Critique of Nanocosm by William Atkinson
By Chris Phoenix, Director of Research, Center for Responsible Nanotechnology - June 2003
The Atkinson-Phoenix Nanotech Debate in detail.
William Atkinson has an agenda. This self-described "professional rhetoretician" (p. 134) has written a book on the emerging science of nanotechnology, with special emphasis on debunking. The reader will learn many details about science and nanoscience--and much of it will be wrong.
For example, Atkinson believes that in 10-15 years, "Maxwell Demons" will be developed (p. 277), enabling "heating, cooling, and material sorting, at zero energy cost." This is of course perpetual motion. The error is not merely sophomoric; it is glaring. Atkinson apparently believes this because someone told him that Cool Chips (pp. 76-79) trap heat on one side. Maxwell's Demon, in theory, does the same thing with no energy, and even Maxwell knew that this implied perpetual motion. What makes Cool Chips work is a constant stream of electrical power--a fact that Atkinson seems to have missed in his excitement. If it traps heat, then... then... it must be a Maxwell's Demon!
In addition to incomprehension of basic science, Atkinson does not check his facts. He claims that gallium arsenide is "almost universally found in microchips" (p. 193). In fact, gallium arsenide is a niche competitor to silicon and has no place in most microchips. The "fin" in FinFET transistors is not for cooling, as he claims (p. 283). Any computer scientist could have told him that his description of cellular automata (pp. 261-263) is not only confusing, but wrong: neither cellular automata nor cells are added to a running cellular automaton. He misspells "analyte" as "analete" repeatedly--even in its glossary entry. (*)
So why did he write a book on subjects he doesn't adequately research, in a field he apparently knows little about? Open the book to a random place, and you'll probably find the word "Drex." Apparently part of a professional rhetoretician's job is to make up childish nicknames for people he wants to criticize. Drex, of course, is Eric Drexler; "Merk" is Ralph Merkle. It's funny the first time. The 20th or 30th time, it's simply tiresome. But this is the purpose and focus of the book: debunking "Drex." Science, accuracy, and honesty are apparently less important.
Atkinson's treatment of Drexler crosses the line between rhetoric and intellectual dishonesty. He complains that Drexler's book Nanosystems discusses cooling systems (p. 130). In the very next paragraph, he asserts that Drexler believes that nanobots will produce no excess heat. What are the cooling systems for, then? Likewise, he gives a description of Drexler's "Stiff-Arm Nanomanipulator" design (p. 127), ending with the complaint that the design makes no concession to "the otherness of the nanocosm." But the whole point of Drexler's design--as reflected in its name--is that things must be built much stiffer at the nanocosm.
So the reader can't trust the book's basic science, or nanoscience, or descriptions of molecular nanotechnology. Even the anti-Drexler rhetoric is too heavy-handed to be useful. On page 131, under the heading "The Church of St. Drex" (see what I mean?) he refers to "the movement called Drexlerianism." Alas, that word doesn't exist, at least not on Google's three billion web pages. Atkinson seems to have made it up to support his assertion that a "movement" even exists. He spends the next few pages complaining that Ralph Merkle is a good speaker. He really, really wants you to believe that there's nothing to Drexler's work but hype. To make this point, he spends five pages (126-130) making fun of Nanosystems--for being too technical. "This is too smart for you, so it must be bogus," he says, in essence. "Move along, nothing to see here. Now let's laugh at 'Drex' and his 'movement' some more."
Atkinson's technical criticisms of Drexler's work are often demonstrably wrong, and the rest are questionable at best. Many of his criticisms are supported only by his own authority: for example, "staggeringly complicated--unworkably so, in my opinion" (p. 255). This particular criticism rests on the claim that nanobots would have to respond to picosecond events--which shows a basic misunderstanding of Drexler's work. Drexler's designs control the conditions so that they do not ever have to respond to rapid events. This leads to Atkinson's next criticism: that Drexler's designs require the complete subjugation of the natural world, and so can't work in practice. On page 256, he claims that "Drexlerians" refuse to consider using chemicals to pass messages between nanobots because "wet nanotech" is "anathema" to them. This is untrue. Nanomedicine, a thoroughly "Drexlerian" book, considers exactly that possibility in section 18.104.22.168: "Ideal Messenger Molecule."
Later on that page, Atkinson gives an absurd description of molecular manufacturing: a vision of humans who "think themselves atom-sized" in "control booths." Then he describes more advanced designs. It seems that nanoassemblers "would themselves choose how, where, and when to work." With onboard memory and planning, "The nanoassemblers would then, by any definition, be a species of living individuals." No nanotechnology expert claims that nanoassemblers, with or without memory, would be alive--that stretches logic far past the breaking point. It's hard to tell whether Atkinson is making this claim, or attributing it to Drexler. Either way, Atkinson's statement is incorrect.
A few of Atkinson's criticisms deserve an answer. On page 9, Atkinson claims, "to Drexlerian engineering, carbon is a disaster. The instant a nanomanipulator arm touched carbon, it would become as immobilized..." Many of Drexler's proposals involve the manipulation of carbon, so it is simply untrue that "carbon is a disaster" in "Drexlerian" engineering. But would the arm actually become immobilized? The science is new, and a little skepticism toward molecular manipulation may be appropriate. But it's worth noting that a few weeks ago, scientists in Japan used a nanomanipulator arm (an atomic force microscope) to remove an atom from a sticky, covalent silicon surface--and then put it back--and then scan the surface to verify their work.
But many of Atkinson's attacks on Drexler's work are simply not worth answering. On page 33, he makes the false claim, "Now, according to Eric Drexler and his ilk, nanotechnology will make us omnipotent." On page 69, he compares "Drexlerian" work to "belt sanders and escalators at the atomic level," and then to "a time machine out of tin cans and Plasticine." It gets worse. Page 145: "K. Eric Drex, K. Eric Drex /The man who dispensed with reality checks."
I am reminded of a saying of Mahatma Gandhi: "First they ignore you, then they laugh at you, then they fight you, then you win." Atkinson is trying to fight, but most of his attacks are little more than mean-spirited mockery. The strongest attack he can muster is a reference to Smalley's criticism of Drexler in Scientific American. Unfortunately for his case, this has been thoroughly answered in "A Debate About Assemblers," and "An Open Letter to Richard Smalley" by K. Eric Drexler.
To be fair, Atkinson spends a lot of the book reporting genuinely exciting advances in nanotech. If it were not for the scientific and technical errors scattered liberally through the book, I might recommend that people ignore the Drexler-bashing and read the book for the nanotech news. Anyway, better sources of nanotech news can be found online for free: for example, TNT Weekly or Nanogirl News.
A nonfiction technical book should be accurate. As Atkinson states in the Preface, one of his goals is to give the venture capitalist "a thorough briefing in the science and technology emerging from the nanocosm." Unfortunately, his science and technology explanations are too inaccurate to be useful; readers will come away badly misinformed.
(*) At this point the review originally read: "He gets brains wrong too: "White matter handles reflexes and intuition." (p. 42) In fact, gray matter contains the chunky part of nerve cells, and white matter contains the long skinny part of the same cells." But Atkinson was correctly reporting the statement of a primatologist, Roger Fouts. Fouts' usage is nonstandard, but Atkinson shouldn't be blamed for it.
Chris' interest in nanotechnology began when he took Eric Drexler's class "Nanotechnology and Exploratory Engineering" at Stanford University in 1988. He has followed the field continuously since then, attending numerous nanotech conferences, contributing frequently to several on-line discussion lists, and helping to review a major book and a Ph.D. thesis on nanotech. He is a Senior Associate of the Foresight Institute and co-moderator of the sci.nanotech newsgroup.
Other papers by Chris
See also these other reviews on Nanocosm:
Nanotech: Fact and Fiction