Feb 18, 2007

The Bulletin of Atomic Scientists Displays Gaps in Nanotechnology Understanding

Posted by in category: nanotechnology

Like the Lifeboat Foundation, The Bulletin of Atomic Scientists is an organization formed to address catastrophic technological risks. In catastrophic risk management, vision and foresight are essential. You take at technological, social, and political trends which are happening today — for example, steps towards mechanical chemistry, increasing transparency, or civil atomic programs — and brainstorm with as many experts as possible about what these trends indicate about what is coming 5, 10, or 20 years down the road. Because catastrophic risk management is a long-term enterprise, one where countermeasures are ideally deployed before a threat has even materialized, the further and more clearly you try to see into the future, the better.

Traditionally, The Bulletin has focused on the risk from nuclear warfare. Lately, they have expanded their attention to all large-scale technological risks, including global warming and future risks from emerging technologies. However, the language and claims used on their website show that the organization’s members are only just beginning to get informed about the emerging technologies, and the core of their awareness still lies with the nuclear issue.

From The Bulletin’s statement regarding their decision to move the clock 5 minutes to midnight, from the “emerging technologies” section specifically:

The emergence of nanotechnology — manufacturing at the molecular or atomic level — presents similar concerns, especially if coupled with chemical and biological weapons, explosives, or missiles. Such combinations could result in highly destructive missiles the size of an insect and microscopic delivery systems for dangerous pathogens.

“Highly destructive missiles the size of an insect”? Depressingly, statements like this are a red flag that the authors and fact-checkers at The Bulletin are poorly informed about nanotechnology and molecular manufacturing. To my knowledge, no one in the entire defense research industry has ever proposed creating highly destructive missiles the size of an insect. Highly destructive missiles the size of an insect are impossible for the same reason that meals in a pill are impossible — chemical bonds only let you pack so much energy into a given space. We cannot improve the energy density of explosives like we can improve the speed of computers or the resolution of satellite imagery. There can be incremental improvements, yes, but suggesting that nanotechnology has something to do with highly destructive missiles the size of insects is not just dubious from the point of view of physics, but particularly embarassing because it seems to have been made up from scratch, and was missed by everyone in the organization that reviewed the statement.

The general phrasing of the statement makes it seem like the scientists that wrote it are still stuck in the way of thinking that says “molecular manufacturing has to do with molecules, and molecules are small, so the products of molecular manufacturing will be small”. This is also the bias frequently seen displayed by the general media, although early products based on nanotechnology (not molecular manufacturing), including stainless pants and sunscreen, also subtly direct the popular perception of nanotech. It’s natural to think that nanotechnology, and therefore, molecular manufacturing, means small. However, this natural tendency is flawed. We should recall that the world’s largest organisms, up to 6,600 tons in weight, were manufactured by the molecular machines called ribosomes.

Molecular manufacturing (MM) would greatly boost manufacturing throughput and lower the cost of large products. While some associate MM with smallness, it is better thought of in connection with size and grandeur. Although microscopic killing machines built by MM will definitely become a risk by 2015–2020, the greatest risk will come from the size, performance, and sheer quantity of products. Because a nanofactory would need to be able to output its own weight in product in less than a 12 or so hours or it wouldn’t have been developed in the first place (scaling up from a single molecular manipulator to many trillions requires 33 or so doublings — which could take a long time if the product cycle is not measured in hours), these factories, given raw materials and energy, could produce new factories at an exponential rate. Assuming a doubling time of 12 hours, a 100 kg-size tabletop nanofactory could be used to produce 819,200 kg worth of nanofactory in only a week. As long as the nanofactories can support their own weight and be supplied with adequate matter and energy, they can be made almost arbitrarily large. Minimal labor would be necessary because the manufacturing components are so small, they must be automated to work at all. Regulations and structural challenges from excess height can be circumvented by fabricating nanofactories that are long and wide rather than tall and fragile. Once created, these factories could be programmed to produce whatever products are technologically possible with the tools at hand — at the very least, products at least as sophisticated as the nanofactories themselves. Unscrupulous governments could use the technology to mass produce missiles, helicopters, tanks, and entirely new weapons, as long as their engineers are capable of designing diamondoid versions of these products. Their rate of production, and quality of hardware, would outclass that of non-nano-equipped nations by many orders of magnitude.

Because unregulated, exponentially replicating molecular manufacturing units would create a severe threat to global security, it seems prudent to regulate them with care. Restrictions should be placed on what products can be manufactured and in what quantity and quality. Just as permits and inspections are required to operate industrial machinery, restrictions should be placed on industrial-scale molecular manufacturing. In some cases, preexisting regulatory infrastructure will be sufficient. In others, we’ll need to augment or expand the purview of historical regulations and customize them to address the specific challenges that MM represents.

Further Reading:

30 Essential Nanotechnology Studies
Lifeboat Foundation NanoShield
Nanotechnology Category on Accelerating Future


Comments — comments are now closed.

  1. Phillip Huggan says:

    MNT in 7 years?! This is where I mark my break from this blog. Your delusions are very marginally impacting real surface science research. *Real* surface scientists are laughing at you but there may be impressionable funding sources that will be burned by this estimate; won’t front capital when this technology really does become mature. Stop brainwashing people (I wish you all the best and will “affiliate” with you again when you learn basic solid-state physics). Political differences are fine but this is inexcusible.

  2. “Although microscopic killing machines built by MM will definitely become a risk by 2015–2020, …”

    I agree with Huggan on this one, I read Kurzweil’s “Singularity is Near” and Richard Jones’ “Soft Machines” at the same time and was amazed by the distance in “optimism” on molecular manufacturing. MM (in it’s true sense) by 2015 sounds highly improbable. I also believe it to be a very positive step of The Bulletin to include emerging technologies as a factor in the Doomsday Clock (even if they envision nanoscopic artillery)

    Drexlerians and Kurzweilians believe MM to be imminent but still there’s very little actual research into this (expect maybe Freitas and the US military) while the conservative British are running full speed ahead with Ideas Factory Software control of matter.

    See a post on the Soft Machines blog on this, Keeping on keeping.

    I hope the members and ideas of the Lifeboat Foundation can bridge this gap in understanding.

  3. The highly academic blog Overcoming Bias from the Future of Humanity Institute at Oxford is interesting in this context.

  4. Like the folks at the Center for Responsible Nanotechnology and Eric Drexler himself, I think that MNT between 2015 and 2025 is feasible if not quite likely.

    Not a Drexlerian or a Kurzweilian myself per se, but after discussions and reading the work of J. Storrs Hall, Chris Phoenix, et al, I think MNT is easier than it seems to most people, although certainly not easy, but the incentives to develop it are great, and billions are already going into nanotechnology in general, which provides the prerequisites for MNT.

    Ole, you say that no one is working on MNT except for Freitas, then contradict yourself by bringing up the Idea Factory… dozens, if not hundreds of others are working on MNT as well, at corporate labs as well as in academia.

    Phillip, sorry you find the timeframe inexcusable. I recommend “Design of a Primitive Nanofactory” by Chris Phoenix:

  5. This ends up with the definition of what MNT really is. General research into nanotech (sensors, electronics etc.) will incrementally lead to MNT as you say, but to say that “hundreds” are working on MNT now would be an overstatement. There’s a big difference between writing reports like “Design of a Primitive Nanofactory” or designing computer models of molecular machines and doing the actual hands-on laboratory research. It may be that I’m not aware of all the research going on out there.

    What I meant when you say I contradict myself is that the US-nanotech crowd seems to be more concerned with thinking up really advanced uses of nanotech, while the british are doing the actual research. I know this is a bit unfair towards US researchers, but the Ideas Factory seems to be more coordinated than most MNT-research in the US.

    Let’s hope we’ll see a nanofactoy by 2015, be it in the US or in the UK.

  6. Nanorex was recently on the front page of Machine Design magazine… Zyvex has been around for a while and is doing research on MNT simultaneously with less advanced nanotech stuff like bulk carbon nanotubes. IBM is working on mechanosynthesis and molecular self-assembly in an effort to keep Moore’s law going past 2015 or so.

    I know what MNT is (diamondoid mechanosynthesis), and I know that dozens of labs are working on it, albeit the “Drexlerian vision” is certainly stronger at, say, Nanorex, than it is at IBM. The disparity between the US and the UK you’re pointing to is non-existent — the US has been a leader both in designing advanced nanomachines and doing the chemical simulations and CAD work necessary to lay the foundation for further progress. The Ideas Factory thing is incredibly recent and new.

    There is a big difference between reports like “Design of a Primitive Nanofactory” and computer models. The difference is that without the report, you have little idea what you’re going for, and any models you build may have little or no bearing on an assembler design that is scalable to a human-scale nanofactory. Computational chemists run millions of molecular simulations per day, but do they necessarily contribute to MNT? Only if they have MNT in mind (from reading reports like Chris’). Both simulating and report-writing are important at this point, but there is so little “vision” in the field of MNT, aside from Nanosystems, that we need reports more than we need what one would call “actual” hands-on laboratory research. Today, the vision is more “actual” in terms of progress gained than the lab work.