Who is betting on what in AI, and against whom.
Researcher
Harvard professor and co-director of the Kempner Institute, whose work spans the theory of reinforcement learning and, lately, the empirical science of scaling: which optimizers, precisions and self-improvement loops actually move the loss curve. One of the few theorists both camps 1 and 4 cite.
Known for Kempner Institute; Scaling laws for precision; Policy gradient theory