the theorem was true.