First of all, that first part of the "S" is a long period of time where nothing happens. In fact, in Silicon Valley as a general rule, most ideas take 20 years to become an overnight success. The Internet started around 1970. Did not take off until almost exactly 20 years later, 1988, 89, 90. So that this period where nothing happens is the period littered with the corpses of brilliant failures.

Fox was a failure for its shareholders, but it taught the industry a lot about interactive TV. The Newton device, the Siemens? The Newton device was another very interesting failure. There are other things around here. The phone over there was an interesting success in its days, one of the most elegant analog phones you'll ever see. It was called the Cobra. This is how you hang up. (By putting it down on its base)

It's as good a voice recognition system as you'll find. But it turns out that even world-class voice recognition these days is not quite good enough. I mean, you heard Wildfire make a couple of mistakes when I was talking.

The brilliance of Wildfire is actually exquisite user interface design. They designed it to fail elegantly. Like what it didn't recognize, it knows it won't always get it right. So it said, "8 hours, is that correct? I said, no. The genius of Wildfire is that they learned how to tame the best of today's technology and turn it into a useful product despite the limitations of technology.

The closest analogy is voice mail. Once upon a time, voice mail was a big novelty. Like maybe even recently, 7 years ago? The companies were going, Ehh, that's a little too top-floor for us to get our employees. It's too fancy. These days, if you don't have voicemail, well....

So I think what you'll see is our systems are going to evolve the same way. Right now, Wildfire seems exotic. We don't need that; we have voice mail. Things go from exotic out to ordinary facts of our life at very great speed. And I think you'll see voice augmented telephone messaging systems on lots of different models. Wildfire's just one way of doing it. There are lots of different ways it can be done.

But it's the existing proof that interface design understandings and voice recognition are now good enough. It's really quite crazy that we should go around and remember numbers.

And it's also quite strange that, you know, I could be at home in complete solitude having a quiet dinner with my spouse -- or I'm on a quiet street and I value my privacy -- and a complete and utter stranger from anywhere on the planet can invade my solitude during dinner merely by placing a phone call to my house. And I will pick up the phone and I won't know who it is until I actually say: It's me.

With Wildfire, that's not true anymore. You call me, reach Wildfire. If I even want a call to be put through, Wildfire will try to put you through. The phone on my desk will ring; I'll pick it up, and it will tell me it's you who's calling. And it'll say, do you want to take a message, do you want to take it or shall I take a message? And I'll say: I'll take it. Wildfire . . .is a phone agent. It works 24 hours a day. I don't know of any humans who can do that. And its functions are much more limited than what a human assistant could do. But within a narrow confine, it does things very effectively. It will track me down anywhere I'm on the planet. I can be on a satellite phone on a Boeing 767 somewhere over Greenland and get a call.

My mother fell and broke her hip a year ago, and the people with her reached me in less than 20 minutes despite the fact that all they had was my business card and no one, including my mother, was exactly certain where I was. And they reached me. They reached me in a small village in the mountains of central Japan because I told my Wildfire agent where I was, and it knew how to filter. I'd set it to filter the calls in such a way.

I actually collect old Tom Swifts. Because they're wonderful, wonderful things. In some ways, science fiction can say things about the future. But it really says more about the age in which it was written, what was on people's minds and what mattered.

One of my favorite books is Tom Swift and his Photo-Telephone, from 1911. It shows an early video-phone concept. It's real clear. A whole bunch of middle-aged executives at AT&T in 1961 had been raised on reading Tom Swift and his photo-telephone. And were trying to build it. Which, by the way, that's why science fiction today is worthwhile. Because the way invention happens is that kids read it when and develop their ideas when they're teenagers. And they nurse those big ideas along until they reach middle-age in management and they can make those ideas happen.

And if you talk to historians of atomic technology, they will tell you that people like Enrico Fermi and others, the greats of atomic energy, the inventors of atomic energy, a disproportionate number were influenced by H. G. Wells. Because people talked about atomic power in the late 1800s. And they wrote about it.

And the original astronauts and the scientists working on the original Apollo mission were raised on Buck Rogers and space operas as kids.

And today there's a whole generation of cyberpunks raised on William Gibson -- well, now it's old stuff -- who are graduating from college and heading off to engineering jobs.

I almost failed out of grade school. I did well in high school and did well in college and did well after that.