While your general points are good, there's some flaws in your reasoning. I agree that there is a bottom limit to the size of an autonomously replicating unit. However, using a bacterium as the bottom size limit is not an airtight case. Contrary to popular belief, evolution does not find the globally optimal solution to a problem. Rather, evolution often finds itself bound by the history of previous changes and is locked into suboptimal solutions.
For example, if our genetic code had more bases, our genome could be stored in a much smaller volume and be replicated with much less energy required. However, evolution happened to pick 4 bases and now we're stuck with it. (although some researchers at the Scrips institutes, I think, are working on adding some new bases for bioengineering purposes)
Furthermore, biology is largely limited to 20 amino acids for its building blocks. If the fundamental building blocks of proteins were more chemically diverse or allowed things like branched proteins, it is quite likely that most enzymes would be radically smaller. The functions of most enzymes are contained in a tiny fraction (a few %) of the total residues, the rest is largely filler that results from having a randonly generated linear polymer act as the backbone.
If we could rationally design life, the minimal size for a self-relicating organism would be drastically reduced. What's the size of this organism? I don't know. AFAIK, noone's ever worked out the mathematics for self-replication. Obviously, there is some mathematical value of complexity which represents the absolute minimum requirement for a self-replicating organism in a non-hostile environment (obviously the complexity differs from a construct that replicates in a lab vs one that replicates in the 'wild') which corresonds to some spercific atomic structure.
BUT, you say, what about an arbitrary assembler? This too is probably possible - I point towards human society in general. It can be treated as a giant organisms of sorts that is capable of producing, eventually, just about any imaginable material object. Obviously, we can't make arbitrarily ordered materials just yet but we're well on the way there, nanoassemblers or no. So, let's take a hypothetical human race 200 years from now and take that as an example of a universal assembler. Clearly, there must be some sort of minimal general assembler as there is a minimal self-replicator. Will that general assembler be a single tiny robot arm picking up atoms like Drexler envisions? Almost certainly not. The informational density required is simply too much for any simgle small assembly of atoms. However, it is easy to envision tiny robot arms that are specific to particular tasks - working around the sticky fingers problem.
You need to move medium size +2 valence transition metal atoms? Use arm 32-A94. Want to start assembling branched olefins? Use arm 22-C543. A general purpose assembler is probably going to more closely resemble a factory than a cell when all's said and done.
Drexler's purely mechanical approach to nano is, IMO, rather naive towards how messy real chemistry is. These robot arms I mention will probably look more like tethered enzyme complexes than anything out of a car factory. Smalley is correct to an extent - Drexler's vision is way too overoptimistic. However, I do think that some sort of assembler tech will be available within the next century.
OK, I'm gonna be a nitpicker here but this is buggin' me:
1: viruses don't eat *anything*, they're really not even alive. Bacteria eat stuff, not viruses. 2: Silicon is what sand and computer chips are made of. Silicone is what boob implants and bathtub caulk is made of. Silicon != silicone!!!!!1!! ARRRGH!!!
Ah, my bad, I thought that the SPL canned their sale this year as well.
Of course, if I wanted to be pendantic, I could just say that the KCLS is the suburbS of Seattle rather than a suburb but that would be, well, pendantic.
Yeah, and Adams also predicted that we'd use the EPR effect for instantaneous communication and that evolution would be proven wrong which I'm not holding my breath on.
Wow, I'm amazed,/. and no one's mentioned the link at the bottom of the article about S. Korea going open source?
As for N. Korea, I'll sit this one out as my Mom grew up in Pusan and can still remember the war. I'll just leave it at - Plato's republic is *not* a good basis for a real-world government.
Good article, As much fun as it is to be able to just walk down to a book sale, it makes much more sense to do this sort of thing online for both the customers as well as the library.
However, it was the Seattle public library which moved its annual book sale to Amazon starting this year, not a suburb.
Uh, the feature size of the ship has absolutely nothing to do with the radiation coming out of it. Your monitor releases X-rays (mostly blocked by the lead they put in the CRT glass) because of Brehmstrahlung (sp?) radiation from the interaction of high energy electrons with the inside of the CRT. The same process is used in an X-ray machine at the doctor's office w/o shielding. What you'll get is radio frequency emissions with the same frequency as the clock speed of the CPU. At a THz, your emissions are in the microwave band which will be nicely contained by the case. (although it might give a whole new meaning to the ability to cook an egg on a CPU) A very rough calculation I just did in gives ~300-500 THz as the clock speed recquired to even emit visible light.
No need to pull out the lead apron or tinfoil hats just yet.
Actually, IIRC, the flaw was more like an inch or two. They had a set of spacers or something that were added in behind the primary mirror but they forgot to account for the focal length change when grinding the mirror. And, of course, they never bothered to actually point the telescope at something and check if was in focus...grr.
Something as small as a paint fleck wouldn't have affected the primary mirror, it's too big. A lens or mirror inside the camera assembly, *that* might be a different story since the components are smaller there. But, the camera was fine. The fix required that the problem be corrected in the camera as it was replaceable, but I think that the fix had to be somewhat incomplete beause of certain optics limitations and gave us the somewhat truncated field of view we now have.
Re: 9) TEAM
I'm pretty sure this is a typo, modern TEMs can easily see down to about 0.2 nm. The aberration correctors that have been recently developed should be able to resolve images down to 0.05 nm, not 0.05 microns.
However, this is misleading. You won't be able to see many materials like proteins at this resolution because of radiation damage. The electron/energy flux in a FEG TEM is roughly equivalent to holding your sample 30 feet from ground zero of a nuclear detonation. (yes, I realize that a nuke puts out X-rays, not electrons, this is just raw energy flux.) STM and AFM regularly 'see' under 1 nm but the same caveat applies, the sort of stuff you can image at 1nm resolution is very limited. eg: imaging flat Si wafers at 0.1 nm using STM is relatively easy but looking at soft, squishy proteins, you really can't get below 1nm regularly and even then it usually looks like a ruined fried egg.
I worked with a postdoc who was doing high res EEL spectoscopy of carbon nanotubes and even on those, electron radiation damage was high, he had to drop the 200 keV scope down to 100 keV - he realized that the nanotubes were disintegrating from the radioation damage - including electron-positron pair formation.
What I don't understand is why this is listed as a large facility sort of thing. A 200 keV TEM with monochromator and aberration corrector only (lol) costs a few million, not the billion dollar range we're seeing for the other projects.
Slashdot Mirror
While your general points are good, there's some flaws in your reasoning.
I agree that there is a bottom limit to the size of an autonomously replicating unit. However, using a bacterium as the bottom size limit is not an airtight case. Contrary to popular belief, evolution does not find the globally optimal solution to a problem. Rather, evolution often finds itself bound by the history of previous changes and is locked into suboptimal solutions.
For example, if our genetic code had more bases, our genome could be stored in a much smaller volume and be replicated with much less energy required. However, evolution happened to pick 4 bases and now we're stuck with it. (although some researchers at the Scrips institutes, I think, are working on adding some new bases for bioengineering purposes)
Furthermore, biology is largely limited to 20 amino acids for its building blocks. If the fundamental building blocks of proteins were more chemically diverse or allowed things like branched proteins, it is quite likely that most enzymes would be radically smaller. The functions of most enzymes are contained in a tiny fraction (a few %) of the total residues, the rest is largely filler that results from having a randonly generated linear polymer act as the backbone.
If we could rationally design life, the minimal size for a self-relicating organism would be drastically reduced. What's the size of this organism? I don't know. AFAIK, noone's ever worked out the mathematics for self-replication. Obviously, there is some mathematical value of complexity which represents the absolute minimum requirement for a self-replicating organism in a non-hostile environment (obviously the complexity differs from a construct that replicates in a lab vs one that replicates in the 'wild') which corresonds to some spercific atomic structure.
BUT, you say, what about an arbitrary assembler? This too is probably possible - I point towards human society in general. It can be treated as a giant organisms of sorts that is capable of producing, eventually, just about any imaginable material object. Obviously, we can't make arbitrarily ordered materials just yet but we're well on the way there, nanoassemblers or no. So, let's take a hypothetical human race 200 years from now and take that as an example of a universal assembler. Clearly, there must be some sort of minimal general assembler as there is a minimal self-replicator. Will that general assembler be a single tiny robot arm picking up atoms like Drexler envisions? Almost certainly not. The informational density required is simply too much for any simgle small assembly of atoms. However, it is easy to envision tiny robot arms that are specific to particular tasks - working around the sticky fingers problem.
You need to move medium size +2 valence transition metal atoms? Use arm 32-A94. Want to start assembling branched olefins? Use arm 22-C543. A general purpose assembler is probably going to more closely resemble a factory than a cell when all's said and done.
Drexler's purely mechanical approach to nano is, IMO, rather naive towards how messy real chemistry is. These robot arms I mention will probably look more like tethered enzyme complexes than anything out of a car factory. Smalley is correct to an extent - Drexler's vision is way too overoptimistic. However, I do think that some sort of assembler tech will be available within the next century.
OK, I'm gonna be a nitpicker here but this is buggin' me :
1: viruses don't eat *anything*, they're really not even alive. Bacteria eat stuff, not viruses.
2: Silicon is what sand and computer chips are made of. Silicone is what boob implants and bathtub caulk is made of. Silicon != silicone!!!!!1!! ARRRGH!!!
Ah, my bad, I thought that the SPL canned their sale this year as well. Of course, if I wanted to be pendantic, I could just say that the KCLS is the suburbS of Seattle rather than a suburb but that would be, well, pendantic.
Yeah, and Adams also predicted that we'd use the EPR effect for instantaneous communication and that evolution would be proven wrong which I'm not holding my breath on.
Wow, I'm amazed, /. and no one's mentioned the link at the bottom of the article about S. Korea going open source?
As for N. Korea, I'll sit this one out as my Mom grew up in Pusan and can still remember the war. I'll just leave it at - Plato's republic is *not* a good basis for a real-world government.
Good article, As much fun as it is to be able to just walk down to a book sale, it makes much more sense to do this sort of thing online for both the customers as well as the library.
However, it was the Seattle public library which moved its annual book sale to Amazon starting this year, not a suburb.
Uh, the feature size of the ship has absolutely nothing to do with the radiation coming out of it. Your monitor releases X-rays (mostly blocked by the lead they put in the CRT glass) because of Brehmstrahlung (sp?) radiation from the interaction of high energy electrons with the inside of the CRT. The same process is used in an X-ray machine at the doctor's office w/o shielding.
What you'll get is radio frequency emissions with the same frequency as the clock speed of the CPU. At a THz, your emissions are in the microwave band which will be nicely contained by the case. (although it might give a whole new meaning to the ability to cook an egg on a CPU) A very rough calculation I just did in gives ~300-500 THz as the clock speed recquired to even emit visible light.
No need to pull out the lead apron or tinfoil hats just yet.
Actually, IIRC, the flaw was more like an inch or two. They had a set of spacers or something that were added in behind the primary mirror but they forgot to account for the focal length change when grinding the mirror. And, of course, they never bothered to actually point the telescope at something and check if was in focus...grr.
Something as small as a paint fleck wouldn't have affected the primary mirror, it's too big. A lens or mirror inside the camera assembly, *that* might be a different story since the components are smaller there. But, the camera was fine. The fix required that the problem be corrected in the camera as it was replaceable, but I think that the fix had to be somewhat incomplete beause of certain optics limitations and gave us the somewhat truncated field of view we now have.
Re: 9) TEAM
I'm pretty sure this is a typo, modern TEMs can easily see down to about 0.2 nm. The aberration correctors that have been recently developed should be able to resolve images down to 0.05 nm, not 0.05 microns.
However, this is misleading. You won't be able to see many materials like proteins at this resolution because of radiation damage. The electron/energy flux in a FEG TEM is roughly equivalent to holding your sample 30 feet from ground zero of a nuclear detonation. (yes, I realize that a nuke puts out X-rays, not electrons, this is just raw energy flux.) STM and AFM regularly 'see' under 1 nm but the same caveat applies, the sort of stuff you can image at 1nm resolution is very limited. eg: imaging flat Si wafers at 0.1 nm using STM is relatively easy but looking at soft, squishy proteins, you really can't get below 1nm regularly and even then it usually looks like a ruined fried egg.
I worked with a postdoc who was doing high res EEL spectoscopy of carbon nanotubes and even on those, electron radiation damage was high, he had to drop the 200 keV scope down to 100 keV - he realized that the nanotubes were disintegrating from the radioation damage - including electron-positron pair formation.
What I don't understand is why this is listed as a large facility sort of thing. A 200 keV TEM with monochromator and aberration corrector only (lol) costs a few million, not the billion dollar range we're seeing for the other projects.