An Interview with Nanotechnologist K. Eric Drexler

By Michael Berry

K. Eric Drexler thinks small. Incredibly small. So small you'd need one of the world's most powerful microscopes to see what he's been thinking about lately.

An interdisciplinary scientist who now resides in Los Altos, Drexler spends most of his time ruminating on nanomachines, infinitesimally tiny arrangements of atoms hypothetically capable of performing all sorts of complex jobs. If Drexler's theories are correct, it's almost an understatement to say that his miniscule "engines ofcreation," as he calls them, woul have a gigantic impact on the human condition.

For example, they could make precisely engineered automobiles as cheap per pound as potatoes. Or patch the hole in the ozone and strip the excess carbon dioxide out of the atmosphere. Or grow human body parts from scratch. Or single-handedly turn the planet into a big, uninhabitable ball of gray goo.

Welcome to the amazing -- if still largely theoretical -- world of nanotechnology.

Born in Oakland in 1955 and raised in Oregon, Eric Drexler first started thinking seriously about molecular machines in 1977, during his senior year at MIT. While rooting around in the campus library, he read up on the latest advances in biology, chemistry, physics and computer science. What interested him most were the small-scale applications, the techniques for manipulating very tiny amounts of matter, rather than the bulk quantities with which most engineers and scientists deal.

In fact, Drexler's goal was to work with things whose dimensions could be measured in nanometers, billionths of a meter. Seeing how biotechnicians were learning to alter individual cells, bacteria and viruses by editing their DNA, Drexler began to wonder whether something similar couldn't be done with non-living material.

In 1981, he published a paper on molecular engineering in the Proceedings of the National Academy of Sciences. The paper generated a small amount of interest, enough to spur Drexler on to write "Engines of Creation," a full-length examination of the possibilities of nanotechnology, published in 1986 by Doubleday.

The logic of "Engines of Creation" runs something like this. Picture an atom as a physical object the size of a marble. A fairly complex molecule would then be a clump of linked atoms about the size of a fist. Depending upon the chemical properties of the various atoms and how their bonds "snap" and "unsnap", the molecule could be shaped in a way similar to how kids sculpt working toys from an Erector Set. The molecule could take on the shape and function of tools familiar to us in the visible world -- levers, motors, gearsand the like.

Suppose you could somehow fashion a sturdy, submicroscopic device from carbon atoms, for example, equipped with a robotic arm controlled by a computer. This "assembler" would then be able to push around atoms and place them precisely where it wanted them.

Drexler speaks softly and carefully as he explains the implications of these proposed assemblers. "In biology, enzymes stick molecules together in new patterns all the time. In every cell, there are programmable, numerically controlled machine tools that manufacture protein molecules according to genetic instructions. If you can build one molecular machine, then you can use it to build an even better molecular machine.

Although we presently lack the means to make them follow our precise bidding, proteins may prove to be the starting point for creating these inorganic assemblers. Where might increasingly better and smaller molecular machines lead? Drexler says, "The conclusion I came to was that the process leads to the thorough and inexpensive control of the structure of matter."

What Drexler is talking about sounds like alchemy but isn't. Nanotechnology will never allow anyone to transmute lead into gold. However, it could conceivably be used to change a lump of coal into a diamond. All the essential ingredients are present in the coal; the carbon atoms only need to be put in the proper arrangement to produce a gem to rival the Kohinoor.

Drexler's nanomachines would operate essentially as their large-scale counterparts do. Their moving parts, however, would be fashioned from a small number of atoms and held together by the power of their atomic bonds.

"The really big difference is that what you make with a molecular machine can be completely precise, down to the tiniest degree of detail that can exist in the world," Drexler says. "And that because the moving parts are a million times smaller than the ones we're familiar with, they move a million times faster, just as a smaller tuning fork produces a higher pitch than a large one. On the molecular scale, you find it's reasonable to have a machine that does a million steps per second, a mechanical system that works at computer speeds."

Granted, a single nanomachine would be incredibly fast and precise. But due to its size, it would take a long time for it to do anything of use to the average human being. That's why you need a whole bunch of them operating at once. And the best way to get a whole bunch is to teach the machines how to replicate themselves.

Drexler explains, "If you take all the factories in the world today, they could make all the parts necessary to build more factories like themselves. So, in a sense, we have a self-replicating industrial system today, but it would take a tremendous effort to copy what we already have."

A single, specially-designed molecular machine, however, could copy itself quickly by using simple fuel and raw materials supplied in a vat of "replicator cocktail." Its twin could do the same thing and, as their numbers increased exponentially, the troop of nanomachines would soon be ready to turn its attention to more complex operations, such as manufacturing objects like chairs, boats, guns and rocket engines. As long as they had sufficient fuel and raw materials, the nanomachines could make almost anything imaginable, with minimal labor and energy costs.

"With molecular manufacturing, the process can be put into a smaller, faster package," says Drexler. "The basic parts, the start-up molecules, can be supplied in abundance and don't have to be made by some elaborate process. That immediately makes things simpler. The other advantage is that in conventional manufacturing processes, it takes a long time for a factory to produce an amount of product equal to its own weight. With molecular machines, the time required would be something more like a minute."

The products manufactured by nanomachines would not only be cheap and plentiful, but precision-crafted as well. Free of the microscopic defects that presently plague all man-made goods, they would be designed to be extremely durable. Some would be improved beyond all current expectation.

Take computers. Drexler has a proposal for building nanocomputers that make today's most powerful Cray look like an abacus. They wouldn't, however, work electronically. Like the steam-driven calculating machine proposed by Charles Babbage in the 1800s, they'd have lots of moving, mechanical parts.

"Computers are based on switches that turn other switches on or off," says Drexler. "In electronics, current flows through a conductor and a transistor has the opportunity to block or unblock it.

"In a mechanical system, you can have a rod that will slide or not slide, and when it is in one position, it blocks or unblocks another rod. It turns out that it's just as good at representing mathematics and logic as electrons are."

According to Drexler, mechanical nanocomputers would be moderately faster than today's electronic ones. Because the information within them travels across such miniscule distances, nanocomputers would be able to handle a billion instructions per second. He also predicts that they would be small enough so that an entire central processing unit would fit into a transistor on one of today's microchips, a trillion processors into a desktop computer. With such vast memory storage capabilities, the contents of the Library of Congress could fit into the speck smaller than a grain of sand.

Drexler says, "Eventually the whole integrated circuit technology base is going to be replaced."

One area in which Drexler sees these super-smart mechanical computers playing a vital role is medicine. Loaded with software bestowing upon them artificial intelligence, nanomachines would bring undreamed-of health benefits. Surgical precision would be brought down to the molecular level.

Drexler says, "Today we have big, crude instruments guided by intelligent surgeons, and we have little, stupid molecules of drugs that get dumped into the body, diffuse around and interfere with things as best they can. At present, medicine is unable to heal anything. It can help the body heal itself, but it can't, for example, take an incision and join the two sides together in such a way that the wound is healed as soon as it is closed."

According to Drexler, the simplest applications will be those which depend on the selective destruction of things that shouldn't be in the body. "The immune system may not be able to recognize that a cancer cell is abnormal, because it's not able to look at enough different characteristics to identify the cell properly. If injected into the body, a nanomachine with an on-board computer could look at 20 different characteristics of a cell before doing anything to it."

Maybe, just maybe, nanotechnology can even offer added hope to those folks who are planning to have themselves frozen for posterity, the cryonics enthusiasts. Until Drexler and his ideas about nanotechnology came along, no one had a clue how to effect the obviously massive corrective surgery required to reanimate a human body after it's been bobbing around in liquid nitrogen for centuries. The job would require programming skills far beyond the ken of present-day software engineers, but Drexler thinks that molecular machines might eventually be made smart and agile enough to get the job done.

In fact, when the Alcor Life Extension Foundation, a cryonics lab in Riverside, ran afoul of the law a few years ago by allegedly removing and freezing a client's head before she was declared legally dead, Drexler supplied a deposition in their defense. Since he was lecturing on nanotechnology at Standford University at the time, his "technical declaration" made the goings-on at Alcor seem a little less macabre.

Lest one get the impression that molecular machines will automatically cure all the world's ills, it's important to recognize their dark side as well. Drexler certainly does, and that's part of the reason for the four-year lag between his dreaming up nanotech and finally publishing his ideas.

He says, "Some of the consequences of the potential abuse of this technology frankly scared the hell out of me. I wasn't sure I wanted to talk about it publicly. After a while, though, I realized that the technology was headed in a certain direction whether people were paying attention to the long-term consequences or not. It then made sense to publish and to become more active in developing these ideas further."

What's so scary about nanotech falling into the wrong hands? For one thing, there's the "gray goo " problem. Suppose some madman engineers a molecular machine that can reduce _all_ matter down to its basic components. The final result wouldn't necessarily have to be either gray or gooey, but everything from trees to people, from buildings to mountains, might wind up as one tremendous, amorphous lump of dissolved, dead stuff.

Of course, that's a worst-case scenario. But nanotech also has more limited, though no less insidious, applications as an agent of chaos. Not only could it manufacture better, smarter conventional weapons like guns, tanks and missiles, it would make biological weapons much more attractive to certain folks. Instead of killing the enemy willy-nilly, you could program a deadly virus to kill only those with certain skin colors or ethnic backgrounds.

"Any powerful technology can be abused," says Drexler. "A technology that can make large amounts of high-quality products quickly and inexpensively, can be used to do just that for military products. My greatest concern is that the emergence of this technology without the appropriate public attention and international controls could lead to an unstable arms race."

Drexler, however, doesn't think we need be too concerned about catastrophic accidents involving runaway nanomachines. "One misconception is that anything small that can construct more molecular things is going to be alive, a flexible, capable entity that might get loose. Confusing a nanomachine with a bacterium is like confusing a radio-controlled model car with a rat."

Drexler finds another automotive metaphor to describe the chances of stray nanomachines running wild. "It's hard enough to design, build and maintain an automobile that runs when you give it carefully refined gasoline and oil. It would be much, much harder to build one that could go out into the forest and refuel itself from tree sap. To make something that can 'get loose' would be a very difficult engineering challenge."

Speaking of engineering challenges, once you get used to thinking about nanotech's promises, it's possible to lose sight of the fact that, for the most part, these machines still exist only in Eric Drexler's fertile imagination. Pushing aside all the fascinating speculation, can these things be built and will they actually work?

Drexler contends that the answer lies in what we can already easily observe. "We know the basic principles of molecular machinery will work because they do work," he says. "Some bacteria can swim because they have reversible, variable-speed motors mounted on their cell walls which turn a helical rod of protein that works as a propeller. That's a molecular machine.

"We know it is possible to store a billion bits of data in a few cubic microns because that's how much data is stored n digital tapes of DNA in the nucleus of each cell in your body."

Some members of the scientific establishment do indeed think that Drexler is on to something important. For example, John G. Cramer, a professor of physics at the University of Washington in Seattle, says, "When I read 'Engines of Creation,' my basic reaction was, 'Of course.' It's the sort of thing I had seen looming on the horizon for a long time. On the other hand, it's not clear how long it's going to take for there to be signifant societal impacts from the things Eric is talking about."

Cramer adds, "Predicting the future is always a risky business, but I think Eric has done as good a job as most people who have tried similar things."

In fact, it's a lot easier to find people who'll speak in support of Drexler than it is to find anyone willing to criticize his work publicly. That doesn't mean there are no dissenting opinions. Rather, some scientists don't take Drexler at all seriously, viewing nanotechnology as sheer science fiction. Consequently, they simply don't want to be involved in the debate.

Simson Garfinkel, a doctoral candidate at MIT's Media Lab with undergraduate degrees in chemistry and the history of technology, did take Drexler to task in the Summer 1990 issue of Whole Earth Review. Although Garfinkel refers to Drexler as "a visionary," one thing that bothers him is that Drexler has been, so far, vague on the nitty-gritty details of how to build his molecular machines.

Among Garfinkel's objections to Drexler's proposals is that atoms and molecule really don't snap together like itsy-bitsy Tinker Toys, that the rigidity a nanomachine needs would be extremely difficult to achieve. He believes the gears and levers in nanocomputers, for example, might be too limp and unstable to do their jobs.

"Drexler says that things will be possible. It's hard to argue against that," says Garfinkel. "But he doesn't put a proposal on the table of how to do X and Y. He doesn't know. Nobody knows."

Drexler, on the other hand, claims to be frustrated by what he perceives as "consistently off-target and empty criticism." He says, "I've encountered a lot of people who sound like critics but very few who have substantive criticisms. There is a lot of skepticism, but it seems to be more a matter of inertia than it is of people having some real reason for thinking something else."

Regarding the stiffness of gears and levers in nanomachines, Drexler responds. "I've done the calculations, using the standard computer models. These things will work."

According to Drexler, who made an eight-day speaking tour of Japan last spring, the Japanese are far more receptive to the concepts of nanotechnology than are researchers in this country.

He says, "In basic technology, it's not clear that they're very far ahead of us. In terms of organization and commitment, having identified the right goals and pulled together interdisciplinary teams to work toward them, they're about five years ahead.

"We haven't really begun yet. You can find academic and industrial groups doing some relevant work, but there isn't a focus on building complex molecular systems. In that respect, Japan is first, Europe is second, and we're third."

To promote discussion about nanotechnology and other future technologies, Drexler and his wife, Chris Peterson, have founded the Foresight Institute. Despite its imposing title, the Institute consists mostly of a few phone lines and home offices scattered across the Peninsula, but it does produce an informative newsletter, Foresight Update. It also co-sponsored the well-attended First Foresight Conference on Nanotechnology, held at Stanford in October, 1989.

The next conference in slated for this fall. A second book for the general reader on the implications of nanotechnology is nearly complete, and Drexler is hard at work on his long-awaited technical treatise on the subject.

Whether or not everything Drexler forsees comes to pass, nanotechnology is capturing the interest of a wide variety of professionals and laypeople, providing a springboard for serious discussion of how we might live our lives ten, fifty or more than a hundred years from now. At a time when so many predictions about the future are dire, it is gratifying to contemplate a technology that, fraught though it is with dangers, offers some hope of the miraculous.

"The kind of transition that lies ahead will be very painful, very disruptive in many different ways," says Drexler. "But if we can manage it so people don't have things forced on them that they don't want, I think there's every reason to believe things can settle out in a situation that is recognizably better than the one we're stuck in today."

(c) 1991 by Michael Berry