PUBLISHED BY THE NEW YORK ACADEMY OF SCIENCES · NOVEMBER/DECEMBER 2000

letters from readers

PEER REVIEW


Alan Magee, Luftpost, 1998

MINI MENACE?

I share Robert A. Freitas Jr.’s enthusiasm for the possibilities of nanomedicine [“Say ‘Ah!” July/August], but I also share specific concerns that Mr. Freitas has voiced elsewhere—though, surprisingly, not here. As I pointed out in April in an article in Wired magazine, “Why the Future Doesn’t Need Us,” nanotechnology poses a new and dangerous kind of risk, both because it is information-based (no hard-to-obtain raw materials or special production facilities will be required to practice it) and because nanomachines have the potential to replicate themselves.

K. Eric Drexler, chairman of the Foresight Institute, which studies the anticipated impact of advanced technologies, pointed out more than a decade ago in his book Engines of Creation that a runaway accident involving nanotechnology could have dire consequences—including the destruction of the entire biosphere. In addition, if access to the proper information and designs is not carefully restricted, violent individuals or small terrorist groups will be able to adopt nanotechnology to carry out large-scale destructive acts.

Mr. Freitas discussed that problem in a recent paper titled “Some Limits to Global Ecophagy by Biovorous Nanoreplicators, with Public Policy Recommendations,” which is available on the Web at www.foresight.org/NanoRev/Ecophagy.html As far as I know, that article by Mr. Freitas was the first detailed, published analysis of the so-called “gray goo” problem—the possibility that miniature, self-replicating robots could devour all the natural resources on earth, thus obliterating all life. Mr. Freitas was right to analyze the danger, which should give any sober person pause, but, regretfully, he failed to mention his analysis or these concerns in his article for The Sciences.

Thus, although I applaud Mr. Freitas for publicizing the enormous potential benefits of nanomedicine, I hope the editors of The Sciences will consider participating in a wider discussion of the ethical and safety issues raised by the new self-replicating technologies: genetic engineering, nanotechnology and robotics. We need to figure out how to reap the enormous benefits of those technologies while reducing their dangers—dangers that are far too likely to lead to catastrophe if we do not take steps to impose sensible limits.

Bill Joy
Sun Microsystems, Inc.
Aspen, Colorado

Robert Freitas writes: “To make the most complicated structures, . . . self-assembly is less than ideal.” I suggest that he look in the mirror. I can think of no structure more complicated than a human being, yet each of us has self-assembled.

Regarding the positional control that Mr. Freitas advocates as an alternative to self-assembly, it is important to recognize that the end-products of chemical reactions are under thermodynamic control. It may be possible to avoid kinetic traps or to produce metastable species through positional control, but thermodynamic considerations will direct the ultimate products of all syntheses. The main role I can envision for positional synthesis is to do chemistry at locations that are spatially distinct but chemically identical; such symmetric sites are found on the surfaces of crystals, carbon nanotubes and viruses.

Nadrian C. Seeman
New York University
New York, New York

 

Like the clock in Shakespeare’s Julius Caesar, there is an anachronism in Robert Freitas’s article—or perhaps it might more aptly be called a “nanachronism,” in view of his subject matter. Mr. Freitas describes a young man going to a physician in the year 2030 to reap the benefits of nanomedicine. But once the technology Mr. Freitas describes is viable, physicians will become largely obsolete. Such services may well be delivered via the Internet.

Nanomedicine will also have a profound impact on politics and the economy. Even if the delivery of nanomedicine does not reduce total health-care expenditures—which it surely would—it would free up billions of dollars that are now spent on premiums for private and public health-insurance programs.

Barry Kriegel
Atlanta, Georgia

 

Robert A. Freitas Jr. replies: I share many of Bill Joy’s concerns about the potential for abuse of nanotechnology-based replicators, and I join his call to reduce this danger. The main conclusion of my ecophagy paper, which he cites, was that the best defense against deliberate abuse is preparedness and vigilance. Neither one is possible if a ban on nanotechnology drives its development underground, or into the hands of irresponsible governments or secret organizations. Open development ensures that the same technology that causes a problem can also be employed in its solution, thus leveling the playing field.

As for my article in these pages, the specific topic was nanomedicine, not nanotechnology in general. There is a broad consensus that replicating systems belong in sealed factories, not in people’s bodies. Medical nanorobots, like aspirin pills, have no need to reproduce.

Nadrian Seeman is correct that self-assembly is a powerful manufacturing method that can make a wide range of structures, that it has so far demonstrated more capability at the molecular level than has positional assembly, and that it is likely to prove crucial in the development of nanotechnology. But self-assembly, by itself, does not make human beings. Critical to the construction of humans are ribosomes—programmable manufacturing systems that are present by the millions in each cell. Ribosomes make proteins by means of a limited form of positional assembly. Both the new amino acid and the growing polypeptide chain are positionally held, and the reaction between them takes place at a specific site because of that positional control.

Positional assembly is useful whenever reacting structures might have prematurely reacted elsewhere if they had not been positionally constrained. The use of highly reactive compounds enables the synthetic process to be both simple and direct, and that in turn makes for enhanced flexibility and allows a greatly increased range of structures to be synthesized. But attempting to synthesize various reactive components via self-assembly, even if the components are chem- ically distinct, must fail in the absence of positional constraints, because those components can collide with each other in undesired but still reactive combinations.

I agree with Barry Kriegel that costs may fall. But physicians and hospitals will remain in demand for cases involving acute trauma, emergency care, major surgery, rare disease conditions, significant cosmetic modifications and many other complex procedures in which professional guidance is essential. Often a disease condition can remain unrecognized by a patient until it is diagnosed by a physician. Trivial treatments can and will be delivered via the Web, but given the large numbers of VCRs blinking 12:00 across the land, I suspect new service occupations (e.g., “personal nanomedics”?) will arise to help fill the knowledge gap. 

 

 

WEIGHING THE ODDS

In his article about the evolutionist-creationist debate, Keith Devlin points out that one powerful argument often put forward against evolution is mathematical [“Snake Eyes in the Garden of Eden,” July/August]. “The probability that evolution by natural selection could have given rise to human beings,” as he cites the argument, “is simply too small to support credible belief.” But Mr. Devlin also cautions that “inferring design—or ruling it out—on the basis of mathematical probabilities alone is not a simple matter.” He is certainly right, but just because the task is not simple does not mean it is impossible or that it shouldn’t be attempted.

There is no lack of reasonable estimates for important parameters that might govern evolution. For example, mutation rates have been estimated for a variety of organisms, and theorists can model the speed with which beneficial or neutral mutations might spread through a population. Yet biologists have historically been reluctant to quantify the probabilities of postulated evolutionary events. The theory of evolution is backed by surprisingly few rigorous calculations.

“Just how small does a probability have to be before it is judged to be ‘too small’?” Mr. Devlin asks. A good question. Let’s work on getting good probability estimates so that we can begin to address it.

Michael J. Behe
Lehigh University
Bethlehem, Pennsylvania

 

One fact Keith Devlin neglected to mention—and which would have strengthened his argument substantially—is that evolution by natural selection is not solely a matter of chance. Although mutations are random, the ones that provide a competitive advantage to the organism possessing them will in time disseminate throughout the population. When that fact is understood, it becomes much easier to believe that life could have come into being without the intervention of a divine creator.

Creationists are not trying to convince scientists; they are trying to convince relatively uninformed laypeople.

Arthur Vibert
San Francisco, California

 

Keith Devlin is right to describe evolution as a series of cumulative events. Consider an analogy: A crack meanders across a road from edge to edge. A small vine begins to grow inside the crack at one edge of the road. As long as the road remains empty of traffic, the chances that the vine will follow the crack exactly all the way to the other side are very small. But if cars can travel along the road, the picture changes radically. Any exploratory shoot from the vine that lands on the roadbed will get squashed. Now the vine must grow inside the crack.

In the parable the growth of the vine is the evolutionary history of humanity, and the cars are natural selection. Humans made it to the other side of the road not because of design, but because all other evolutionary attempts died.

Michael J. Campa
Duke University Medical Center
Durham, North Carolina

 

Keith Devlin’s argument establishes a false opposition between randomness and design. As anyone familiar with human behavior in Las Vegas can testify, an agent capable of forming an intention may choose (deliberately, albeit foolishly) to rely on random means to implement that intention. When all is said and done, it is just possible that God does play dice with the universe.

The weakness of the religious hypothesis is that no empirical standard exists against which its claims can be weighed. By contrast, the theory of evolution puts forth a natural, non-miraculous scenario for the emergence of life and the diversification of species. Ultimately, it is not vague speculations about the probability of an unlikely event, but the judicious use of Occam’s razor, that will clip the beard of the demiurgic Jehovah.

Thomas A. Reisner
Université Laval
Quebec, Canada

 

Keith Devlin replies: My article “Snake Eyes in the Garden of Eden” was not intended to contribute directly to the evolution-versus-creation debate. Rather, as a mathematician, I set out to examine the creationists’ use of probabilistic arguments, in particular William A. Dembski’s design filter.

Michael Behe’s letter seems to suggest that I endorsed Dembski’s argument. That is not the case. I do, however, judge it of sufficient scientific importance to merit careful attention, which unfortunately is not the case for most of the rhetoric coming from the creationists’ camp.

Mr. Behe observes that “the theory of evolution is backed by surprisingly few rigorous calculations.” That is true, and I endorse his request for the development of such calculations. (I did in fact give one such calculation in my article, an example credited to the evolutionary biologist Richard Dawkins of the University of Cambridge.) In the spirit of a balanced debate between two camps, I would also call on the creationists to provide some rigorous calculations to support their own case.

Arthur Vibert takes me to task for not saying that evolution by natural selection is not merely a matter of chance. In fact, I spent some time explaining that it is not just a random process, but rather a cumulative process with random elements, as Michael Campa rightly observes in his letter.

I don’t have any problem with what Thomas Reisner says, except that the “randomness-versus-design” dichotomy was not part of my argument. Rather, it was a fundamental assumption of Dembski’s investigation, which is what I set out to examine.

 

Back to the top


© 2000 New York Academy of Sciences, All rights reserved.