Earlier this week, I wrote about singularity, the point at which computers will be able to “out-think” humans. Estimates of if and when that will happen depend right now on whether Moore’s Law will continue to describe the pace at which computing power escalates. The pace of computational advances has been expected to slow for some time, but such predictions have not yet come to fruition. Still, consensus is growing that computers that use silicon-based transistors are reaching their physical limits, and that new materials must be employed to continue achieving performance gains.
Stanford researchers have made significant progress on that front. They have successfully built a computer from carbon nanotubes. Carbon nanotubes are predicted eventually to enable smaller, faster, and more energy-efficient computers than silicon can. The computer the Stanford team unveiled earlier this week is the first Turing-complete computer to be built entirely from carbon nanotubes instead of silicon. A Turing-complete machine is one which can be programmed to solve any problem for which a solution actually exists. Simple calculators like your old Casio are not Turing-complete, because, although they are computers, they can only perform simple arithmetic. More modern calculators like the TI-84 and, of course, modern laptops and desktops are Turing-complete, in that they can execute programming statements in sequence, as a choice among multiple options, or repetitively until some condition is met. In other words, Turing-complete computers support sequence, selection, and repetition, the basic building blocks of all computer programs. This gives them the flexibility to solve any kind of problem that can be solved by hand. The new Stanford computer, although much larger and more limited than today’s sillicon-based computers, achieves this standard, and so it is viewed as a great proof-of-concept for future development. The Stanford group has proven that Turing completeness is not unique to silicon-based transistors and that alternatives like carbon nanotubes can achieve equivalent behavior. This is an impressive breakthrough.
More work needs to be done, of course. Most importantly, the manufacture of carbon transistors must advance significantly to achieve the same small scale and volume as current silicon transistors, of which about 4,000 can be spread across a human hair. Still, the achievement of Stanford’s research, as well as recent successes from IBM regarding the same nanotube-based architecture, suggest reason for continued interest in this technology. So, we can expect more silicon-free innovation to come.
There may very well come a day when we’ll be using carbon-based computers instead of ones based on silicon. And why stop there? Jack up the pressure a bit during production, and, before you know it, we’ll be computing with diamonds. Our computers will double as ear bling and cufflinks. I predict Apple will start selling these sometime in 2016. You should get in line now.