The Science Of The Next 150 Years: 150 Years in the Future
Predicting what next year's (or next week's) ipad is going to be like is hard enough. Knowing what computers in general will be like 150 years from now—an eternity in technology development—is nearly impossible. On the other hand, technology prophets, computer pioneers and researchers have never been known for their reticence on the subject of the future. So we thought it wouldn't hurt to ask them. For starters, will there even be computers in the far future?

“There will definitely be computers,” says nanotechnology oracle Eric Drexler of the University of Oxford. “They're more fundamental than the wheel.”

But Stewart Brand, whose business is forecasting, refuses to even speculate about what they would be like: “Maybe because I'm a professional futurist, specific future look-backs I know are going to be risible (always), so I veer away from them. I don't even like examining those made by other people. It feels like I'm exploring their medicine cabinet—a violation of privacy—learning too much about their frailties and illusions.”

George Dyson, author of books about computers and global intelligence, says, “I can tell you a lot about computing 50, 100 and 150 years ago but really nothing about computing 50, 100 or 150 years in the future. It's just truly impossible to predict: all I can guarantee is that any prediction will be wrong!” He then relents and makes one: “In 150 years most of the important computation will be analog computation (for the same reason that most of the important numbers are real numbers but not integers), and the notion of all-digital computation will be a quaint relic.”

Ivan Sutherland, who invented Sketchpad, the basis for today's ubiquitous graphical user interfaces, says, “I have no clue about the state of the world 150 years from now. If you want to know the future, ask the young people who will create it.”

“I suspect they don't know, either!” says his friend Vinton Cerf, one of the “fathers of the Internet,” who today works for Google. “Actually there may be some clues in studies to assess minimum power required from the quantum perspective for any kind of computation. There is also the possibility that the kind of asynchronous parallelism we see in brain function may find its way into ‘hardware,’ although I am tempted to believe that some computations will prove to be more readily accomplished using more conventional hardware structures.” (An asynchronous computer is one whose operations are not governed by a central clock that times operations.)

Danny Hillis, inventor of the Connection Machine, a massively parallel supercomputer, says, “We will have computers, but they may not be made out of electronics. They will be more intimately connected to our minds than today's tenuous linkups through screens and keyboards. Some parts of them may actually be implanted into us, and it may be hard to tell where we end and the computers begin.”

Nathan Myhrvold, formerly chief technology officer at Microsoft, agrees: “Yes, there will be computers 150 years from now, but they may be hard to recognize. If you asked Edison or Tesla about electric motors, they would probably have said yes, too, and they were right: there are hundreds of tiny electric motors built into everything we have. You still occasionally have a big electric motor that is recognizable, but mostly they have dissolved into the fabric of our lives. The same will be true of computers in 150 years. In a few cases, we'll find that there is something very recognizable as a computer, but mostly they will be inside of everything else.

“In that time frame, computers will be vastly more powerful. I would be surprised if they aren't much smarter than people. That weirds some people out—they have this view that we ought to be the smartest things around. But at one time, they would have said that about strength, and humans are very weak compared to machines. We've coped with that. Computers are already smarter than we are at narrow tasks. That will broaden until they are smarter than us at everything.”

Michael Freedman, a researcher at Microsoft Station Q, which is focused on studies of topological quantum computing, says, “Implanted devices will not be popular: as now, beauty and style, not computational power, will dictate the choice of bodily modification. But devices will be small and have direct communication to the brain. Special sunglasses or hats may confer the ability to muddle through with a foreign language by directly interacting with speech centers.”

Freedman adds that “computation will be pervasive in the environment, with difficult tasks (like sunglasses translation) being done in low-power, cryogenic, Josephson logic computers scattered all about. The golden age of mathematics that we currently live in will continue to flourish as human-machine collaboration heads toward a seamless perfection. Science-fiction writers will worry about human obsolescence, but 150 years from now people will have more to do and better ways of doing it than ever before. The world best in the marathon will be one hour, 58 minutes and 59 seconds, and the Nose on El Capitan in Yosemite will be climbed ropeless.”

Well, maybe. The problem with all such predictions is that they run up against the principle of computational irreducibility, an epistemological barrier to knowledge of the future. According to Stephen Wolfram in his book A New Kind of Science, a system is computationally irreducible when “in effect, there can be no way to predict how the system will behave except by going through almost as many steps of computation as the evolution of the system itself.” In other words, “there is no general shortcut: no way to find the outcome without doing essentially as much work as the system itself.”

The technological pathway to the computers of the future seems to constitute a system of this type. It will be a product of countless human decisions, technological innovations, market forces and consumer choices, among other things, and there does not seem to be any way of knowing in advance how those forces and decisions will mutually interact to create the future of technology—which means that there is no way to know what the computer of the future will be like other than to wait 150 years and find out.