What will be the most important advances in business technology in the next five to 10 years? I don't see anything on that time scale that's likely to change technology qualitatively. A lot more businesses are going to take real advantage of what's out there now. Looking probably further forward, when we truly get good speech recognition, I think that will be a dramatic change in the way things get used.
What do you see as the real promise of speech recognition? There's some simple things, like you talk to it in English and it'll spit it back out in French. I think the computer stops being a passive tool. If you had to interact with other people by poking buttons, if people just responded blindly to instructions you gave them, then it would certainly be a different world.
You have said you foresee amending Moore's Law to a four- to five-year cycle. Does that suggest an industry that's maturing? We're approaching atomic sizes where, as you shrink further, things don't behave properly. We've got a couple more generations, maybe more, before we enter some limits there. But the industry will still be moving at a phenomenal rate.
You've been skeptical about DNA chips and quantum computing. But they seem like natural extensions to Moore's Law. Quantum mechanics is so far from a practical device that I can't imagine it ever having a significant impact on what we consider the computing world today. For the DNA chip, again, it's hard for me to see how you can input data and get data out of something that's dependent on DNA for the execution. You can demonstrate these things in laboratories. But it's far from being, or even showing a pathway to, a practical solution.
Is there a time where technology development might be post-Mooreian, if you will? [Chuckles.] I'm writing a paper right now entitled "No Exponential Is Forever." Any physical thing that's been growing exponentially eventually comes to an end; it can't do it anymore.
Fitzgerald is a freelance writer in Oakland, Calif. Contact him at firstname.lastname@example.org.