On April 23 2001, the London based Financial Times published an article by Fiona Harvey about nanotechnology. The byline was:
INSIDE TRACK: A miniature world of tubes, balls and locusts: NANOTECHNOLOGY: Scientists are learning how to build tiny machines a few atoms wide. Industrial applications may be a decade or more away but progress has been astonishing, says Fiona Harvey.The URL is: http://globalarchive.ft.com/globalarchive/articles.html?id=010423000540
This critique is an attempt to correct misconceptions or errors in the article in the hope that it will find its way onto the web so that future reporters will "get it right".
"Invisible machines with components only a few atoms thick are swarming over a lump of coal."This depends what one means by "components". Most envisioned components in diamondoid machinery have to be much thicker than a "few atoms". To obtain the necessary positional accuracy in assembling materials atom by atom, envisioned assembler arms containing ~4 million atoms have to be ~30 nanometers (nm) wide. If made out of carbon atoms, this diameter would translate to ~2000 atoms across. Only sub-components such as gears or bearings (like the single-walled buckytubes the article mentions) might be as thin as a few atoms or even single atoms.
"Nanotechnology has its roots in 1985, when two US scientists, Richard Smalley and Robert Curl, and Britain's Sir Harold Kroto discovered a new molecular formation for carbon."Actually, Nanotechnology was first envisioned in a concrete way by the Nobel Prize winner in 1959 by Richard Feynman [Fey59]. It was brought to scientific attention again by Eric Drexler in [Dre81]. The public was made aware of some of the consequences by his book Engines of Creation in 1986.
"These balls and tubes have been created in laboratories in ways that suggest they could be made to grow themselves. That holds out the promise of manufacturing tiny machines that would form themselves into the desired shape without manipulation - machines that would be, in effect, self-assembling. Such an ability is crucial, as it would be unthinkably expensive to craft each one individually in a laboratory."This is incorrect. Everything in nature is both "assembled" and "self-assembling". The central dogma of biology -- DNA gets translated into messenger RNA which then gets translated into proteins requires "directed assembly", fiirst of the mRNA and subsequently the proteins. These are the primary functions of the nanoscale assemblers biologists call RNA polymerase and the ribosome. The subsequent 3-D protein folding process and the binding together of various proteins into workable machines is the "self-assembly" part. As the raw materials (energy from the sun, water from rain, and carbon dioxide from the atmosphere are essentially free), products "grown" using molecular nanotechnology need not be expensive. We will design nanoscale machines either by intentional design with atomic scale computer modeling programs or perhaps use use computer-aided genetic evolution [Bra00]. As documented by Drexler [Dre81] the cost of anything we desire to manufacture using molecular nanotechnology should fall to approximately the cost of the raw materials. That should be around $1.00/kg if we are able to avoid licensing or manufacturing fees. Having competing open source designs supported by government or foundation based development efforts may be one way to limit what private industries may charge.
"That is an important caveat, because the proponents of nanotechnology sometimes get carried away. Eric Drexler, who founded the Massachusetts Institute of Technology's Nanotechnology Study Group and whose 1986 work Engines of Creation became a cult classic, has, for instance, come up with outlandish suggestions, such as the "grey goo". This is a collection of many billions of nanomachines that are self-replicating and thus require new material to make more copies of themselves. Like so many locusts, these machines would, in theory, swallow up every object in their path so as to turn it into more nanomachines, dissolving the whole world in a sea of grey goo."Eric did not get "carried away" when he wrote about "grey goo". He was justifiably concerned over the possible accidental uses or even misuses of nanotechnology. Bringing it to the attention of people early, allowing people to think about the consequences of the technology allows the development of approaches such as the Ralph Merkle's "Broadcast Architecture" [Mer92] that separate the instructions for how to build a copy of a nanobot from the machinery that assembles nanobots or the nanobots themselves. This helps to prevent accidents. This discussion of the hazards continues today, 15 years later, leading to publications like the Foresight Institute's "Molecular Nanotechnology Guidelines". Hopefully this approach will allow us to have the benefits of nanotechnology without the worries and wasted resources that went into the development and deployment of nuclear weapons and the Cold War.
"This is the bizarre end of nanotechnology and it has attracted the scorn of fellow scientists. Stanley Williams, a fellow and director of the Hewlett-Packard Laboratories, scoffs: "I don't believe in this grey goo stuff." He derides Dr Drexler's work as insufficiently scientific."This is a classic case of people expressing 'disbelief' regarding Eric's work without expressing any credible objection. This was first dealt with over 5 years ago in the infamous commentary on nanotechnology by Gary Stix in Scientific American [Sti95]. This has been dealt with in great detail by the Foresight Institute [For96].
Dr. Williams is a physical chemist and computer scientist (bio here) and while his work in some areas such as fault-tolerant computing is excellent, when it comes to self-replicating systems or Eric's work he doesn't know what he is talking about. Anyone who has even briefly looked at Nanosystems [Dre92] knows that Drexler is extraordinarily scientific. A recent discussion by several people who know Eric and his work in detail (all of whom had IQ's over 130) concluded that Eric's IQ had to be above 160. If there was any problem with Eric's work, it is that he published Engines of Creation 6 years before Nanosystems, perhaps causing "serious" scientists lacking his insight to disbelieve his ideas. Eric's work 15 years ago was so far ahead of what everyone else was thinking or doing that people can only pronounce unsubstantiated claims that it is "insufficiently scientific". People should keep in mind that Eric spent at least 11 years (1981-1992) developing the ideas of molecular nanotechnology and most other people have only been thinking about it for a few years and therefore they haven't had time to absorb it all yet.
"Dr Williams, too, works in nanotechnology - but in a different field, trying to create computers out of the same microscopic components that would be used in the nanomachines. Most lay observers would find his work just as arcane."It is precisely because Dr. Williams is working at the "component" scale (e.g. 1-100 nanometers) that he cannot see the big picture. All currently envisioned self-replicating systems based on nanotechnology (nanobots), require billions of atoms [Fre98, Fre99, Fre01], not the dozens to thousands that Dr. Williams works with. If one wants "valid" opinions regarding the "grey goo" scenario, one should talk to microbiologists who work with self-replicating systems every day. For example, the Nobel prize winner Joshua Lederberg has written about the epidemics humans have faced in the past [Led00a] and the "World Wide Web" of microorganisms that continues to rapidly evolve [Led00b]. Given facts like the 15 people died from E. coli food poisoning in Scotland in 1985, 11 people died from Necrotizing fasciitis caused by Streptococcus infections of the so-called "flesh eating bacteria" in England in 1994 and the recent rapid spread of hoof and mouth disease, it is highly surprising how any reporter in England can doubt the reality of "goo" scenarios. The definitive paper on nanotechnology enabled goo scenarios was delivered by Robert Freitas, Jr. in 2000 [Fre00] in response to Bill Joy's "Why the Future Doesn't Need Us" [Joy00]. It documents some very scary scenarios that are possible but concludes that they can be addressed if we are prepared. Disbelief by individuals, such as Dr. Williams, is precisely what Eric was trying to combat by discussing such scenarios Engines of Creation. Individuals in our era will not be able to maintain that they were not warned about the potential dangers that could be created.
Bacteria are self-replicating systems that do contain billions of atoms. Each human being contains ~40 trillion bacteria on, or in, their body [Fre99]. Fortunately, almost all of these are "friendly" bacteria and they occupy the various ecological niches that we provide and usually manage to crowd out (or eliminate) unfriendly bacteria. We currently have the genetic blueprints for over 100 of these organisms [Bra97]. While this allows us to design new approaches to killing the unfriendly microorganisms, it also potentially provides the information that bioterrorists would need to create a "green goo" scenario. Predictions by leading bioethicists such as Arthur Kaplan who have looked at what is known as the "minimal genome" problem [Hut99, Mus96], predict that scientists will be able to construct these artificial microorganisms within the next few years. The only difference between "green goo" and "grey goo" is the fact that in the "grey" scenario, the self-replicating systems are built out of stronger materials.
If one wants to ask scientists about a feature of a technology, one should determine what the critical aspects of those technologies are and who has knowledge about them. For the "grey goo" problem it is the "self-replication" aspect of molecular nanotechnology that is the dangerous component. The people who understand self-replication, other than the previously mentioned microbiologists, are computer scientists who study either self-replication or artificial life, like Moshe Sipper who maintains a bibliography on the net about self-replicating systems [TASRP]. One should not always assume that an expert in one area, even a closely related area, may be a reliable source for comments about technologies he is unfamiliar with [Bra01].
"Making springs and belts is just the first stage - it may yet turn out not to be possible to assemble machinery out of them at a bearable cost. If so, nanomachinery may be consigned to the history books as an interesting technological diversion."Actually, the assembly of modified microorganisms is quite inexpensive compared with many other forms of scientific research. To sequence genomes such as the human genome requires the creation of millions of novel microorganisms, each of which carries a small fragment of the genome being sequenced. We are assembling modified "nanosystems" on a very large scale even today. I personally am currently circulating a business plan for a company that would design and assemble bacteria-sized "biobots" and it is clear to me, from developing strategies for how that should be accomplished, that it can be done quite inexpensively.
On the whole Ms. Harvey's article was quite well written and presented a good picture of the current "state-of-the-art" in nanotechnology research. My recommendations for improving such efforts in the future are: