« You should never bet against anything in science at odds of more than about 1012 to 1. »