An anonymous reader, reacting to my Tech Central Station piece on the new assembler animation, writes:
- Has anyone actually run a *Gasp* molecular simulation of these devices to study how or if they might work. The technology is there and the fact that none of the major proponents have performed these computational experiments is dismaying: it means that they either are not smart enough to perform the simulations or they have performed them and found the results proving their beliefs wrong.
I spent three years helping to lead the charge in bringing these disparate business efforts together and calling them a nanotech "industry." But, ultimately, this kind of business-boosting coverage falls short. Uncritical coverage is not only bad journalism, but it's also bad for the industry. The stories NOT being told are screaming on the sidelines. So, circumstances pushed me into returning to what I wanted to do from the beginning -- cover nanotechnology as a broader business, political, cultural and societal story, regardless of whether there's a five-year profit plan involved. For that matter, regardless of whether it fits neatly into the U.S. government's nano plans or those of the few star researchers who have the government's ear.
This is a long-winded way of answering the criticism, I know, but I thought it was important to spell out some of my base assumptions when I cover these stories. I'll sit back and enjoy the debate over what is physically possible and what is not. I'm not qualified to decide who is correct. My job is to draw attention to different points of view. Despite the negative PR generated by various interests against molecular manufacturing, I've spoken to enough, and varied, sources to realize that this vision of nanotech is being painted as physically impossible not because there is any proof that it is, but rather because it is politically and economically convenient for a few people to say that it is. And if you want to get in on all this government nanotech funding (other research projects might be up on the chopping block, but nanotech is certainly not), you'd better stay silent on "nanofactories" and forget about getting it funded, no matter what you really believe.
However, like I write in Tech Central Station, there is a nanotech "network-in-exile" working on this vision, including the molecular simulations that the reader describes. You simply haven't heard about them because they're doing it quietly, without fanfare, with very little government funding and with no media coverage.
Stay tuned to NanoBot. That situation is about to change.
Introduction to Online Journalism: Publishing News and Information
3 comments:
At the moment, doing molecular simulations of a system that large and that complicated is extremely nontrivial at best, and completely impossible at worst. If we want to do a simulation which takes a reasonable account of bonding, then we're pretty much limited to systems of maybe a thousand atoms. If we were to focus our attention on a very small part of the system at a time (say, to study whether two gears can be made to mesh together without, say, sticking together or wearing each other flat) then we might be able to do it, but we'd be looking at many, many, many years of work just to figure out whether the design as shown would work. And if it didn't, there'd be many more years of work trying to figure out whether the problems with the design as shown can be overcome. So my main point is that the previous anonymous shouldn't be under the impression that we can just stick this system into our simulation codes and see whether it works.
I don't think this animation should be seen as a fully-fledged design for something that would actually work, but rather as a sort of wild-assed guess as to the sorts of things that might be possible. If and when an assembler is built, it will probably bear the same sort of relationship to this design as a 747 bears to a 17th-century sketch of what a flying machine might look like.
In conclusion, it's a cute concept, and a cute movie, but it's not yet a real design for an assembler, and it's unlikely to be turning into one any time soon. Designing a working assembler will no doubt take about a zillion man-years of serious research, it's not the kind of thing you can come up with in an afternoon in the animation department.
Perhaps the best feature of this animation is that it helps to communicate the idea that molecular assembly need not involve nanobots.
I'm the original anonymous poster. (I prefer anonymity here more because of the inconvienance of creating an account on Blogger than anything else).
The previous comment has a point. Right now, it is impossible to put an entire 'molecular assembler' into simulation software and get results within a reasonable amount of time (~6 months of 24/7 computing). But we don't need to put the entire device in to see if the concept works.
What is the concept of mechanosynthesis? As far as I can determine, from the Drexlerian description of mechanosynthesis, is that by exerting mechanical work (physically pushing two molecules together) one can catalyze a reaction, breaking or forming bonds, to create a new molecule. A lot of chemists and physicists don't see this happenening for a variety of reasons (and none of those reasons have to do with businesses, the government, money, or God). If someone were to perform a computational experiment where they demonstrate that mechanosynthesis is possible then that would go a LONG way towards convincing people that the overall goal of a molecular assembler is possible. So start small: Can two molecules attached to a scaffold be forced together to catalyze a reaction? It can be less than 1000 atoms (especially in vacuum). One could even perform ab initio calculations, the highest 'grain' simulation. (For the layperson, that's where you calculate the densities of electrons floating around the atoms and see how the densities flow from one molecule to another..the quantum definition of a reaction).
So performing a useful simulation is possible and should be done soon. The whole idea of mechanosynthesis needs to be explored and it can be done with current resources and technology. So my question to the audience: If you perform ab initio calculations and can donate your expertise (and a few thousand SUs), why not try to answer this interesting question?
But, beware: You might find that all of those chemists and physicists out there are skeptical for a reason. It may be that future molecular assemblers don't use mechanosynthesis at all. They might rely on more traditional catalysis chemistry coupled with nano-scale movement of reagents. But that's where academia and business is moving towards: microreactors and microfluidics. And they'll keep making them smaller and smaller because it's more efficient. So we probably won't need a revolution in technology to get to nanoreactors and nanofluidics. It'll just be the slow progression of understanding how physics works as things get smaller and smaller and smaller. And then we might just get our molecular assemblers. Or not.
Good point. I immediately focussed on some of the extraneous details of the device (eg the gears) as these seemed to me to be those least likely to work. But certainly if we're just trying to evaluate the very concept of mechanosynthesis then simulations should be possible.
But the question "does mechanosynthesis (as the previous poster defined it) work?" isn't really a yes-or-no question. It will undoubtedly work fine for some reactions (for instance, bring two O atoms close enough together and they will most definitely form an O_2 molecule) but may not work for others (eg some reactions which involve major configurational changes in reagent molecules, or involve some bizarre intermediate products). I think we'd really need to ask "does mechanosynthesis work for this reaction?" on a reaction-by-reaction basis.
So if a mechanosynthesis device is going to work, it can't just chuck atoms together willy-nilly, it will have to follow a plan which ensures that every step of the process is actually possible. This may or may not place some limitations on what can actually be built, and these limitations may or may not be a serious problem. Again, it will probably be many decades of research (including lots of ab initio simulation) to figure out what is and isn't possible via mechanosynthesis.
- Jorge
Post a Comment