Hacker News new | past | comments | ask | show | jobs | submit login
Drexler–Smalley debate on molecular nanotechnology (wikipedia.org)
24 points by luu on Aug 10, 2015 | hide | past | favorite | 15 comments



There is something ironic in Drexler's invocation of biology to defend the plausibility of his vision, as it also suggests that biochemical and biologically-inspired methods might well be more appropriate than his approach of attempting to scale down human-scale machinery to a size where molecular forces and quantum effects are significant. So far, it is the former, and only the former, that has demonstrated progress.

Drexler's approach seems akin to building a computer by using micro-fabrication techniques to create a network of tiny relays.


A computer made of nanoscale mechanical latches, made using nano-fabricators, was in fact one of the examples Drexler proposed. Supposedly it would perform better than transistor-based ones. [http://www.halcyon.com/nanojbl/NanoConProc/nanocon2.html]

But isn't current computers an example of microscale fabrication that worked without being biochemical or biologically-inspired? They are made using optical etching, completely different from anything biology does.


> But isn't current computers an example of microscale fabrication that worked without being biochemical or biologically-inspired? They are made using optical etching, completely different from anything biology does.

Yes, and as such, this is another example suggesting that the road to nano-technology is not through the shrinking of macro-scale machinery.

This may be moot, however, as this is not a self-replicating technology, unlike either Drexler's hypothetical or biology's actual molecular assemblers.

Drexler's claims about the feasibility, let alone the performance, of his proposed computer strike me as highly speculative and sometimes self-serving (see, for example, the section on thermodynamic reversibility.) Furthermore, computer performance has made huge advances since the devices (Z80, Motorola 68000) he was comparing to.


Well, Drexler has always, or at least with the publication of his Engines of Creation, used analogies like the nanomachines known as sperm and egg was an existence proof that at some level his style of nanotech was demonstrably possible (or something like this, not sure I've re-read it since it came out in the '80s).

His take on Smalley's "fat fingers" criticism was that it was a straw man, something he never proposed, and the ribosome proved that Smalley's general approach contradicted it.

This all should be in the context of Drexler's claim that the chemistry community co-opted the publicity he created with his expositions on his style of nanotech to get/keep almost all the funding it might attract, especially from them. Plenty of people, including at least some conventional chemists, also feared grey goo and other possible threats from Drexler style nanotech, no doubt including the macro ones, famously Bill Joy in the programming community, and are quite determined to suppress research in it. So without the right background, which I only started to get, it's hard to judge their claims of it being "impossible" or the like.


I agree that Drexler won the debate (or at least that aspect of it), mainly because Smalley overstated his case.

I think the grey goo argument was directed against Drexler not because of his 'mechanical' approach to nanotechnology, but because of his claims that it would produce replicators that are far faster and more efficient than biological ones.


Errr, I don't know who came up with the concept, but it was prominently featured in Engines of Creation by that name.

It is a legitimate concern, but a few years later, when the nanotech conference was at MIT, I remember him saying that it looked easily solvable for the good guys, e.g. make assemblers depend one or more chemicals not found in nature in significant quantities. The problems of bad actors, well, no one has any better solution to them than we've ever had.

That's one of the problems I see with the chemistry etc. community's suppression of nanotech development. The first entity that both develops nanotech and decides to take over the world will do so unless we get our act together (something he's mentioned as well, I think in Engines).


Smalley was obviously right because Drexler thought that nanotechnology would eventually become something like mechanical engineering. That does not work, because quantum mechanics are the default nanoscale rules.

However, Drexler was able to contradict himself enough to make an interesting point about molecular assembly in ribosomes; it just does not have the same atomic precision as peptide synthesis. Therefore, nanotechnology is now firmly in the discipline of chemical engineering.

In that article, I can find one quote from Smalley which was certainly untrue at the time:

"biology is wonderous in the vast diversity of what it can build, but it can't make a crystal of silicon, or steel, or copper, or aluminum, or titanium, or virtually any of the key materials on which modern technology is built."

Shellfish nucleate crystals of silicon, and bacteria nucleate magnetic nanoparticles of iron. That much was obvious to geologists during Smalley's lifetime, but one could still construe chemical skepticism.

May the victor of the debate rest in peace; and let K. Eric Drexler correct himself as much as he wants.


> Smalley was obviously right because Drexler thought that nanotechnology would eventually become something like mechanical engineering. That does not work, because quantum mechanics are the default nanoscale rules.

Non sequitur. First off, quantum mechanics apply at all scales, not just the nanoscale. The question is whether one can model these those machines using classical models instead of the much more computationally intensive quantum mechanical models. Part of the reason the proposals focus on diamondoid structures is that the rigidity of the bonds greatly limits the degrees of freedoms, making molecular mechanics approach accurate enough to model various designs.


Talk about Non sequitur!

It does not seem like you understand my use of the word 'default'; because scales are nested, the nanoscale is everywhere. One can design molecular machines using classical models; but those models are chemical rather than mechanical. The diamondoid structures are the fruit of carbon research; and the molecular mechanics has gotten much computationally cheaper, especially with regards to biomolecules. The motivation for using carbon is the modularity, not rigidity; albeit biomolecules are even more modular than carbon alone.


The rigidity buys you a lot. It's not just easier modelling, it's also higher reliability. If your bond is rigid, you don't have to worry too much about heat causing errors in the synthesis.


The rigidity is the expensive part. Modeling has gotten easier with parallel distributed computing. Reliability is of qualia that I would only attribute to platinum group metals. If heat causes errors in an organic synthesis, then that reaction mechanism is not selective enough for the desired purpose. The flexible bonds in biopolymers serve as counterpoint to the notion that bond rigidity is essential to stochastic nanotechnology. On the contrary, self-assembly requires some bond flexibility.


Rigidity helps you for the same reason using a jointed arm is easier than controlling a tentacle. Limiting the number of degrees of freedom can make things a whole lot easier.

Yes, there are interesting designs with biopolymers, but they probably cannot get you the type of atomically precise manufacturing that you need for most "cool" MNT applications.


It isn't fundamentally easier to control an arm than a tentacle; That is human intuition based on the fact that we have arms. Limiting the number of degrees of freedom is something done at the level of experimental design for the production of chemical libraries. DNA is already a programmable medium for nanorobotics, and skilled genetic engineers can do atomically precise peptide biosynthesis for the sake of NMR spectroscopy. It does not seem as though you are informed of bionanotechnology in the present.



Octopodes do not beg to differ, they tergiversate to genuflect.

The thing to recognize is that tentacles predate arms, legs, or fins by millennia. Jellyfish tentacles have fewer kinds of parts than any arm, hence they constitute simpler machines.

In the paper you cite, humans are describing cephalopods with a model based on their own anatomy; it is not the other way around, as you claim.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: