> If you already have a paper proof (which is the case in the original post), and you know it’s correct

You may think it's correct until you realize that your argument transitively depends on some obscure results from the 70s that nobody had ever questioned but that turned out to be flawed. Probably salvageable, but now you suddenly need to prove something completely different that you might not even understand very well. There's an interview with Kevin Buzzard somewhere where he talks about how that happened to them (couldn't find a reference rn).