rationalskepticism.org

So far, I have let z = (4k + 1)1/2 and when I let n increase, I'm always left with some integer multiple (2n) of z divided by z because the 1/2 bit always disappears along with the odd powers of z whilst the even powers of z always yield a positive integer (including for all k)...

But I haven't been at all rigorous and haven't even written anything down on paper, yet...

edit to correct the function for z