[auggierose]

Good point. So then the question becomes: can you prove in HoTT that there exist irrational x, y such that x^y is rational?

[fmap]

Yes. You can translate the classical proof, assuming excluded middle for propositions.