When Things Start to Think
Published by Henry Holt
1999, 224 pages
Buy it online
The Next Computer Revolution
Reviewed by Brad Murray
Neil Gershenfeld likes things. He likes building physical, real devices that do things. He also has a rare genius for knowing what kinds of things are useful as opposed to what things are just nifty. Consequently, he has brought an air of practicality to the Media Lab at MIT that has, paradoxically, made its productions that much more fantastic.
Neil's book is broken into three sections: What, which discusses some of the actual projects that Neil has pulled off; Why, which outlines his philosophy regarding objects that sense and interact with their environment and how they fit into our world; and finally How, which examines how the Media Lab actually works -- how it fits into the academio-economic niche between the ivory towers and the satanic mills.
Gershenfeld and his students and colleagues have wired Yo-Yo Ma to a computer to let him explore digital music space with the interface he knows best--- his cello. They have developed computers that augment their memories and eyesight, allowing them to communicate with something approximating telepathy and experience heretofore unexplored sensory scenarios. One student sees only through digital lenses. He finds he likes to explore the world with different filters, sometimes tuning his eyes so that they can only detect motion, like some predators. Sometimes he prefers everything sideways. Sometimes he mounts the cameras on his shoes to experience the world like an insect. They have built things that boggle the mind, using the human body as a private local area network, transmitting signals between wristwatch and shoe using the body as a communications medium. They can exchange information with each other with a handshake. They have coffee mugs that tell the coffee machine what kind of coffee they like. They have reusable laser printing paper and soon will have paper that can physically move itself to the filing cabinet.
In short, Gershenfeld is exploring not the abstract space of what could possibly be done in the future, but the far more interesting space of what we can do right now if we want to. His philosophy regarding computers is simple: they should be useful, they should be reliable, they should not make demands of us, but rather we should make demands of them, and they should be dirt cheap. This is a million miles from the current philosophy of computer software and hardware design, which somehow sacrifices our needs to the constraints of clumsy input devices and unreliable software. He wants to know why your VCR can't go out and find out what time it is rather than ask you to program it. And most importantly, he does not just talk this philosophy. He does not talk about how these things should be thus in a perfect world: he describes how to make them thus now. And his students are busy doing the constructing.
In many ways the most interesting aspect of When Things Start to Think is not the technology itself, but the environment that spawns it. The Media Lab survives on support from corporations that do not do their own blue-sky research. And they sponsor the lab because it constantly delivers products and devices that change the way the world thinks and runs. NEC dropped by for a magic show that the Media Lab Physics group put on with Penn and Teller using a device that could make music by detecting humans moving in the space in front of it, and came away with a derivative device that can detect the nature of the front passenger in an automobile, allowing the air bag control to inflate (in the case of an appropriately-sized passenger) or not (in the case of a reverse-mounted baby seat). Future versions will go further and actually control the rate of inflation appropriately. These are the connections Gershenfeld and his students facilitate. They start with ideas and then they build them rather than theorize about them.
The book itself is an ugly affair, with a crude cover and uninspired layout, but the content is some of the most inspiring I have ever read (I am a geek, of course). Gershenfeld writes passionately and eloquently whether narrating an anecdote about Yo-Yo Ma misbehaving with his computer, or outlining his philosophy of device design. This book is a real gem, one that makes you want to go out and build things yourself. More importantly, it makes you wonder why you take for granted the demands your computer places on you. It's time for another computer revolution, but it's the users who need to rise up, not the computers. | February 1999
Brad Murray is a self-taught geek and school-taught philosopher. His loves include science, philosophy, philosophy of science, computer tinkering, and sweet manic depressives. He is an explorer of operating systems and programming languages and other computer esoterica, and a large French company pays him to design complex control systems. He views the planet as something here for his amusement, and therefore tries to skip the unamusing bits whenever possible. No degree, lots of school. He can be reached at firstname.lastname@example.org but rarely reads his e-mail.