Comment by VladimirGolovin

I feel that long, unexplainable, hard-to-comprehend proofs may become commonplace when we get to AIs that are capable of discovering new proofs on their own.