One of the most successful recent applications of quantum information has been the generation of "certified" random numbers. The unique properties of quantum devices allow the generation of numbers that are provably random, offering a potential major benefit to cryptography and other fields.  Constructing the right theory to prove quantum randomness can be challenging.  Quantum devices and adversaries are modeled by complex vector spaces whose dimensions may be arbitrarily large, and proving security in such a setting leads to some difficult mathematical problems.  This talk will be a retrospective look at one successful approach to quantum randomness that is based on the convexity of the Schatten matrix norm.