Slashdot Mirror
Nanotechnology: Are Molecular Assemblers Possible?
Roland Piquepaille writes "Two experts in the field of nanotechnology, K. Eric Drexler, Ph.D., cofounder of the Foresight Institute in Palo Alto, Calif., and the person who coined the term "nanotechnology," and Richard E. Smalley, Ph.D., a professor at Rice University and winner of the 1996 Nobel Prize in Chemistry, exchanged open letters about "molecular assemblers" -- devices capable of positioning atoms and molecules for precisely defined reactions in almost any environment. These letters are making the -- long -- cover story of the current issue of Chemical & Engineering News. At the end of this rich exchange of four letters, they still disagree about the issue. Drexler thinks "molecular assemblers" are possible while Smalley denies it. Who is right? Don't count on me to give an answer. This summary contains some forceful quotes from the original letters."
If, in the future, copying physical objects is nearly as easy as copying information on a computer, will corporations lobby to pass laws that make it illegal to do so? In other words, will I be arrested one day for making a copy of my friend's Ferrari?
Richard Feynman talked about nanotechnology way back in 1959--before "nanotechnology" was even a word.
It kind of irks me that the person who coins a word gets more credit than a person who talked about the actual process--nearly thirty years prior.
Read Feynman's talk at the Zyvex Web site.
There is a fundamental obstacle to creating moleular assemblers: What do you make them out of?
Imagine that you were given the task of designing a machine to lay bricks. This probably would not be all that difficult, considering all of the things we already do with robots.
However, the problem becomes much more difficult if I add the stipulation that the machine be constructed entirely from bricks and mortar.
Comment removed based on user account deletion
*sigh* I'm touched.
Also I found it interesting that the usage of Nanotechnology was changed so greatly that the creator of the term accepts the newer phrase 'molecular assemblers' for that process.
The grass is only greener, if you don't take care of your own lawn.
They say its impossible, but isn't DNA essentially just that, and I'm quite sure some lab recently built a transitor from DNA so I'd say its definatly possible.
From reading the letters I don't think Drexler has really addressed the problems raised by Smalley fingers at all, he just tries to brush the problems aside.
Panurge has posted for the last time. Thanks for the positive moderations.
Ribosomes are essentially molecular assemblers that build proteins out of amino acids using instructions from messenger RNA (originally transcribed from the DNA in the nucleus). So, it's not only possible, your cells are doing it as you read this.
[Insert pithy quote here]
Now, I have only a vague understanding of the subject, but from what I read, I was lead to believe that you didn't have one little agent running around like a little gnome (or group thereof) building some complicated structure. You had a sequence of these things which acted like an assembly line. Each agent knows how to slap a specific atom or subset of atoms onto some atomic structure it receives, and only does something when it receives that atomic structure. So there wouldn't really need to be any memory, or very little, since it only does a specific task repeatedly. The thing could almost be stateless.
Again, this is my dim recollection from something I read awhile back, so I bow down before more informed heads.
It doesn't need molecular technology. They already try to come after you even today. See this nice (and real) example:
_ sl r-pl_us.htm
http://www.mb-portal.net/html/news/special/2003
Some guy from Poland "copied" the new Mercedes SLR, long before the real car hits the market. Mercedes tried to buy it from it to get it off the streets. Because that failed, they sue him.
Marc
Of course molecular assemblers are possible. Your body contains billions of them -- ribosomes.
A ribosome (a combination of several large protein molecules) constructs arbitrary protein molecules from individual amino acids according to the instructions on a strand of RNA (copied from DNA). Sounds like a molecular assembler to me.
Now, as to whether they can be made smaller and more flexible than that (nanotech's "universal assembler") is another question -- ribosomes may turn out to be the minimum possible assembler. Or not.
-- Alastair
However, there's also Asimov's Corollary to Clarke's First Law (1977):
Maybe when everyone has their own personal assembler, no one will feel compelled to buy anything anymore. Therefore, the only people in business selling a product will be those selling assemblers. But maybe there will also come a day when the government provides them, too... kinda like phone booths. The phone book could be a directory of things it can make. Anyway, the market would die, but only products. People selling services would still be valued. ;). Maybe if everyone had anything they wanted at the touch of a button, we would all shift our new attention to creating new and better things, instead of the pursuit of money. Of course, money would have to be kept track of electronically, if it still existed, since it could easily be copied. True, the serial numbers would be the same, but it could be spent before it was caught as a double.
Of course, I can imagine that someone would get the idea to copy a person so that they wouldn't even have to pay for services... just make a servant. I think in this case the government would make a law against copying a human- much in the same lines with the cloning issue today- so as not to devalue the human life
Well, that's the end of my rant.. tell me what you think. Also, I have a question. Forgive me for not RTFA, but from what the poster said, it seemed to point at the fact that the assemblers simply rearranged matter. On what level does this happen? i.e. would i be able to make an apple if i threw in some raw glucose, pure water, etc.? would i need even that?
The power of Christ compiles you.
A Random Blog
I read the letters, and skimmed Drexler's "Nanotechnolgy: ..." book.
...". Nanotech is going to make heavy use of indexes like this. Storing all those enzymes, and shipping them one after another to the right place, is going to be SLOW. I suppose you could pipeline your enzyme fetches.
I think Smalley's argument is that for a specific reaction between two molecules, you need something like an enzyme designed specifically for that reaction. The number of possible molecules is astronomical, and the number of pairs astronomical squared.
I hear you can treat most of molecules mechanically except for a few dozen atoms surrounding the reaction site. That limits it to, let's say, 2^^30 possible molecules, so 2^^60 reactions you need specific enzymes for. Designing any one of those 2^^60 enzymes or reaction paths is feasible. Making an index iwth 2^^60 entries is feasible, given atom-scale memory, although it isn't microscopic. Drexler suggested such an index for diamondoid struts of different sizes in "Nanotechnology:
The real number of enzymes needed is much smaller than 2^^60. To get a self-assembling molecule, assuming you feed it the right basic building blocks, you don't need a universal assembler. DNA limits itself to 4 molecules with a single type of connecting part. Proteins limit themselves to 24 molecules (I don't know if the connecting parts are standardized but I suspect they are). Ribosomes can construct ribosomes, so we already know self-assembling machines are possible.
An interesting question is, given an assembler that knows how to do some fixed set of assemblies, what can be built? How big a set is needed? The smaller the set, the less work is needed to get the correct configuration for each reaction. Perhaps we need specialized factories for some building blocks with standard connectors, then just a tape-reading assembler that can connect standard connectors? Standards simplify things.
Neil Stephenson's Diamond Age is probably a more reasonable assesment of where things will go. People will still be employed in the design of new machines, and will be able to afford better pieces of land, and more electricity (Stephenson also suggested that perhaps handmade items would become status symbols). The (unemployed?) masses will live in unparalleled comfort from a historical perspective, being well clothed, fed, etc.
"The question of whether a computer can think is no more interesting than that of whether a submarine can swim" -EWD
Apparently my design for a molecular assembler is deemed impossible for the sticky or fat finger aspect, but I need help understanding why this is impossible. I think I can circumvent that problem.
Anyways here's the design. It is simple and achievable. It is not conducive to building/replicating itself easily though.
The core is an STM microscope-like device, with many parralel tips each working on its own square millimeter (or smaller) area. Needle Tips or fingers doped with the next mollecule (properly oriented) to be inserted move along a conveyor belt where each are inserted into one of the parallel heads, the head then sticks the molecule in place, then the needle is sucked out, and sent to be refilled.
I don't really have a process for making mollecules, and placing them in the proper orientation on a needle.
The one good thing about this design, is that there probably exists a needle material than can react "properly" with any given mollecule, such that it can 1) capture it, and 2) release it. (One method of releasing could just be to jab the needle quickly forward, flinging the package into place).
There's one problem with Drexler's universal assembler theory, with little publicity, that is only partially solved by my design. There is not an infinite number of universal assemblers created instantly, and as a secondary problem, programming them to work and move around cooperatively is not easy, and increases their required size if only because of the massive cpu they need to operate with.
The Other issue only partially addressed is speed. If everything is built using 3d tetris-like merging of 10nm building blocks (mollecules), then finishing a square milimeter takes 10B blocks. A quadrillion blocks makes a cubic milimeter of something. Even at 1 billion blocks per second, it takes 11.5 days to make 1 milimeter thick object. The billion blocks per second seems outrageously high to reach, but another way to increase throughput (but increase congestion of feeder needles) is to have denser parallel heads. If each head works on a square micrometer area, then building a cubic micrometer object takes 1 million 10nm blocks. At (only) 100K blocks per second, a 1mm thick object takes 10,000 secs = 2.77 hours
There's one other big problem. Like building a house of cards on an uneven surface, mollecules won't necessarily maintain a desired orientation without simultaneously placing adjoining molecules to couterbalance them. Seems like there would be a solution to this, with all the arms in such close proximity, but it would also slow down the process.
Most people do not really understand the potential impact of mature nanotechnologies. And it's easy to see why - even Drexler's book Unbounding the Future: The Nanotechnology Revolution gives some really ridiculous examples (may be to make it simplier to understand). Here are my responses to two of particularly misleading comments in this thread.
:
:
2BorgDrone
However, if molecular assemblers ever become mainstream I'd rather design my own car and let it assemble that. If everyone is driving a Ferrari I'd rather have something different.
When molecular assemblers become mainstream, having a car would be rather pointless. First, there are unlikely to be any streets where you can impress chicks, since everyone would be able to live wherever on Earth they like. Second, there probably won?t be any roads, since you don?t need to transport goods (they can be manufactured from CO2 on the spot) and it?s easier to fly people from A to B. Third, designed cars would be as old-fashioned as horse carriages now ? smart completely transformable people-movers would be all the rage. And forth, you will be able to drive any kind of car in your personal virtual reality simulation, so you don?t need to actually design the car (just program how it should behave) and the issue of IP would be moot.
2jchoyt
Money will still have value. Someone has to create and/or design food, clothing, medicine, entertainment, etc.
Strong AIs will be able to create and or/design everything, including these things you describe. Furthermore, people will not need food, because it will be easier to just get energy from the environment without any conscious actions like eating from your side. Clothing is likely to be designed for the sake of it. Most couturiers are not in this business for money, they do it because they like it, and when all fabrics and basic production operations will become free, as well as everything they need personally, they are unlikely to charge you anything for their latest fashionable clothes. Medicine will not be used, because our bodies will be redesigned to include a smart AI-based immune system, capable of fixing most problems, except, may be, for being in the epicenter of the thermonuclear explosion. So most things you mention will not be needed and those that still will be needed, will be done by professional volunteers for free.
Future Wiki -- If you don't think about the future, you cannot have one.