In fact, we loaned the big machine to some students at Dartmouth. And it... so, it disappeared up to Dartmouth and about 10 years later, I thought it would be nice to collect it and see if it still worked, but no one could find it. So, this is the only trace – along with some photographs – of this machine, which was called the SNARC: Stochastic Neural Analogue Reinforcement Computer or something. Anyway, then I did my PhD thesis on theories of how to make a smarter machine which didn’t just make responses at random, but would look ahead and predict what would happen and compare different predictions and... sort of... that eventually turned into my six layer... Freud’s theory had three layers of this... this id, which is a very elementary kind of conditioned reflex learning machine and a superego, which is a self conscious high level machine; and the ego, which – in Freud’s view – is a rather passive process, which negotiates between ambitious goals that you want to do and primitive goals that... or fears or things that you have to do or... or try to prevent. Freud had wonderful... wonderful cognitive theories and... I had six layers instead of just three and that was... that came out in the Society of Mind book in the 1980s. But in the 1950s I worked on this... these reinforcement theories, which were inspired by the early theories of Skinner and... and those pioneers.