Gold’s Theorem – it is impossible to learn an infinite class with 100% certainty from a finite training set. Any finite training set is a subset of infinitely many infinite and finite sets. No finite training set contains enough ‘information’ to allow learning the class with absolute certainty.