Give Me That Diamondoid Tower of Babel Saturday, May 30 2009
nanotechnology 3:02 am
For some interesting analysis of MNT and the Singularity as it pertains to Biblical prophesy, check out Britt Gillette’s site.
Reading a few of the articles, I find them more MNT-aware than many transhumanists. Most transhumanists like to focus on cyborgs, some Star Trek crap, and whatever else is the flavor of the week on WIRED, io9, Digg, and Engadget. In contrast, Mr. Gillette is actually making some sort of effort to figure out who might develop MNT first. He correctly notes that Prof. Kaku’s appearance on Fox News last month was one of the first straight-faced mentions of MNT in the mainstream media since Bill Joy’s scary WIRED article.
His summary of MNT is not terrible, and probably contains info that at least half of the readers of this blog never thought of. (Except for the part where he seems to take grey goo seriously and goes blank on specific nanoweapons.) Mr. Gillette apparently thinks that September 11th will push the USA to develop MNT, which is wrong. Essentially no one is developing MNT. MNT is basically a non-issue with regards to current research. Our capacities for molecular manipulation are obviously increasing, but no one outside of the Nanofactory Collaboration is even trying to build nanomachines that can build other nanomachines. (Zyvex doesn’t really count in my opinion, sorry.) Gillette points out the following:
As Eric Drexler states in Chapter 11 (”Engines of Destruction”) of Engines of Creation, “a state that makes the assembler breakthrough could rapidly create a decisive military force – if not literally overnight, then at least with unprecedented speed.” Able to replicate swiftly, assemblers can become abundant in a very short period (if the self-replication period for an assembler is 15 minutes, then a single assembler can replicate into two to the ninety-fifth power assemblers in the first 24 hour period). Those assemblers can then be used to create weapons pre-designed in anticipation of the future development of molecular manufacturing, weapons capable of enormous destructive power – weapons that most people would find difficult to imagine. The leading force will hold unprecedented power over the nations of the earth. How it chooses to exercise that power should be humanity’s greatest concern.
This might be a radically dangerous statement, but I would assert that those seeking to minimize the risk of MNT or nanowar might be well-advised to consider sketching out designs for nanoweapons in advance. Why? Because someone will design them anyway, and if fundamentally benevolent and considerate people have high positions in MNT development activities in leading nations developing the technology, they will be able to use their influence to minimize the risk. The perfect is the enemy of the good, and stubbornly refusing to believe that nanoweapons will be developed is the enemy of minimizing the risk of nanoweapons when they are inevitably developed. (If we don’t build Friendly AI first, which I hope we do.)
In one of his posts, Gillette says that the Bible implies that the development of MNT could be followed by the construction of a very large tower, mirroring the Tower of Babel. I don’t consider this implausible at all. With MNT, a large diamond tower could be constructed. This would be a very visible application of the technology. The diamond nanoblocks used to construct the tower could even be made very light and filled with vacuum or hydrogen to provide buoyancy. I haven’t bothered running the numbers to see a plausible height, but see no reason why at least 5-10 miles of height would not be physically possible.
Britt has a radio interview up, which I haven’t listened to yet, but is probably pretty entertaining.
On a similar note, I think we should be just as worried about someone developing nanodisassemblers, nanos that disassemble materials into their component materials or elements. These could be justified for development under the idea of environmentally safe garbage removal and resource recovery, but in the wrong hands with the wrong programming could cause unprecedented devastation. Conventional bombs, crop dusting equipment or even the drop-off fuel tanks of fighter jets could be retrofitted to contain payloads of such disassemblers and dropped over any target. They could be selectively programmed to disassemble only for certain targeted materials such as certain alloys, certain kinds of plastic, plant life, or human tissue. They could be seeded into clouds and carried on the wind. They could be carried by rainwater. They might be detectable but there would be no adequate defense. If programmed to go dormant and activate when they come in contact with organic matter, they could probably dust a few dozen square miles and create an uncrossable no-man’s-land. Anything organic that tried to cross or establish itself would be disassembled.
Even without a “gray goo” replicator scenario it’s a frightening prospect.
All these wonderful apocalyptic scenarios depend on MNT working flawlessly 100% of the time, forever – one failed replication and you have an evolutionary system. Rogue goo doing ungooey things, disguising itself as normal goo to avoid the normal goo (or even the goo killing goo that would evolve in that environment). Then there’d be multicellular goo, and cordate goo, and goo that would lay goo eggs, etc …
I quite like the idea that our entire ecosystem could be the result of gray goo that accidentally created instead of destroyed. Hilarious.
People like to think of gray goo as an easy mistake – but the truth is that the level of perfection required in both the construction and programming of the devices can never exist.
We see plagues and infestations frequently in both natural and synthetic systems – but catastrophic unstoppable replication is rare (I would go so far as to say cannot truly exist). Resources run out, physics get in the way, other organisms, processes, etc. act as checks. This scenario is not as easy or as simple as it first appears.
@Carol:
Yes, I believe this scenario is more likely. It protects the attackers from their own weapons — sort of a selective napalm. Only an apocalyptic cult (ahem) would want to release a weapon that might erase all biological life.
I think it is a huge leap to say that someone who develops MNT will suddenly be able to magically develop anything or create a super army within days/weeks/months.
For instance a robotic car factory can easily make different types of cars, but that doesn’t mean that they necessarily know how to make other types of products like tanks or jets. Being able to make products requires knowledge of how molecules are arranged, which will take time to learn. The same rules apply to MNT as they do to any other factory. Imagine people arguing before the first robotic factory (RF) created that the new RF would allow a nation to make everything/anything and give that nation a decisive advantage. This is clearly not the case, and will not be the case for MNT factory either. Just having the capacity to arrange molecules precisely is not a guarantee that you can make whatever you want.
There is way to much unrealistic thinking as to what MNT will do. “Particles” deterministically move according to the quantum universal wave function. So any future arrangement of particles will be due to the progression of this universal wave function. This molecular movement does not entail magically making any product in seconds.
I also think that chemical type reactions are rather slow. Will nanotechnology factories speed up the building of macroscopic objects like a car? I somewhat doubt it. You have to look into the logistics of building things and also the fact that there are diminishing returns to using nanotechnology for a product too.
Stuart, computers work “flawlessly” for years at the circuit level because there are redundant parts. For the same reason, nanofactories could work flawlessly due to redundancy. You idea might sound cool, but it’s totally implausible — the vast majority of all variants on a nanoproduct design would just cause it to break. Even if mistakes did occur, they wouldn’t be passed on, because most nanoproducts will not self-replicate, grey goo would be extremely difficult to build, and I doubt it would ever be very effective. (Too much thermal signature, and it would be slow.) Natural evolution operates due to many different factors, all of which have to be in place for adaptations to be passed on. Study population genetics.
Mike, clearly someone would need a precise design to manufacture something on a nanofactory, but with that design, what’s to stop them? It would not be a huge effort to make designs based on simple, well understood parts such as wheels, pulleys, levers, rods, springs, etc., leading to a wide range of possible products.
I am very confused by what you mean about the quantum universal wave function. No one ever claimed that MNT could make any product in seconds. Extrusion speeds may be limited to a foot an hour or less. Still, specialized nanofactories may be able to achieve higher extrusion speeds for narrow categories of parts. The Drexler replicator speed numbers are probably exaggerated by at least a factor of 4, maybe more. The main point is still exponential replication.
Manufacturing speed will, of course, be based on reaction speed and throughput. Many “chemical type reactions” are actually quite fast, and mechanosynthesis would certainly be faster than chemosynthesis. You seem to be operating on a lot of intuition here, where when we want a real answer, we have to run the numbers. For one place where that has been done, see Chris Phoenix’s “Design of a Primitive Nanofactory”.
I wouldn’t say I’m operating on that much intuition. I do have a degree in chemical engineering and have taken college level nanotechnology courses. So I’m aware of certainly reaction kinetics and how things operate at the atomic level. I don’t have as much experience in nanotechnology as some people do, but it is probably way above the average person.
I’m not saying you couldn’t in theory lead to more/better products, but I don’t think this will happen automatically and quickly after a nano factory is created (if it is made in they way many futurist claim, which I have my doubts about). I like Brian Wang’s blog, and usually agree with everything he says. However I found myself somewhat disbelieving this article he wrote about nanofactories below. Especially when he makes so many grandiose claims like 24% gdp growth.
http://www.kurzweilai.net/meme/frame.html?main=/articles/art0672.html
I think he is a little unrealistic about what will actually be possible. Like now they can create carbon nanotube sheets. However, that doesn’t automatically mean we now have cars made of carbon nanotubes. There are logistics involved and practicality issues along with expenses. Think how long it has taken since the discovery of carbon nanotubes to create certain products composed of those nanotubes. Scientists are still discovering new properties of them and how they could be used.
Ok, sorry about the “taking seconds” part, I just figuratively meant that futurist claim things can be made extremely fast. I tend to think it won’t be as quick as some futurist claim. The quantum thing is that all particles follow deterministic physics laws (quantum multiverse many worlds theory). If you don’t know much about that aspect of quantum physics then just forget I mentioned. Its probably not that relevant.
You say that it will be easy to create well understood items like wheels, pulleys and levers. However I’m sure you know that a lot of things at the atomic level do not act anything like they do at a macroscopic level. Like this paper (below) on an autonomous artificial nanomotor is fairly complicated.
http://www.pnas.org/content/103/5/1178.full
Figuring out these things may take a considerable amount of time. Lets assume they can make these sorts of items with a nanofactory. So what? Just because you are able to make an atomic pulley doesn’t mean you can automatically make a super powered macroscopic object either using those atomic things. I think you may be underestimating the difficulty of creating products. We can already create things like atomic wheels and pulleys/levers. Even if you are able to make those things, it doesn’t mean you can construct a larger item made of those constituent parts. The nanofactory is probably going to have limitations too. Nanotechnology has already created tons of microscopic devices. Some of these things do eventually get integrated into certain products, however this doesn’t happen as fast as certain people would like. In a free market this often happens according to practicality or necessity and not what could theoretically be possible. This is something that Kurzweil has more difficulty understanding and why his futurology tends to be unrealistic at times (practicality vs. technologically “possible”).
I think the future of nanotechnology factories will be analogous to how robotic factories currently operate. Different companies will specialize in creating different microscopic products and there won’t be this one nanofactory that will be able to assemble anything imaginable. Just like now there is a lot of equipment that a university/company needs to own to be able to reliably do nanotechnology research. Car companies may not make some of their own microchips. They buy these those things from companies that specialize in them and integrate them into their product. There will probably be no nanofactory device that will accomplish everything. Even if there was such a device it doesn’t mean it will be able to create anything or that you’d be able to construct something larger out of many constituent smaller things. I also think there are some diminishing returns to this stuff that are not taken into account.
I’m all about wild speculation and have no problem with doing it. I usually like to think 4th dimensionally about all possibilities however weird. I just find this MNT to be slightly unrealistic about what products you can create.
Made a few typos on my post sorry. Third sentence should read “So I’m certainly aware of reaction kinetics”. First sentence, second paragraph should read. “I’m not saying this couldn’t, in theory, lead to more/better products”. Sorry if I made any other mistakes.