I liked how it didn't matter whether or not the previous case was rational, because if it were then you've already proven it, and if it wasn't then your current case proves it. ie:
If this rational then it is proven, but if it's irrational then we can just say:
Which is rational. So whether or not the first case is rational has no impact whatsoever on the fact that a^b can be rational.