Principal Fellow, Zyvex

Originally published in the January 2001 issue of IEEE Spectrum. Published on KurzweilAI.net April 27, 2001.

Nanotechnology will make us healthy and wealthy though not necessarily wise. In a few decades, this emerging manufacturing technology will let us inexpensively arrange atoms and molecules in most of the ways permitted by physical law. It will let us make supercomputers that fit on the head of a pin and fleets of medical nanobots smaller than a human cell able to eliminate cancer, infections, clogged arteries and even old age. People will look back on this era with the same feelings we have toward medieval times — when technology was primitive and almost everyone lived in poverty and died young.

Besides computers billions of times more powerful than today’s, and new medical capabilities that will heal and cure in cases that are now viewed as utterly hopeless, this new and very precise way of fabricating products will also eliminate the pollution from current manufacturing methods. Molecular manufacturing will make exactly what it is supposed to make, no more and no less, and therefore won’t make pollutants.

When nanotechnology pioneer Eric Drexler first dared to publish this vision back in the early 1980s the response was skeptical, at best. It seemed too good to be true, and many scientists pronounced the whole thing impossible. But the laws of physics care little for either our hopes or our fears, and subsequent analysis kept returning the same answer: it will take time, but it is not only possible but almost unavoidable. The progress of technology around the world has already given us more precise, less expensive manufacturing technologies that can make an unprecedented diversity of new products. Nowhere is this more evident than in computer hardware: computational power has increased exponentially while the finest feature sizes have steadily shrunk into the deep submicron range. Extrapolating these remarkably regular trends it seems clear where we’re headed: molecular computers with billions upon billions of molecular switches made by the pound. And if we can arrange atoms into molecular computers, why not a whole range of other molecularly precise products?

It has taken decades for the bulk of the research community to accept the feasibility of this vision. But when the President of the United States in January 2000 called for a $500 million National Nanotechnology Initiative we knew nanotechnology had reached critical mass.

Some people have recently, publicly (and belatedly) realized that nanotechnology might create new concerns that we should address. Any powerful technology can be used to do great harm as well as great good. If the vision of nanotechnology sketched earlier is even partly right, we are in for some major changes — as big as the changes ushered in by the industrial revolution, if not bigger. How should we deal with these changes? What policies should we adopt during the development and deployment of nanotechnology?

Drexler discussed these issues extensively in his 1986 book Engines of Creation, and in a remarkably prescient essay first published in 1988 outlined some of the major issues and positions that have since come to the fore. One response, proposed by Bill Joy, cofounder and Chief Scientist of Sun Microsystems, would be to “relinquish” research and development of nanotechnology to avoid any possible adverse consequences.

This approach suffers from major problems: telling researchers not to research nanotechnology and companies not to build it when there are vast fortunes to be made, glory to be won, and national strategic interests at stake either won’t work, or will push research underground where it can’t be regulated while depriving anyone who actually obeys the ban of the many benefits nanotechnology offers.

If a ban won’t work, how should we best address the concerns that have been raised? The major concerns fall into two classes: deliberate abuse and accidents.

Deliberate abuse, the misuse of a technology by some small group or nation to cause great harm, is best prevented by measures based on a clear understanding of that technology. Nanotechnology could, in the future, be used to rapidly identify and block attacks. Distributed surveillance systems could quickly identify arms buildups and offensive weapons deployments; while lighter, stronger and smarter materials controlled by powerful molecular computers would let us make radically improved versions of existing weapons able to respond to such threats. Replicating manufacturing systems could rapidly churn out the needed defenses in huge quantities. Such systems are best developed by continuing a vigorous program of research and development, which provides a clear understanding of both the potential threats and the countermeasures available.

Besides deliberate attacks, the other concern is that a self-replicating molecular machine could replicate unchecked, converting most of the biosphere into copies of itself.

While nanotechnology does propose to use replication (to reduce manufacturing costs to a minimum), it does not propose to copy living systems. Living systems are wonderfully adaptable and can survive in a complex natural environment. Instead, nanotechnology proposes to build molecular machine systems that are similar to small versions of what you might find in today’s modern factories. Robotic arms shrunk to submicron size should be able to pick up and assemble molecular parts like their large cousins in factories around the world pick up and assemble nuts and bolts.

Unfortunately, our intuitions about replicating systems can be led seriously astray by a simple fact: the only replicating systems most of us are familiar with are biological self-replicating systems. We automatically assume that nanotechnological replicating systems will be similar when, in fact, nothing could be further from the truth. The machines people make bear little resemblance to living systems, and molecular manufacturing systems are likely to be just as dissimilar.

An illustration of the vast gulf between self-replicating biological systems and the kind of replicating robotic systems that might be made for manufacturing purposes is exponential assembly; currently under investigation at our company, Zyvex Corporation, in Richardson, Texas. Zyvex is developing positional assembly systems at the micron, sub-micron and molecular scale. At the micron scale, using existing MEMS technology, we are developing simple “pick-and-place” robotic arms that can pick up relatively complex, planar, micron scale parts made with lithographic technology; and assemble those planar parts into simple three dimensional robotic arms that have the ability to pick up specially designed MEMS parts. Called exponential assembly, this replicative technology starts with a single robotic arm on a wafer which then assembles more robotic arms on a facing wafer by picking up parts already laid out in precisely known locations.

While the number of assembled robotic arms can increase exponentially (up to some limit imposed by the manufacturing system), this assembly process requires (among other things) lithographically produced parts as well as externally provided power and computer control signals to coordinate the complex motions of the robotic arms. Cut off from power, control signals and parts, a micron-sized robotic arm would function about as well as one of its larger cousins taken from one of today’s automated assembly lines that was dropped into the middle of a forest.

To avoid any possible risk from future (and perhaps more ambitious) systems, the Palo Alto based non-profit Foresight Institute (motto: preparing for nanotechnology) has written a set of draft guidelines to inform developers and manufacturers of molecular manufacturing systems how to develop them safely. The guidelines include such common-sense principles as: artificial replicators must not be capable of replication in a natural, uncontrolled environment; they must have an absolute dependence on an artificial fuel source or artificial components not found in nature; they must use appropriate error detection codes and encryption to prevent unintended alterations in their blue prints; and the like.

Building on over a decade of discussions of a very wide range of scenarios, the first version of the guidelines were based on a February 1999 workshop in Monterey, Calif. They have since been reviewed at two subsequent Foresight conferences. Because our understanding of this developing technology is and will continue to evolve, the guidelines will evolve with them representing our best understanding of how to ensure the safe development of nanotechnology.

Nanotechnology’s potential to improve the human condition is staggering: we would be shirking our duty to future generations if we did not responsibly develop it.

Copyright © 2001 IEEE. Reprinted from the January 2001 issue of IEEE Spectrum