well, similarity on circles is a phony business
in essence, all circles are similar by definition, I mean, hell they're all circular, what's so unsimilar about that?
however, the silly proof asks to translate it about and then dilate it to be superimposed on the other.