Molecular Manufacturing Review

Someone asked me in email about the difference between nanopowders and molecular manufacturing. My response seems suitable for a blog post:

For a very brief overview of molecular manufacturing, see http://crnano.org/whatis.htm.

Nanopowders are small, relatively simple structures made by relatively large machines. Molecular manufacturing uses small machines (in fact, nanoscale machines) under computer control to make intricate products via molecular additive techniques. The range of products includes small machines, and it turns out that you can scale up production fairly quickly, due to scaling law advantages of small machines.

Computer-controlled additive manufacturing at the level of molecular bonds implies general-purpose manufacturing: a single manufacturing system could build a wide range of products depending on what blueprints were fed in. Complete products, that is; not merely shapes like today's rapid prototyping systems. And, again thanks to scaling laws and a few physics tricks, the products would have orders of magnitude higher performance than today's versions.

Molecular manufacturing is, so far, in the early stages of research; as an industrial process, it doesn't exist yet. But when it's developed, watch out! I expect it to do for manufacturing what computers did for data processing. We used to have rooms full of people doing arithmetic and even writing by hand. In the future, we will "used to have" factories full of people and special-purpose machines making individual components and then assembling them into products.

The implications of molecular manufacturing go beyond the economic impacts. One implication is rapid R&D of highly advanced aerospace technology, including weapons--it looks to me like this could be a major factor in geopolitics. It is implications like this that are CRN's reason for existence.

Chris

Comments

[Sam]

On what do you base your conclusion? You state that, "The range of products includes small machines, and it turns out that you can scale up production fairly quickly, due to scaling law advantages of small machines" what study, with empirical not theoretical, experimentation shows this? I haven't seen anything that definetly proves that Molecular Manufaturing as you state it is viable.
It just seems that your ideas are based more on wishful thinking and a optimistic ideology that in the future all things will be figured out, than true hard facts. Seriously I'm not trying to bash it, I just don't see enough scientific conclusions for you to make the jump to where you are.

[Jay]

If you think nanotech isn't possible... let's see the math to prove it.

If you want empirical evidence, just take a look at the mirror to see a functioning nanomachine right there.

[michael vassar]

The phrase "with empirical, not theoretical, experimentation" implies an epistemology which holds all things impossible until they have been accomplished. It's usually worthless to argue against an epistemology. This is the flip side of the belief that experimental data is no evidence without a theoretical argument, a belief common among so-called "skeptics" in the face of placebo effects and medical procedures that aren't endorsed by the AMA. In fact, both experiment and theory can provide strong Bayesian valid evidence for a proclamation, and if they were honest, proponents of weird epistemologies would admit this. In fact, they do something equivalent when they refuse bets which take advantage of the holes in their epistemologies.
Sam, I can propose a number of machines which I have great confidence can be constructed and will achieve certain functions. None of these machines have ever been constructed before, so there is no empirical evidence that they work. According to your stated beliefs, I have no basis for thinking they will work. I, disagree. Why don't we place a wager on whether these designs can in fact be constructed and accomplish the goals that my "theoretical studies" tell me they can?

Sadly, while it may be possible to win a debate with a person in this manner, the scientific community is not so easily pursuaded. Oh well. Complaining doesn't help anyone.

[Sam]

I know nanotech is possible, that wasn't what I was saying. I was addressing Molecular Manufacturing as stated above.
And unfortunately, looking in the mirror I do not see anything that resembles a man made nanomachine, to be honest. Don't get me wrong I believe molecular nanometer technologies exist and expand from here. What I'm saying is that, it isn't good to refer to something that hasen't been tested or even created yet and know what the ramifications will be.

[Sam]

Michael, I support you entirely on using the hypothesis you've clearly and eloquently mentioned. But shouldn't we see if we can make a nanodevice that can be used in molecular manufacturing as stated above? Chris is basing his future expectations as if molecular manufacturing (of what type I don't know) is here today.
I say let's spend the time and energy on finding the tools and the process to actually make molecular manufacturing instead of "putting the cart before the horse" so to speak as I believe Chris is doing.

[Chris Phoenix, CRN]

Sam, it would be nice if we could afford to wait until the situation developed before we made any plans. However, that is not always safe.

CRN is founded on the idea that the capabilities of molecular manufacturing can be broadly predicted even before it is implemented. This is the same way that people knew the Golden Gate Bridge would work before they built it.

The scaling law advantages of small machines are not a matter of debate--at least I've never seen it debated. Power density scales inverse-linearly with size: a 100-nm device can have a power density well above 10^12 W/m^3.

Time for a molecular fabrication machine to make its own mass scales as the fourth power of size: a 100 nm machine could make its own mass in about 100 seconds.

The only reason we don't see this kind of performance in today's nanoscale systems is that they are crippled by being immersed in water.

Chris

[Sam]

Chris,
There's a huge difference between knowing we can make the Golden Gate Bridge, which the technology already existed and proven at the time, and one hypothesising about a revolution with a technology that doesn't exist and no one has yet a clear idea how to achieve it.

[Chris Phoenix, CRN]

I'm not sure whether the difference is categorical or just an (admittedly large) matter of degree. Lots of problems had to be solved on-the-fly to build the GG Bridge. The span was the longest in the world at the time, the bridge cables the thickest ever, and many of the architect's peers said it couldn't be done.

Even more problems had to be solved to develop the A-bomb. They didn't even know which uranium enrichment technology they were going to use!

Whether the technology will work as planned, and how to achieve (develop) it, are two separate questions. We were talking about whether it would work as expected. There are several development pathways proposed, but let's stick to one conversation at a time.

Chris

[Sam]

I thought I was. I'm questioning the fact that , at the moment, there is no clear path to creating a device for MNT, How ever that may be defined. I would never say it's impossible, given time, all things are, it's just not as yet.

[Chris Phoenix, CRN]

Sam, you started by questioning my statements on scaling laws and range of products for a relatively mature molecular manufacturing technology.

Now you are questioning the "path to creating a device for MNT."

I'm happy to have either conversation, but I will not try to have them both at once. Pick one.

Chris

[Sam]

My apologies. What I'm trying to state, without having a clear path to MNT, we have no idea what tools and how complex it will be to achieve MNT.
Breaking it down into facts helps define the argument.
1) It must be extremely complex to achieve MNT, since no University/company has achieved it as yet, or is actively pursuing it.
2) Without knowing the true complexity of the matter, we do not know the cost's associated with achieving MNT. It may take several billions of dollars just to be able manufacture one potatoe, not very efficient obviously. And I do know your argument about replication, and I believe that it is to much of an unknown to be factored into. Replication may have costs or complexities that make it unreasonable, in itself.
3) From 1 and 2, we can determine that if the tools and the complexities make MNT to costly, at least at first, the dangers are reduced.

I think the idea of cheap MNT should be taken out of the picture, at least in the beginning. It's sounds to much like what they said about nuclear energy and it being "to cheap to meter". The reality has never lived up to the hope.

[Chris Phoenix, CRN]

Sam, without exponential manufacturing (what you're calling replication), molecular manufacturing basically can't work. It's not that we'll develop MM, and it'll have a low impact, and then maybe later we'll develop EM. If that were the case, I'd be inclined to agree with most of your argument.

But it can't happen that way. Pre-EM, MM will have basically zero impact. Post-EM, MM will have huge impact. In fact, once you've solved EM, you're mostly home free--there's room for improvements in CAD software etc., but the steps from a Merkle fabricator to a Phoenix nanofactory to a flood of products are pretty small.

The technical arguments that I have given for why MM works are directed at EM. You may not believe them, but unless you can find a problem with them, you won't have much luck convincing me to stop using them.

I don't think point 1 of your argument is valid. The universities/companies have never evaluated MM, so they cannot know how difficult it will be to develop.

Chris

[Sam]

Chris,

You have to ask yourself why any university or company has never evaluated MM?! If it (MM) is as inevitable as you say in the next few decades, why do all the experts say otherwise? The idea of Drexlerian MNT has been around since early 80's. and yet no one has come up with even the idea on how to achieve it. I don't say that it won't happen, It will take many many more years than I think most are willing to admit to.

[Chris Phoenix, CRN]

Sam, I don't have to ask; I watched it happen. It's a combination of politics and scientific inertia and bad luck. Some of the top nanoscale tech people were afraid that MM would threaten their funding, due to either competition for funds or public fear. Meanwhile, several top scientists criticized it without ever studying it, and everyone else jumped on their bandwagon. For a while, students who were interested in it were told not even to read about it because it would hurt their careers.

It's not true that all the experts say otherwise. A few experts got a lot of publicity. But even that is tapering off in the past year or two.

NASA Ames studied it for several years, but the program was shut down in the middle of interesting work. At this time, the NAS has a committee studying the technology and its implications carefully. Battelle is also studying the technical side, developing a detailed roadmap with the help of Dr. Drexler.

So far in this discussion, you have generalities and skepticism, and I have facts. That's not very productive.

Chris

[Richard Jones]

Chris, I don't particularly want to get into another flame war with you, but you know that it is not true to say that no one in universities has thought about MNT.
Sam, you might want to look at http://www.softmachines.org/wordpress/?p=130 for another perspective. The debate between Chris Phoenix and Philip Moriarty, an experimental nanoscientist from Nottingham University who has looked at proposals for MNT in some depth, is particularly worth looking at.

[Sam]

Pardon? I do not see any generalities that I've stated? They are facts.
1) Most respectable scientists will not even approach the subject of MNT. Some fear of respectability is in their. I worked on my masters in industrial chemistery, the professor who, back in '95, stated that drexler ideas were "absolute nonsense" (and yes I'm cleaning it up). As he stated Scientist's work on what is, not what could be. To even get MNT out the door you have to get past this impedement, that is a fact. I do not see what anyone has to worry about if no one will even work on MNT.
2) You tout potential issues from, what is basically an idea. By your own admission you've seen respectable scientists turn their collective backs on MNT, you say due to politics and a battle for funding, I say due to the fact that they would loose funding to basically what comes down to an idea. I will sum up my few facts, MNT does not exist, and no one knows how to yet achieve it, or even if its achievable in the short term (10-15 years out).
It's odd, you say you are basing what you've said on facts and yet what you are basing everything your saying on is just an idea.
How can you derive facts from something that does not yet exist?

[Jay]

"I will sum up my few facts, MNT does not exist, and no one knows how to yet achieve it, or even if its achievable in the short term (10-15 years out)."

This is a non-issue. It does not make sense to argue with 'no one knows how to achieve it'.

If anybody knew how to achieve it, we'd have MNT tomorrow.

Whether we know how to achieve it, is irrelevant. What is important, is that people have ideas on how to get to the next level. On every level. Because that is what constitutes progress.

Based on the steady progress in MNT in these recent years, and based on what I've seen from CRNano, I'd say science has a good idea on how to get to the endgoal: MNT.

In Tihamer Toth-Fejel's own words: "What is happening right now is that top-down methods are meeting bottom-up methods, and that is why Chris Phoenix and I, among others, are confident that powerful molecular manufacturing technologies will emerge sooner rather than later."

[benji]

"They are ill discoverers that think there is no land when they can see nothing but sea."
-Francis Bacon

[Sam]

If you want to know your past, look into your present conditions. If you want to know your future, look into your present actions.
--Buddhist Saying