I think that it's possible to study human intelligence, to...

I think that it's possible to study human intelligence, to know what it's made up of, to break it down into modules in a sense, and then -- little by little -- to build those modules into machines. And that's part of what I do.