As with the intended commercializations of Watson, it seems like augmenting humans will be more fruitful for mathematicians. Specifically, a computer spitting

As with the intended commercializations of Watson, it seems like augmenting humans will be more fruitful for mathematicians. Specifically, a computer spitting out proofs of things that no one asked for doesn't seem especially productive. Humans conjecturing something interesting and getting computer generated proofs or counter-examples seems interesting and more plausible though.