Drexlerian nanotech

[JohnnyYesterday]

What I mean to point out here is that Drexler made no allowances for progressively larger manufacturing implements. This is the crux of the problem: molecular assembly is good for things like transistors, but bad for things like a block of steel. Molecular assembly MUST be integrated with macroscopic processes to be effective. Which is why in some cases, a reconfigurable manufacturing machine may be better than a traditional universal assembler, as it can do things like build said ruler without the added task of molecular synthesis.

Drexler's thoughts moved on from 1986.

He wrote a highly technical book called Nanosystems: Molecular Machinery, Manufacturing, and Computation in '91 that talks about convergent manufacturing to produce macroscopic objects.

This animation contains some errors (don't take it too literally) but it conveys the basic idea of convergent manufacturing.

[Rynn]

I never said that nanotech doesn't work in the first place. I merely noted that the initial ideas were flawed. And as for biology not using progressively larger systems-what do you think digestive organs are for?

Todd didn't say that there was no macroscale assembly/disassembly in nature. There are obvious examples from eating to the construction of insect hives. Just that they both have a place.

And we're nowhere near making wormholes either, yet we know exactly what does and doesn't work. Finally, remember that nanoassembly in air is pretty difficult from a mobility point of view-they don't have little thrusters, after all. And finally: biological cells are a highly controlled environment! Would bionano work outside of them? Has anyone seriously analyzed its feasibility? Don't get me wrong-I really hope it pans out, but I have doubts. After all, all the transhumanists and futurists before us had a bad tendency to get carried away.

Air? Who said anything about air? A fabricator would contain a myriad of chambers with different controlled environments containing different specialised machinery (molecular-macroscale). If you're talking about self replicating technology that can operate in the wild then again you can take a leaf out of nature's book: have cellular constructs with controlled internal environments that alter their external environment as best as possible. "Grey goo" could be a lot like slime mold, but artificial (note that I don't think Grey goo as originally conceived makes sense, a universal eater is orders of magnitude more complicated than a single self replicating breed. Drexler also discussed this after his original publication).

I guess this is pretty far from the original topic by now, but I have a tendency to minirant. Sorry 

It would be good if you could define your terms here. It feels like we're talking about different things and then suddenly moving on with no indication why. We were talking about fabs specifically, now we seem to have moved on to air without noting any of the previous discussion. This would also really help if you're going to start a thread by saying something categorically "doesn't work" without really discussing what you mean.

[Alphadon]

Let me clarify, then. By "drexlerian nanotech" I mean not things derived from Drexler's ideas, but those ideas themselves. Which, as I understand it, is exactly the individual nanoscale robots operating independently that you mentioned, often in an uncontrolled environment. However, my knowledge of his work comes entirely from Engines of Creation, so I suspect that my idea of it is quite incomplete. Anyway, you must understand that I wholeheartedly agree with the idea of an OA nanofab, and I am simply pointing out that Drexler's original work, and by extension most depictions of nanotech, do not incorporate these ideas.
Also,

(note that I don't think Grey goo as originally conceived makes sense, a universal easter is orders of magnitude more complicated than a single self replicating breed. Drexler also discussed this after his original publication).

[Rynn]

Let me clarify, then. By "drexlerian nanotech" I mean not things derived from Drexler's ideas, but those ideas themselves. Which, as I understand it, is exactly the individual nanoscale robots operating independently that you mentioned, often in an uncontrolled environment. However, my knowledge of his work comes entirely from Engines of Creation, so I suspect that my idea of it is quite incomplete. Anyway, you must understand that [i]I wholeheartedly agree with the idea of an OA nanofab, and I am simply pointing out that Drexler's original work, and by extension most depictions of nanotech, do not incorporate these ideas.

I read Engines in my early teens so I really don't remember much of it but...really? He only talked about nanobots operating in the wild? Regardless I don't agree that "most depictions of nanotech do not incorporate these ideas". Sure a lot of science fiction concerns nanobots but usually in an environment that is appropriate for them, such as for medical purposes. I can't think of any examples off of the top of my head in which swarms of nanobots act as a fab.

(note that I don't think Grey goo as originally conceived makes sense, a universal easter is orders of magnitude more complicated than a single self replicating breed. Drexler also discussed this after his original publication).

Oops

[Drashner1]

Let me clarify, then. By "drexlerian nanotech" I mean not things derived from Drexler's ideas, but those ideas themselves. Which, as I understand it, is exactly the individual nanoscale robots operating independently that you mentioned, often in an uncontrolled environment. However, my knowledge of his work comes entirely from Engines of Creation, so I suspect that my idea of it is quite incomplete. Anyway, you must understand that I wholeheartedly agree with the idea of an OA nanofab, and I am simply pointing out that Drexler's original work, and by extension most depictions of nanotech, do not incorporate these ideas.

Actually, this isn't correct (or may be seen as a matter of interpretation). I have Engines of Creation and Drexler's other books in hardcopy and while it's been a while since I last read them, as I recall, Drexler addresses this as follows:

E of C:

a) An example of a single replicator floating in a vat of chemicals - he states that, within a matter of days, the replicator and its progeny could consume a mass equal to the sun and all the planets - if the vat of chemicals didn't run dry long before.

b) A description of the creation of a rocket engine via nanotech in which a 'seed' (mainly used for command and control) is lowered into a vat which is then filled with a mix of feedstock and nanobots. The bots proceed to link up first to the seed and then to each other under the seed's command to form a sort of 3D mold or armature (similar to OA descriptions of engenerator tech) which then proceeds to grab feedstock elements out of the mix and assemble them together into the overall rocket engines. Other elements of the armature act to move feedstock through the structure and to pump liquid through for cooling, etc. Different nanobots lay down feedstock matter in different types and ways to form diamondoid, sapphiroid, etc. depending on the structural and temperature needs of the engine and all based on the design coded into the seed. After less than a day, the engine is complete, the nano and feedstock mix is pumped out, the engine is hosed off, and removed from the vat to be sent where it is to be used. Drexler also mentions that in principle you could just make an entire space shuttle this way.

c) A single line in the section on possible dangers of nanotech in which Drexler states that a nanotech based 'plant' that operated more efficiently than natural plants could outcompete them and essentially consume the biosphere.

Unbounding the Future:

a) Drexler spends more time talking about possible dangers of nanotech, during which he explains that he no longer feels that 'grey goo' is a serious threat, basically because the odds of a replicator accidentally escaping in 'the wild' and being able to survive and flourish there are about the same as an automobile somehow getting loose in the woods and figuring out how to survive off of tree sap rather than gasoline. While he doesn't say that grey goo replicators or nanobots operating outside of a controlled environment are impossible, he does make it clear that such devices would both represent a much higher level of technology than what he is considering and that they would almost certainly have to be deliberately and carefully designed to operate 'in the wild'.

b) Drexler devotes quite a bit of space to 'Desert Rose Industries' a mom and pop factory complex that manufactures a huge range of goods. There is a scene in which a big batch of tents are made to send to the site of a disaster. The factor is described as containing big vats into which feedstock (some of it pre-fabbed at other dedicated facilities using specialized equipment to make components -some of it just raw materials for construction or chemical energy) if fed and then converted into finished products by various forms of nanobots.

c) Drexler spends a portion of the book talking about various more 'factory floor' type nanodevices, including mills and motors and such.

Nanosystems:

Drexler goes into deep details and lots of math about how nanotech might work, focusing nearly all of the book on individual devices that would operate in controlled environments and more like RL robots in that they are fixed to a location and materials and components pass by them via some form of conveyor system rather than free floating nanobots moving themselves around.

Putting these together, Drexler does initially talk about using free moving nanobots, but these are still operating in an enclosed environment. While he does mention what has come to be called 'grey goo' it was first in the context of a thought experiment intended to demonstrate the power of exponential growth and self-replicating technology - not as a serious threat or way that nanotech would work (unless solar system masses of chemicals are something we have lying around and I just missed the meeting? ). He does mention the danger of self replicating tech (plants) based on nano, which could be seen as a kind of grey goo - but again he doesn't spend much time on it and he definitely doesn't describe it in detail or make it his main point - the popular imagination of the general population has done that by fixating on this one aspect and running with it.

In later works Drexler goes out of his way to argue against grey goo scenarios and while he again depicts more 'free moving' nanobots, they are again operating in a controlled environment of a vat or tank.

By the time he gets to Nanosystems, Drexler isn't talking about nanobots much at all and talks almost exclusively in terms of controlled environment 'desktop factories' - a theme he continues with in later years from his website in which he devotes some articles to the convergent assembly type systems that Rynn mentioned earlier.

The upshot of all this is that Drexler never argued for the idea of free range nanobot swarms creating things out of whatever is handy in the natural environment, although he does acknowledge (briefly) that such are probably possible in principle - but that they would be very advanced designs (possibly matched or exceeded only by artificial immune systems in their level of sophistication) and there would be some question of why such would be needed for most cases. The popular imagination has however run with the free range nanobot idea and so people seem to think that's the only thing Drexler came up with - which is incorrect.

As mentioned, my personal copies of Drexler's works are in hardcopy, but I believe there are (or at least were) online copies available from which I can either pull direct quotes or provide links to if folks would like to look at the relevant sections for themselves. Or I can provide page references and some level of quotations (that's a lot of typing) if I can't locate the online versions.

Todd

[Alphadon]

(unless solar system masses of chemicals are something we have lying around and I just missed the meeting? ).

Well, I'm sure if we ask Binah nicely...
Anyway, who did think up this idea of free-ranging nanoswarms (which we rather egregiously used at one point)?

[Rynn]

Excellent summary Todd Note that the free-floating vat model is the same principle as a lot of industrial biotechnology. We've been producing things like insulin for decades this way, using tanks of genetically engineered bacteria to mass produce human insulin that is then filtered off for medical use. The field of synthetic biology is heavily involved in new methods that could allow for vats creating all kinds of products. Almost all raw chemicals but that would be a huge boon for various fields. There's a company called Evolva that's been running a couple of years that hopes to produce all kinds of ingredients that are usually un-sustainable cash crops in vats of engineered cells that would be cheaper and cleaner (haven't heard about them in a while but the principle is sound).

You could likely do much better in OA but it goes to show that the FFV model is entirely valid in a range of circumstances.

[iancampbell]

IIRC Drexler discussed the point about macroscale objects being made using nanotech methods; the point being that for most purposes, objects would be made from standardised parts which might well have been made elsewhere. This wouldn't quite give the ultimate efficiency that the same sort of object being made atom by atom might result in; but given the enormous design margin (diamondoid is orders of magnitude stronger than steel weight for weight, for example) the result would be good enough.

There is also the point that the standardised parts might well come in a hierarchy of sizes, all the way from the size of large molecules to objects massing tons.

I imagine that another advantage of nanotech is that parts would be joined together much better than currently. The weak points in today's manufactured objects are often the joints between their parts; nanomanufacturing would result in objects where one really couldn't see the join.

[Avengium]

. There's a company called Evolva that's been running a couple of years that hopes to produce all kinds of ingredients that are usually un-sustainable cash crops in vats of engineered cells that would be cheaper and cleaner (haven't heard about them in a while but the principle is sound).

http://www.evolva.com
They still exist (or They keep existing).

[Rynn]

IIRC Drexler discussed the point about macroscale objects being made using nanotech methods; the point being that for most purposes, objects would be made from standardised parts which might well have been made elsewhere. This wouldn't quite give the ultimate efficiency that the same sort of object being made atom by atom might result in; but given the enormous design margin (diamondoid is orders of magnitude stronger than steel weight for weight, for example) the result would be good enough.

There is also the point that the standardised parts might well come in a hierarchy of sizes, all the way from the size of large molecules to objects massing tons.

It's a nice idea though I'm not totally convinced by this lego style of goods construction. but along similar lines I imagine that, in OA at least, fab template libraries would contain a vast amount of alternative designs for the same product using different materials. This would be especially useful for field-fabs that exploration probes and colony projects might use. Area lower on carbon as you expected but higher on silicon? There's a near identical design using that material instead.

I imagine that another advantage of nanotech is that parts would be joined together much better than currently. The weak points in today's manufactured objects are often the joints between their parts; nanomanufacturing would result in objects where one really couldn't see the join.

Quite. It's entirely plausible that objects may be chemically bonded into one solid structure. Whether that's by having smart surfaces that expend energy to bond with other surfaces when held together or through tunnelling bots that eat material in front of them and lay down behind (a process similar to bone healing). A lot of OA projects could be quite seamless.