I attended a talk by Don Eigler, an IBM Fellow today; one of the things it reminded me of (other than Science is Cool) was how much I’ve come to loathe the notion of the “ideas man”.
Dr Eigler, twenty years ago, became the first person to assemble something by moving an atom at a time; his most recent work is on a potential breakthough in replacing conventional silicon transistor techniques: he and his colleagues take carbon monoxide molecules and assemble them by standing them up in the (to use his metaphor) egg carton shaped surface of a slice of copper; by building a particular array of them, they can create a logic gate than provides all the logic and storage functions that can be produced with silicon transistors. Very, very interesting; and very impressive—this technique allows them to fit a single logic element that can produce AND, OR, and majority logic in a 12 nm by 17 nm space (for refence, the current smallest commercial silicon is around 45 nm for the smallest trace). If they’re successful in solving a variety of outstanding hard problems (this is, after all, basic research, not engineering), then it has the potential to offer as much as five order of magnitude improvements in computing density and power consumption over existing circutry.
Possibly the coolest thing about the talk, though, was that Don started out by explaining the logic gate in terms of dominoes, discussing how you could use domino fans to create AND and OR gates. I thought this was a metaphor, but no: the carbon monoxide molecules, as I mentioned earlier, are stood upright on the carbon, one atom atop another. If they’re in a pair, they form a stable, upright construction. Add one, and the tower falls over, knocking down the next tower, and the next, and the next, and so on, until your atom-sized dominos have collapsed into the result you want. Dominoes aren’t a metaphor, dominoes is exactly what they’re doing—tiny poisonous dominoes.
(Star Trek fans would doubtless be delighted to know that Dr Eigler could, when discussing real-world use of left-spin molecules to create cloaking devices, referr by name to the Federation-Romulan treaty which prevents the Federation from utilising cloaking technologies.)
In 15 to 20 years, you might be able to buy something based on it. If everything goes well.
Which takes me to the point I opened with. Ideas were tossed around about the sort of things you could do with this kind of improvement in computing densities: since we’ve had demos of using human cell processes to power electronics, a five order of magnitude improvement in power/computational density could presumably make meaningful (by today’s standards) computers be something that can be embedded in the skin and powered by doughnuts.
Ideas are cheap. Ideas are easy. The kind of person who says, “I don’t do these things, I’m an ideas man” is not only most often a total waste of space—worse in fact, since the injection of “ideas” usually wastes the time of people with work to do—but is perversely proud of doing nothing, as though producing nothing more than a few suggestions places one above the hoi-polloi.
But here’s the thing: Don Eigler was making major advances in the science of the very small twenty years ago, is still doing it today, and could easily work on solving the same group of problems for decades longer. His advances are a result of doing work-hard work that takes time, knowledge, and skill. A room full of moderately intelligent people who work in technical professions can listen to his lucid, articulate explanations of the principles of his work and feel that we understand it but, really, we wouldn’t even know where to begin to actually reproduce it, still less build on it.
Ideas are cheap. Results are gold.
