Title: Feynman on Artificial Intelligence and Machine Learning, with Updates

Abstract: I present my recollections of Richard Feynman's mid-1980s interest in artificial intelligence and neural networks, set in the technical context of the physics-related approaches to neural networks of that time. I attempt to evaluate his ideas in the light of the substantial advances in the field since then, and vice versa. There are aspects of Feynman's interests that I think have been largely achieved and others that remain excitingly open, notably in computational science, and potentially including the revival of symbolic methods therein.