Isn't it obvious that the feasible way to respond is to start developing AIs that rewrite proofs for human understanding