Warwick has spent the majority of his career at the University of Reading, where he landed 11 years ago following a stint at British Telecom and teaching jobs at various British universities. At the University of Reading, Warwick has been puzzling over cybernetics -- defined by him as the topic of "communication and control in human and machines." The "machine" part of his experimentation includes his electronic pets, the "seven dwarves": a troupe of little wheeled robots with white faces and ultrasonic eyes who have been teaching themselves both individual and group behavior. He and his research assistants observe as these robots figure out how to move around corrals, compete in robot soccer matches, and perform other notable robot feats. Like ants (or humans, for that matter), they'll follow each other around, mimicking a leader who has already discovered the secret to not bumping into walls.

So far, Warwick's most notorious robotic accomplishment occurred last year, when one of his fleet was the first robot to autonomously train another robot. One of the dwarf robots learned how to zoom around a corral in Reading; using an Internet link and radio signals, that robot then remotely trained another robot in Buffalo, N.Y., to zoom around in the exact same way. It may sound like a relatively modest accomplishment, but the event went into the Guinness Book of World Records as a robot first. ("Quite why I have no idea," laughs Warwick. "I think it was printed somewhere between Pamela Anderson and the Spice Girls.")

"It was a fairly simple demonstration that robots do learn and adapt, and that to me is quite critical -- that they aren't just programmed to do what you want them to do," explains Warwick. In other words, in the future it is quite possible that robots won't even need humans to program them in the first place. Instead, they will reprogram and rebuild themselves. "Intelligent robots can come up with a large amount of independence and do creative and imaginative things that you never imagined they could do, and communicate with each other and create their own languages."

Warwick describes this not as a step towards artificial intelligence but "machine intelligence" -- a new form of silicon-based smarts different from, and perhaps even superior to, our own carbon-based intelligence. As he matter-of-factly puts it, machine "brains" are going to be far superior to human brains. We already know that a computer can surpass a human, he quickly details, in "number crunching, how quickly it can operate, the mathematical capabilities, the memory, the logic. We know that machine intelligence has a lot of advantages already over the way the human brain works. Looking to the future, in what ways is the human brain going to remain ahead?"

Warwick argues that there aren't very many ways in which humans will be superior to machines, despite the current advantages we boast. "Our senses are restricted, they suit us as humans, but machines have the capabilities of sensing the world in ultrasonics, ultraviolet, infrared, X-ray, gamma ray," Warwick explains. "Machines can sense the world in all sorts of ways that we can't hope to do, unless we start looking at implants. The human body and the human brain are fine as far as being a human is concerned; but they are very, very restricted and limited when we look into the future."

Warwick predicts that a future in which computer brains have roughly the same power, if not more, than human brains is only 20-30 years away. The possibility of a computer takeover is one reason why Warwick has been pursuing his implant projects -- he hopes that by tapping into machine intelligence he'll be able to enhance human brainpower.

Recent Stories