And then when I got interested in making learning machines, I had great fortune to become friends with two young incredibly original professors, George Miller and Joe Licklider, and they were making up new kinds of theories about cognitive psychology. And at that time, in the late 1940s, they were way ahead of everyone else; it was sort of the time that Norbert Wiener’s Cybernetics was beginning to spread, and they understood those theories, and I understood them even better, and so that sort of made my decision for me. That is, here was this new field called cognitive psychology, and I was really good at it, getting new ideas, and except for a few other people like Sigmund Freud and William James, who maybe had better ideas than I did, there wasn’t much competition. Licklider and Miller were… accepted me as an equal, and so you could say that my career was based on cowardice, or… or something. If there was a field where there were five or 10 people who were better at it, why bother? Because in these fields, you know, if you were a very good physicist and you ran across a Richard Feynman, what’s the use?

And I was a… quite a good physicist, actually, but I had the fortune to run across Richard Feynman, again we became friends, and I wrote a… in fact I wrote a nice paper that he liked enough to publish, but… but I couldn’t see being serious about it, because with people like Murray Gell-Mann and Alex Rich, and people like that.