While it is flattering to receive favorable reviews, it is often more instructive to consider critiques, and the nature of a blog is to stimulate honest dialogue which encompasses both. In that spirit, I would like to focus on a point raised by Mike Treder, co-founder and Executive Director of the Center for Responsible Nanotechnology.

On his blog (http://crnano.typepad.com/crnblog/science_technology/) he states:

Alex Fuss, lead researcher for CSC’s Digital Disruptions report, offers some interesting predictions about the future in this interview published by Financial Times. [Hat tip to Nanotechnology Now]   

Among other things, he says:

We will be able to print out toys, parts, furniture, designs and more from the net using three-dimensional printers. Today 3D printers cost $20,000 and can only print prototypes. But maybe beyond 2013 you’ll actually be able to print a pair of sneakers in your size. . .

There will be no more lies. You’ll still be able to have secrets but only if you can keep them off the net. Privacy will be available but only to those who can afford to pay for it. For most people, privacy will end in 2013, or a little beyond that. . .

Quantum computing will shatter current encryption techniques, jeopardising anything that relies on encryption, such as credit card transactions, and requiring new approaches to information security. Using an electron spin as opposed to an electron charge in quantum computing would mean all the cryptography we have today would have to be rethought. . .

Unfortunately, he gets one thing very wrong:

From 2025 to 2050 and beyond, nanotechnology will give us the capability to create anything, molecule by molecule, atom by atom. The technology is at very early stages but it is definitely going to happen. For example, it will be possible to create a piece of wood. But self-replication is crucial, otherwise the technology won’t scale up. What’s needed is nano parts that self-assemble—that way you can mass-produce anything. [emphasis added]

This is a common misconception that still persists although it has been obsolete for a decade and half. Obviously, Alex Fuss has not read Nanosystems (published in 1992), nor any of the hundreds of pages on our website and others that expound on those ideas.

But all he really needs to know is contained in a press release issued by CRN a few years ago, titled “Leading nanotech experts put ‘grey goo’ in perspective.“ Here is the key paragraph:

Contrary to previous understanding, self-replication is unnecessary for building an efficient and effective molecular manufacturing system. Instead of building lots of tiny, complex, free-floating robots to manufacture products, it will be more practical to use simple robot arms inside desktop-size factories. A robot arm removed from such a factory would be as inert as a light bulb pulled from its socket. The factory as a whole would be no more mobile than a desktop printer and would require a supply of purified raw materials to build anything.

Self-replication is not crucial for the technology to scale up. Advanced nanotechnology—exponential general-purpose molecular manufacturing—will be achievable without it. And perhaps even sooner than Fuss expects.

I was given the benefit of the doubt and my position somewhat defended in a comment by Brian Wang posted in response to Treder’s blog entry: 

It might be possible that he was speaking about self-replication in simpler terms. For instance when the first nanofactory is built we will have to start somewhere. We will not build each nanofactory arm by human guidance because it would cost too much time and money. So maybe he just meant that nano systems would construct other nanosystems in a factory setting with no danger of run away replication.

For the record, Treder is correct in saying that I did not read K. E. Drexler’s Nanosystems (Wiley-Interscience, 1992), but I did read Nanofuture: What’s next for Nanotechnology (Prometheus Books, 2005) by J. Storrs Hall, which includes a foreword by K. E. Drexler.

In that book, Hall, following Moshe Sipper, distinguishes between reproduction, self-replication, and autogenous replication.  “[R]eproduction involves making, on purpose, imperfect copies for the purposes of evolution, whereas replication is the similar mechanical process of making an exact copy.” In a self reproducing cell, the parts that the cell factory produces, protein molecules, “are not picked off the production machine and carefully conveyed to meet the next part in the process. Instead they float loose to bump randomly into other parts until they meet one they click together with, in a process called self-assembly. In a nanomachine, as in a human sized ‘factory,‘ parts are never loose to drift around. They are moved by robotic arms, shuttles, and conveyor belts from spot to predetermined spot and are assembled by mechanical force….The nanomachine is a lot faster and simpler in some ways than the machinery of the cell. On the other hand, [the nanomachine] has a big disadvantage: it can’t evolve.”

Once we build the first nanofactory that can build all the parts to produce a second nanofactory, and can assemble that second factory and transmit to the second factory the instructions for doing same, we have an autogenous system where “[t]he system as a whole can extend or repair itself,” even though “it does not contain a single machine that can replicate on an individual basis.”  It is to thit capability that I was referring in the Digital Disruptions report upon which the UK Financial Times article was loosely based. In the report we wrote about “atomically-precise manufacturing and self-assembly”.  In my interview with the UK Financial Times, that somehow became “self-replication,” probably through a slip of my tongue. As Hall points out, something that makes copies of itself is very different than something that makes copies by itself. Thank you, Mike, for pointing that out and noting that exponential general-purpose molecular manufacturing will be achievable perhaps even sooner than I project.

Incidentally, subsequent to the completion of Digital Disruptions, I came across Adrian Bowyer’s Replicating Rapid prototype (www.reprap.org), a 3-D printer that “prints” all the parts needed to build a copy of itself except the screws needed to hold it together. Currently, a human with a screw driver is required to assemble the parts to produce the “son-of-printer” printer, but one can imagine it won’t be long, relatively speaking, before we see truly autogenous factories on both the macro and nanoscale.