> It's entirely reasonable that the set of useful mathematical proofs is a proper superset of human intelligible useful proofs.

If you can't explain something in a way that a child could understands it, you don't fully understand it either.