Walter, a champion for robot education rights, wanted to reduce the amount of homework machines have to do to become intelligent. He was interested in temporally hierarchical approaches to robotic learning, and in remembering experiences in a way that allows them to be drawn on in new situations, across different platforms. His fascination with the history of AI development luckily lived alongside a healthy desire to come up with entirely new ways of doing things. Two centuries ago, a mechanical chess machine toured the world and played at a high level. It transpired later that there was actually a small chess master squeezed inside the contraption. Embedding a human in a machine might be slightly too strong a prior, but Walter is a firm believer that the more we can bring from our own brilliant brains to AI systems, the better.