Despite Bernd Matthias' somewhat apocryphal suggestion, theory has played an important role (misguided or otherwise) for those searching for new classes of superconductors. In this talk, I will review some crazy, and not so crazy, ideas for new superconductors: heavy holes playing the role of ions, filling up the Brillouin zone with Fermi surfaces, searching for materials with large superexchange,hybrid systems that play off the pairing scale against the phase stiffness scale, and attempts at large scale data mining. I will end with a thought offered by Mac Beasley ' even if we had a high temperature superconductor, would it really make a difference?