What we discovered was that the things that we thought were hard about intelligence like playing chess or solving calculus problems were actually really easy. It was easy to get the computer to do those. It was much harder to do the things that were so easy for us that we didn't even notice that they were happening, like recognising a face. And so the irony is the easy stuff turned out to be hard and the hard stuff turned out to be easy. But even the easy stuff that turned out to be hard we're really starting to do now because we've gotten the computational capacity to do that.

But I don't think we understood that when I first arrived at the AI lab. We did have this tremendous excitement that we were onto something. And we were also onto something about computers and the Internet and, you know, using computers. We were using computers to write our term papers. We called it 'the million dollar typewriter'. It was... Because a computer was a million dollars and using it to write a term paper was, you know, really not legitimate use of funds or we did things like some people had programmed a video... What we would now call a video game on the computer. Space War. So we would sneak in and work on video games and word processing and email systems and chat systems. And all of this stuff that we weren't supposed to be doing. That's the stuff that turned out to be, you know, billion dollar industries. Was all the things that we weren't supposed to be doing. And the stuff that we were supposed to be doing which was the AI is still kind of a research problem. But the things we were doing with it in the spare time, those were the parts of computers that took off because it turns out the same things that excited us about them excited the world about them.

And so, yes, I'd love to go back there and relook at, realising the things that we did that turned out to be important that we had no idea. I remember for example we would send email around but we realised that you couldn't really convey emotion in an email. And so there was this mailing list that we said, 'Well, you know, let's invent a smile. Let's use like a colon and a parenthesis. Let's invent a wink. Which is semi-colon and...' And we were discussing this. And I remember thinking, you know, wow, this is such a waste of time, we're doing all this silly stuff.

But in the end, you know, all that stuff really ended up mattering quite a lot to people. At the time I don't think we had much of a sense of how important it was.