• Google
    This Blog Web

October 2011

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29
30 31          

RSS Feed

Bookmark and Share

Email Feed

  • Powered by FeedBlitz

« Knowledge in the Making | Main | Virtual Nano Classrooms »

March 28, 2005


Feed You can follow this conversation by subscribing to the comment feed for this post.

Tom Craver

Sure, there should be research into the biological effects of nanoparticles - but associating that with military use just seems to me to be a subtle form of demonization: people who don't like the military will be encouraged to see nanoparticles as a more significant problem if the two are tied together.

In terms of number of people exposed, things like nano-particle sunscreen and "nanopants" are far more significant. And when micro robots become commonplace, likely far more lives will be affected by those in the hands of the police and civilians, than by military robots.

Mike Treder, CRN

Good points, Tom. This particular study was aimed at answering questions for a specific branch of government. But ACUNU certainly will be looking at issues raised by advanced nanotech in other contexts as well.

Mike Deering

Nanoscale remote sensor and surveillance systems will be among the first uses of MM.

Howard Lovy

Actually, there's surprisingly little mystery in what the Pentagon would like nanotech to do for them. This post links to one version of the defense budget. Quite a wish list. Here's another not-so-secret secret. The defense spenders have no idea how this technology will suddenly materialize for them. So, they'll pump more and more money into disparate companies and research groups who live from grant to grant, but have little incentive to provide any kind of finished product, and there's nobody in charge with a vision of what exactly these "nano-enabled" systems really mean. Meanwhile, our troops in Iraq are hunting scrap metal to protect their Humvees.

Stephen M. St. Onge

      With the concerns about nanotechnolgy that have already been expressed, this strikes me as an intelligent effort by the armed forces to get out in front of potential problems.  Now, if anyone questions military research on nanotechnology, they point to this report and say

      'Yes, we know, and we're already looking into ways to deal with those problems.  We aren't the only armed forces on earth, you know, so we may face these problems whatever we do or don't do with nanotechnology ourselves.  Thanks for your concern, and if you have any ideas on other problems that might arise, or possible solutions, let us know.'

      Slick, in other words.

[slogan promoting violence removed by moderator]


Any group affiliated with the UN has to be taken with a huge grain of salt. But when it concerns an assessment of the USA, and especially its military, how can anyone take them serious? The UN has been one of the greatest opponents of American military action.


I am most concerned about nano orgasmic pump beta2 antagonists, which have the potential to depopulate entire nations in just one generation. If the military has no answer to these agents the human population is doomed.

Mike Treder, CRN

NOTE FROM MODERATOR: Thanks to a brief pointer last night from Instapundit, our blog has been visited by thousands of new readers. We're grateful for that, and for their comments. But some, unfortunately, like Boris above, are either making light or are badly misguided.

I won't delete that one, because it seems relatively harmless, and our predisposition is for openness and free speech. But the comment from Stephen M. St. Onge contained an unacceptably offensive slogan that was removed.

We are dedicated to free expressions and new ideas here, but advocation of violence against others crosses the line and will not be tolerated.


Boris' comments are satirical.

I too am worried that when the bridges and tunnels become conscious and demand tolls, we will not longer be able to drive our cars around. It will be the end of Western Civilization.

Why did they focus on military applications? Why did they focus on health issues as the donstream effect? What about the positive effects? Seems very doom and gloom. Did they have a die-in afterwards and burn a photo of Feynman?

How about focusing on how military technology will free more people? Hello! Now thats a topic!

Eliezer Yudkowsky

What a complete farce. This is like talking about military applications of nuclear technology and forecasting that glow-in-the-dark radium flashlights might help soldiers see better at night, and one daring person on the panel suggesting that nuclear-powered submarines might be possible.

What about using molecular factories to assemble tens of thousands of nuclear weapons, perhaps overnight? What about a stealthy first strike with bots that infiltrate every human in the geographic area of the United States and send them all to sleep at a predetermined time? What about the simple threat of nanotechnological weapons being developed provoking a nuclear war among powers who foresee defeat approaching unless they strike while the sides are still even? What about the prospect of a nanotechnological defense against incoming ballistic missiles provoking a nuclear war? What does the prospect of a successful first strike that destroys a population but leaves all the cities and machinery intact, do to the logic of Mutual Assured Destruction among multiple states armed with nanotechnological weapons? What about sleeper weapons, designed with dead-man switches, to strike in the event of a first strike against the home country? What if one of those sleepers triggers accidentally, wiping out one country, and then those country's sleeper weapons go off...

Moving the battle to the molecular level *remakes* the battlefield. The molecular level is as different a military field, as tribes fighting with spears are from formations of aircraft, from attack and defense conducted in cyberspace, and from mutual assured destruction with nuclear weapons. Atoms are a different playing ground. We can't even defend against computer viruses when we have total control of the infrastructure and 'laws of virtual physics' on our home ground - how will you defend against invading bots in the messy physical world?

And you cannot seriously think that nanotechnology is going to produce any less increase in the destructive capability of war than did nuclear technology.

Respirocytes for human soldiers on a human battlefield? *Give* me a *break*!

Chris Phoenix, CRN

Eliezer, you're exactly right.

We're doing our best at CRN to get people to realize just how powerful general-purpose molecular manufacturing will be, and how far-reaching its implications will be. It's a slow process. I've spent the past two years building a technical foundation so that we could start to talk about implications.


Mike Treder, CRN

The fact that there was even a mention of "large quantities of smart weapons" in the report is because we pushed so hard for it. There were a lot of other things we tried to get included, but it turned out this wasn't the time or place.


The problem with nano-weapon applications is that they are inherently offensive. There are few measured defensive responses available against many offensive attacks. You have to trust the goodwill of the 1st entity to develop MNT; that they will use their tech. edge benevolently to depopulate the world of other MNT programmes in progress. Otherwise an arms-race is inevitable. I think decentralized human systems that spread faster than the "nodes of destruction" spread is the only long-term solution that will allow us to harness the nano-benefits. It's why I'm so high on space-elevator technology; earth must be made redundant ASAP. I think it might be a little too late for integrated international nano-efforts...

Chris Phoenix, CRN

I don't think a space elevator will be built before MM arrives. On the other hand, MM will make it pretty easy to get to space with or without an elevator, if you have even a few months to develop aerospace hardware in "startup company mode." On the third hand, once MM arrives, no one can count on having a few months to play.


Tom Craver

Unless the first developer has several years lead (and they don't dare assume they do), shutting down all other MNT projects isn't going to be practical. You'd need to develop systems to reliably locate them all, sabotage them all, and sabotage key weapons (nukes at least) that might be used in retaliation if your attack is traced back to you. This isn't likely to be done in less time than the Apollo project - years.

You'd probably be better off just developing a nano-weapon-of-mass-destruction, giving a small demonstration, and politely telling the world how you're going to use MNT to make the world a wonderful place - so long as no one attacks you and thereby forces you to destroy the world - an extension of the MAD doctrine.


The problem is that MAD doesn't necessarily apply in a nano-war. It might be winnable. There are ways of selective targeting your opponents and their infrastructure while leaving your own assets intact. The physical location of your doomsday weapon could be traced and nuked, or otherwise be rendered impotent. I think with supersonically delivered super-MIRVed small nano-drills and airborne spy devices, you could infiltrate most structures in a matter of weeks or months at the most. Remember, the possessor of the 1st nano-factory will soon have more industrial capacity than the rest of civilization combined. Eliezer might be right above, that a premptive strike by a side losing the Nanhattan Project race is likely. Our current geo-political power structures are based on obsolete paradigms of scarcity and force.

Brett Bellmore

"Remember, the possessor of the 1st nano-factory will soon have more industrial capacity than the rest of civilization combined"

Rapid expansion of nanotech manufacturing capacity to that level would likely be damned conspicuous, since one thing we're fairly certain of is that early nanotech is going to be very energy intensive per unit mass. Unless you were very good with covert weapons, there's that embarassing window between the time you're detectable, and when you're able to out-fight eveyone else.

If it were me, I'd probably covertly build until I could launch the smallest possible nanofactory seed to the far side of the moon; It gets just as much sunlight as the near side, and who's watching it? Then you could grow your industrial base unwitnessed.

Secondly, where are you going to get the designs for all the weapons? And who's going to control them for you?

I can't see a few months' lead translating into an unbeatable advantage, unless you've got good AI, both smart, AND loyal to you.


You're right Brett, about detection of "garage MM" efforts and their paucity of product designs. I was assuming (perhaps incorrectly) that the 1st assembler would be the product of a national Nanhattan-Project, with the associated factory facilities and product design personnel. If power requirements exhaust available generating capacity, I'd guess among the 1st nano-products would be highly efficient power sources. I don't see why a massive hangar or warehouse couldn't utilize dedicated generator capacity for a government effort.
AI couldn't be made loyal in any sense of the word. If it was conscious via some sort of neural substrate, it would experience a lifetime of subjective memories in a mere moment of ours. It would be starved for sensory stimulus; be the most autistic entity ever in existence. One growing AI would soon attain more "value", than would exist in the rest of the biosphere, by whatever quality of life parameters we measure. Its own self-interest would be altruism.

Brian Wang

I agree with Brett that for a first developer of nanotech to gain pre-eminent power (who is not the US military) would have to develop enough of an nano-advantage to overtake the conventional and unconventional capability of the USA and any other force. The US would be able to extend any lead with the comfort of having conventional and unconventional superiority. I also think it would be difficult for a small group to design the systems and products to maintain nanotech advantage and military dominance without strong AI.

Brett Bellmore

Cdnprodigy, I'm not sure your conclusions follow.

First, on power: Generators (Based on known technology) require some sort of fuel, or at least access to some energy store in the enviroment. And unless unreasonably efficient, also require a heat sink. You can't just stuff a warehouse full of generating capacity, you'd need an oil or gas pipeline going into it, a good sized water line, steam coming out... Major generating capacity is not inconspicuous.

Estimates are that initial nanofactory designs, without major energy recovery, are going to use MASSIVE amounts of power per unit mass of created product. Much more than any conventional manufacturing process for bulk products.

(In fact, I think it's inadequately appreciated how energy demands would slow the rollout of nanotechnology, even if it were handed to us on a silver platter; Gotta keep the conventional economy running at least a while while you build all that stuff.)

This is going to be difficult to hide, AND a strain on a nation's energy resources. Which in the meantime still have to be devoted to everything else they were running before. So, sure, you can have generating capacity grow exponentially as part of the system, maybe fuel cells built into the factories, but by the time you're putting out large quantities of product, you'll be straining your conventional economy, AND the world's intelligence services will know you're doing SOMETHING that's using huge amounts of power.

Like maybe isotope separation for nuclear weapons. So they WILL be rather insistant on finding out where the power is going.

Any effort to parlay a nanotechnology breakthrough into military superiority would thus have to concentrate on weapons with an extremely high performance per unit mass, and I'm not just talking about conventional weapons redesigned around better materials. The emphisis would have to be on things like survailance technology, remotely actuated viral dispensers that hide in a person's body, things that inconspicuously position themselves, and do their damage with information, or with micrograms of payload. The sort of weapon where a hundred pounds, properly targeted, wins a war before it starts.

So we're talking major design efforts. Either AI or a LOT of high powered designers. And, frankly, the disappearance of that many engineers would probably be as conspicuous in it's way as a big powerplant.

AND a lot of people controlling the delivery of these products. Monitoring the info coming from the survailance equipment.

So, I don't think the nanotech breakthrough really translates into sudden military superiority, without AI, because you'll get *noticed* before you're ready to strike, otherwise.

Which brings us to AI.

Sensory deprivation would appear to be a significant problem for AIs based on human neural structures, which *expect* constant sensory input. But that need could be satisfied with some kind of virtual reality, I expect. Providing your AI engineers with CGI cubicals to work in...

Extremely long runtimes are going to be problematic for any AI based on humanity; We get *bored*, after all, our interests change. You might, instead, use your resources to run multiple copies of the AI, instead, so that the total runtime on any given copy wasn't excessively long.

For non-human AI, all bets are off. I'm not sure such would actually suffer from sensory deprivation, though it wouldn't suprise me if early versions tended to get lost in odd intellectual obsessions unless constantly disciplined by interaction with outside enviroment. But I doubt we really know enough about the subject to make reasonable extrapolations.

As I've said before, my ideal model for AI is *Amplified*, not Artificial, intelligence: Computational extensions to the human brain, with all the critical motivation remaining in the human part of the system. We've GOT intelligence already, it's just too limited. Improve our own, and we'll at least stay on top.

Chris Phoenix, CRN

Brett: "one thing we're fairly certain of is that early nanotech is going to be very energy intensive per unit mass."

I agree, BUT we don't know how long the time delay will be between early nanotech and late nanotech. We can't even guess within two orders of magnitude, because it depends on choices made by the developers and on design technologies that can't be fully tested until after early nanotech is achieved.

If I were planning to develop advanced nanofactories ASAP, I would develop genetic algorithm design programs that used petaflop computers to do molecular machine design. I'd also develop competence in automated build-and-test implemented via early-nanotech construction. The first thing I'd do with my early nanotech is to build those computers; the second thing is to start refining my models and testing the designs.

An efficient machine-phase mechanosynthesis process might not take very long to develop. I could imagine it being done in a matter of weeks or even days, if the process were mainly automated with high-level human guidance. Of course this would require pre-design of the machinery at a fairly low level, with modular design and multiple options for each component in case an approach didn't work.

Suppose my goal is the atom-level specs for an efficient nanofactory capable of building duplicate nanofactories.
Here is what I could pre-design, and build with early nanotech:
1) The mechanical blueprints for the nanofactory, broken down into simple components (bearings, shafts).
2) A CAD program that could swap components in and out, propagating mechanical constraints/specs. (This may be more advanced than today's CAD, but I don't think by much.)
3) Multiple petaflops of simulators.
4) The ability to build and test a billion variants of components per second.

With tools like these, how long would it take to create efficient components and an efficient nanofactory? I suspect that software engineers are likely to give much shorter estimates than hardware engineers. And

Once you have a nanofactory that's efficient enough to generate electrical power while it operates (from the excess hydrogen in the feedstock), it's not hard to hide the nanofactory's waste heat or usage of hydrocarbons. Just install the nanofactories inside existing hydrocarbon-fired electrical generation plants. Hydrocarbons go in, electricity and heat come out just as they always have... and who's to notice that not all the carbon goes up the smokestack?

As to weapons and nano-first-strike: I think we would need some serious wargaming to figure out how much damage could be done by, for example, a million kilogram-scale remote/AI piloted aircraft. Remember that a kilogram of diamondoid is the equivalent of hundreds of kilograms of today's airframe, computers, powerplants, and sensors.

I'd think that a thousand tons of diamondoid, made into ten million loosely-pre-targeted mostly-automated aircraft, loaded with one hundred thousand tons (~1 large tanker) of semi-processed hydrocarbons for fuel and incendiary munitions, might be able to destroy overnight a large nation that wasn't hardened against such attack. "Mr. President, you have probably been informed of certain radar events earlier today, and you may have noticed the firestorm in Chicago. That was an hour's work by ten thousand of our twenty-pound UAV's. We have ten million UAV's distributed over every city and major town in your nation. Surrender by midnight and blow up all your silo'd missiles by 3AM."



Garage MM efforts wouldn't enslave the world or maybe even destroy it (a standard bio-lab would be a more likely avenue for the latter). As for the power requirements, would it be possible to get a very rough approximation of what those would be per product mass? The waste heat problem is easy: nano heat pipes, nano belt, drop, and sheet radiators based on current designs. A military base on a geologically active locale might avoid heat detection. Could probably harness a few thousand MW's of power from conventional PEM fuel cell generators in the next decade if enough hydrogen could be accumulated. Nano flywheels would be an easily engineerable battery.
A convential nuclear arsenal need not be rendered obsolete early on, by a nano-adversary. 100MT of nukes ground burst (on a nation's own soil to avoid nano-aerospace missile defense) will cause a nuke winter. One submarine could trigger this. Indeed, there are rumors Russia has an H-bomb in the 1000's of MTs. Hmmm, would nano-drills (size of a coffee thermos?) boring through the lithosphere, be detectable with seismic sensors? If they are, an underground airtight shelter with enough nukes to trigger a nuke-winter might be a preliminary nano-deterrant... I just said in my previous post MAD might not apply. This is giving me a headache.


Ps. I'd be very wary of AI that could not potentially experience sensory deprivation. Without stimulus analogous to what we experience, it would have no framework with which to value human life. Augmenting human minds to teach it while limiting its initial growth rates to those of the cyborg-teacher seems a good idea, but AI modeled from human minds would have to precede other forms.

Brian Wang


For the millions of UAV scenario. The design hurdle for the UAVs is that they would have to be able to overwhelm conventional air defences (fighters and conventional UAVs), anti-aircraft, anti-mortar defences. and the follow on possibly near-nano capability of the United States at that time. Any laser, anti-missle defences that might be deployed.

If the creators of the UAVs were another nation. The USA would still have its nuclear arsenal. Any widescale threat or destruction by nano-weapons would be replied with a nuclear strike.

ie. blow up my cities with nano-UAVs. I blow up your cities with my nukes.

You would have to use nano to develop space travel to lift what you want out of reach. That much activity and development toward that would clearly be detectable. That would also be a possible better strategy. Develop nano-lead and leave (without trying to fight here). Staying earth bound leaves you vulnerable.

Chris Phoenix, CRN

Given that these things will be small, fast, containing no metal, and giving off very little heat (and might even be largely transparent), I think they'd be pretty hard to hit with any weapon that wasn't designed for them.

Against a country that had nanotech, big clunky semi-dumb UAV's wouldn't be much of a threat. Just build smaller faster UAV's "upon their backs to bite 'em." But I really doubt we could do much about a UAV attack with our current weapons.

Retaliation might not be practical. You have to know who's attacking, and you have to find out before the war ends "an hour and a half from now." And even if you know who--that doesn't save your own country, it just lets you destroy theirs--and certainly wouldn't be able to stop the attack.

And that's assuming the UAV's can't destroy missiles. What would it take to destroy an ICBM? They can't be armored... If you catch them on takeoff, when they're slow, I'd think that a .45 shell into a fuel tank/grain would be enough. You might not have enough UAV's to catch sub-launched missiles on launch (unless you knew where the subs were), so you'd need some way to stop faster/higher missiles--but not too many of them. Or, just bluff.

By the way, I've thought of a second way to conceal massive manufacturing operations. Oil tankers take weeks to cross oceans. You have all the hydrocarbons you need, for fuel and feedstock, and an unlimited supply of cooling water. And they're metal, so I doubt you could see inside them with radar. You'd have to physically snoop under the tanker to find the cooling water pipes. H2O exhaust would be undetectable. CO2 exhaust, if there was any (meaning, if your process was inefficient enough to need to use carbon as well as hydrogen for energy), could be dissolved in deep water.


The comments to this entry are closed.