I personally think we can accept machine-generated proofs as long as the author of the algorithm is vouching for its results.
On a related note, isn't it possible to fabricate a "peer" algorithm which in turn "reviews" this tremendous amound of data?
I suppose that this would turn on what you mean by "vouching" for a proof. It seems to me that using a computer-generated proof would require a proof of soundness of the program.
I personally think we can accept machine-generated proofs as long as the author of the algorithm is vouching for its results. On a related note, isn't it possible to fabricate a "peer" algorithm which in turn "reviews" this tremendous amound of data?
I suppose that this would turn on what you mean by "vouching" for a proof. It seems to me that using a computer-generated proof would require a proof of soundness of the program.