[carapace]

Does that presuppose that it's computationally expensive to verify proofs? If so, isn't that kind of unrealistic?

[nynox]

Verifying fully formalized proofs is cheap. What is expensive is to produce such formalized proofs. More precisely, to formalize a proof written in a typical math paper is extremely time-consuming and not so informative (that's why it's almost never done in practice).

[carapace]

Ah, cheers. :)

But doesn't that bring us full-circle to Buzzard's point? Isn't that why he's saying, "Hey gang, let's do math with machines." in the first place?