Eric Drexler 1995

THE INCREDIBLE SHRINKING WORLD OF ERIC DREXLER
An interview with K. Eric Drexler, The Foresight Institute
By Anthony B. Perkins
From August 1995 issue

Eric Drexler currently heads the Palo Alto-based, non-profit organization called The Foresight Institute that, in his words, "gathers and distributes information and holds conferences on technologies that are clearly going to have a large impact on society, with a particular focus on nanotechnology." This makes sense, since Mr. Drexler was the first to introduce the word nanotechnology to the English language in 1981 when he was still a graduate student at MIT. Nanotechnology, also known as advanced molecular manufacturing, is the science of creating machines at an atomic level. "Somebody recently completed a computer-based study on the use of the term nanotechnology, and the results followed a pretty steep exponential curve after the publication of my first book. I think the curve is getting even steeper as we speak," Mr. Drexler told The Herring at lunch recently. People generally take one of two distinct views on nanotechnology. Some view it as a serious scientific endeavor that could lead to the creation of atomically precise manufacturing systems, where programmed atoms serve as building blocks, and productivity and data through-put could conceivably increase a trillion-fold. According to this school of thought, nanotechnology could provide, among other things, a personal computer that will pack a trillion transistors, with a current CPU's power in every transistor. Others view nanotechnology as a novelty that a few scientists are a little over-enthusiastic about because it works on a nanometer scale. And scientists all seem to love really, really small, or really, really big things.

Whether the nanorevolution will occur is a moot question. But we do know from a recent conversation with Intel founder Gordon Moore, that the photolithography technology used to build today's semiconductors will ultimately hit the wall, and the number of transistors you can fit on a single chip will reach its limit. What then? Let us introduce you to Eric Drexler--he just might have the answer.

The Herring: So what turned you on to nanotechnology, anyway?

Drexler: From my perspective, the idea for nanotechnology came out of looking at biochemistry and molecular biology. I was a student at MIT back in the 1970s, and doing research on building manufacturing systems from scratch in outer space. When I studied molecular biology and biochemistry, I could see that people were, indeed, using molecules to make a lot of things, including the copying of nature. Workers in the field were basically taking raw material, and putting it together, and building what they referred to as "molecular machines." And like a good engineer I asked: "What will we be able to do when we get good at designing and building these things?" And that led me into building molecular machine systems, and better molecular machine systems, and even better molecular machine systems, and wondering where this was all going to end up! I started to imagine what advanced systems would look like, and what kind of new industrial technologies might result.

The Herring: Were you the only one around at the time that was into this kind of stuff? Or were there others out there working in this area?

Drexler: In completing the research for the first article I wrote on this subject, I found a transcript of a speech that Dr. Richard Feynman delivered to the American Physical Society back in 1959, called There's Plenty of Room at the Bottom, in which he was basically pointing in the same direction I was. Anyway, Dr. Feynman talked about miniaturization in general, foreshadowing many of the later developments in semiconductors, and more recent developments in micromachining, and concluded by saying that there was no reason why, in principal, the miniaturization of the manufacturing process couldn't be taken down to the atomic scale.

The Herring: So what kind of degree did they give you for traveling in this kind of space?

Drexler: My undergraduate major at MIT was interdisciplinary science, and I hung out in the labs long enough to finally get a Masters and Doctorate in molecular nanotechnology as well. So I went through the whole mill. [Chortles]

The Herring: And since you graduated in 1981, you have been cranking out all sorts of academic papers and books full of hundreds of equations. Right?

Drexler: Well, I have pumped out some big-time equations, and put them into books that no one reads, and spoken at conferences full of other brains like me...and, shucks, it may be kind of weird, but how many people do you know who do that for a living?

The Herring: Great to hear someone can make a living that way. But what have the primary motivations behind your work been?

Drexler: The books and the papers I've written have been based on the general thesis that we are going to be developing a new technology based upon molecular machines that is going to be very clean, and very productive. It will usher in a huge leap in computational capacity that will be comparable to the advance that separates hand-cranked data machines from supercomputers. My main goal in research is to present as clear a picture as I can of this future technological opportunity, and to discuss its potential economic, social and strategic consequences in the coming decades.

The Herring: What was the thesis of your first book, Engines of Creation?

Drexler: The book's aim was to first, make a technical argument, and also to be accessible to a general readership. My recent book, Nanosystems: Molecular Machinery, Manufacturing and Computation, is a technical monograph that is being used as a textbook in some places; it's full of equations, diagrams, graphs, and layers and layers of technical arguments and reasoning. I was able to add a lot more detail to the whole debate.

The Herring: What part of your work has been the most controversial?

Drexler: Well, the fact that nanotechnology is a nice-sounding word that can be used to excite funding sources to provide money for research has, of course, been embraced with great vigor. [Laughs] The general notion that technology is heading toward atomic precision, and that this precision will ultimately be achieved through some form of mechanical manipulation of individual atoms and molecules was slower to take off than some of us thought, which fueled some cynicism. The serious controversies surrounding nanotechnology concern how long the development will take, how much it will cost, who will get there first, and what we should do about it. The spurious controversies have to do with whether it makes physical sense or not. And I don't know of any informed person who would argue that it doesn't make sense. This science is really taking off. There is a growing community of people who view nanotechnology as a critical component to any study of science in the future.

The Herring: What universities have programs in nanotechnology?

Drexler: There is a program at Rice University, and the Beckman Institute at the University of Illinois is also working on a program of research that I think is substantially pointed in this direction. There's also work being done at Stanford and Cal Tech, and the Japanese are focusing heavily on this area.

The Herring: We are kind of afraid to ask, but what might some of the social consequences be of atomically precise machines?

Drexler: What we are talking about here is replacing our current physical technology for manufacturing with a whole new set of materials and devices that have far superior computational capacity and power. That is a big deal! The closest thing we have seen to this type of transformation was the industrial revolution, and that comparison somehow doesn't seem adequate. It's very hard to imagine a future that will involve that much change--it's exhausting, so the common reaction is to say: "This is indigestible, therefore I am not going to digest it!"

The Herring: What effect will nanotechnology have on the semiconductor industry, for instance?

Drexler: In the computational universe, the efficiencies gained will be fairly smooth and uniform in two ways. First, in addition to achieving atomic resolution, nanotechnology is inherently built in three dimensions. So far, we have been gaining increased density on a plane where you have a million transistors, a thousand by a thousand. In the third dimension, you'd have another pack of a thousand transistors immediately, so if you extend this to the point where you have even more features to a side, and then get to the third dimension, you experience an even larger step function. The second point is that today's computer devices are extremely expensive on a per-pound basis. If you take the active computational material on a chip and the silicon that supports it, and price that out on a dollar-per-pound basis it is outrageous. It appears that the cost of molecular manufacturing--minus licensing fees, insurance, and all those normal cost-of-business expenses--is comparable to the cost of creating plastic or industrial chemicals, that is to say, ten cents per pound.

The Herring: What does the molecular manufacturing process look like?

Drexler: I can actually point out a process that's making a greater number of devices than Silicon Valley makes in a year, right now, and that is the growth of that tree. [He points to a nearby redwood tree through the window.] That tree is successfully making devices that use the light, and convert this light into electronic energy and move it around to a position where it can turn into chemical energy. And the inputs are not sand and aluminum, but air, water, and some light--all common to the biosphere. This is all very cheap because it's not bulky, it doesn't consume a lot of energy, and the capital equipment that it consists of can be made using the same process. No matter how productive a semiconductor plant is, you can't use semiconductor manufacturing equipment to make semiconductor manufacturing equipment, so it is a process that is always dependent upon other technologies. Molecular manufacturing systems will be built by molecular manufacturing systems. This is a very basic difference in economic structure between the old and future processes.

The Herring: How powerful will my Macintosh be after this is all over?

Drexler: You will have over a trillion transistors, and depending on the power consumption you are willing to expend, you could have power in the order of the present computational capacity of all the semiconductor capacity in the world together--basically a supercomputer behind every pixel on your screen. It will be something that you would have to keep cool with a fan.

The Herring: When do you think that nanotechnology will really take off?

Drexler: When you have molecular machine systems that will be able to take molecules and put them where you want them. In conventional macroscopic manufacturing, a person or a robot moves and assembles the parts. In semiconductor manufacturing, if you want to put a feature on a chip, you have a mask and some optical system that puts the feature in the designated place. In chemistry, if you want to put two things together in a particular way, you have to try to find two things that want to go together and put them into a solution and shake, and hope that will do what you want. This process has been amazingly successful, but terribly limited. If you add back in the physical processes that make molecules rearrange, combine, and build up new parts, you can actually hold these parts and assemble them, much like we see in factories today. This type of molecular manipulation is so foreign to chemistry that it disorients most chemists.

The Herring: Why?

Drexler: Because all their theories about how molecules wander around freely in liquid aren't true if you get a hold of the molecules and put them where you want them. All of the fundamental scientific principles are the same, but from the engineering and systems perspective, it is qualitatively very different. So, if you are able to build molecular systems like what I've been talking about, you will be able to build very complex, precise structures that can serve as parts for the molecular machines. And the characteristic speed of operation for these machines will be something on the order of a billion cycles per second, which will lead to enormous data through-put and higher productivity. At this point you could design computers in such a precise way that every atom is carefully placed and optimized in all three dimensions.

The Herring: And that's how you get your trillion transistors into my Powerbook...

Drexler: Well, you would at least be able to put the equivalent of a modern-day CPU in each of your transistors.

The Herring: And what is standing between us and this brave new world?

Drexler: A lot of hard work, and an unknown amount of money. Depends on how silly you are.