Slashdot Mirror
Real Nanotechnology Getting Closer, Says Drexler
destinyland writes "Sun Microsystems has helped fund a 198-page nanotechnology roadmap — but how
close are we to real nanotechnology? A science writer asked four nano pioneers, including
K. Eric Dexler ('progress is accelerating') and Ralph Merkle ('the exponential trends continue to be exponential') Though we don't have Star Trek replicators yet, the article lists some surprising recent nano developments (artificial tissue, nanoparticle sheets, ultrathin diamond nanorods).
And the roadmap's scientists are envisioning targeted cancer therapies, super-efficient solar cells, high-density computer memory chips and even responsive 'smart' materials."
(... And circular reasoning works because circular reasoning works.)
We've seen this with so many things, including solar cells - Constant assurance that they are getting cheaper easier to make, more efficient, etc; people ranting about how it is finally feasible and will be seen in mass quantities soon... yet we still don't.
IMHO, it's vaporware until the common Joe is purchasing and holding it in their hands.
That _DOESNT_ mean I don't acknowledge the advances, just that I don't get my hopes up.
Evolution - Est. 4500000000 B.C. Don't piss in the gene pool.
Nanotechnology is like cybernetics: Any application that no longer feels exotic no longer falls under the common use of the term. This is why people with cardiac pacemakers or cochlear implants are generally not considered to be cyborgs, and microchips are not considered to be nanotech.
Nope, microtechnology, hence the micro. ;)
Spell check says it isn't a word, but what do spell checkers know? Bah!
Just around the corner! No, really!
I think the problem with nanotechnology is that people make the assumption that it has to be a miniature self aware all purpose robot.
Where really we already have nanotechnology being used in the real world today.
I think we should call it "nanorobotics" instead of nanotechnology to make it more clear to people.
That said, they do have nanobots out there in the research phase which are very promising for chemical delivery for tumors at this point so we are going to see something in 5 to 10 years.
"I am the king of the Romans, and am superior to rules of grammar!"
-Sigismund, Holy Roman Emperor (1368-1437)
Nanotechnology is the artificial intelligence of today. Back in the 1970s, we were all promised that real artificial intelligence would someday exist and we'd all have all-in-one robot maids running around doing our dishes and vacuuming our floors and answering the phone, the door, etc. Lots of things, like natural language processing get called AI, but real AI? A real, self-aware robot with a mind? Forget it. A computer is a billion switches. Even if we turned it into a googolplex switches, it's still nothing more than a googolplex switches.
Nanotech is doing the same thing: we have lots of things we're calling 'nanotech' that aren't what we were promised.
My blog
On the other hand, the goo growing in your bathroom sink, notably without the benefit of a 10+billion dollar fab and cleanroom conditions, is pumping out structures that small more or less continually. Top down is, indeed, beating bottom up in the limited realm of what we know how to do; but bottom up has been kicking ass everywhere else since not so long after the planet cooled a bit.
Bottom-up assembly is certainly a long-term basic research type project(unless you count the sort of temperature and composition control tricks that metallurgists have been using to produce desired crystal structures for centuries, among other things); but it is ultimately a very desirable skill to pick up. As long as we have to fab them top down, nanotech materials are going to be confined to niche applications(Sure, semiconductors are common; but compared to concrete and steel?)
A computer is a billion switches. Even if we turned it into a googolplex switches, it's still nothing more than a googolplex switches.
Our brains are nothing but billions upon billions of neurons, synapses, etc. forming complex interconnections.. yes, any first generation intelligent AI would have to be created by humans, but if we exactly modelled a human brain in software and trained it like any other child (it would probably need the aid of a prosthetic or virtual body to be able to learn), what would really make the resulting AI different from ourselves if it reacted as we do? I know it's a big if, and that there probably isn't much point in creating an AI that has human flaws - but there is nothing in life to indicate that we are anything other than purely physical constructs. Otherwise, why bother with having bodies in the first place - unless perhaps our bodies are as to the soul as cars and aeroplanes are to humans?
which is totally what she said
True, we are already building electronic components at nanometer scales. But when people talk about nanotechnology, they are usually thinking of mechanical devices built from nanometer scale components, or larger structures which exhibit new properties based on manufactured, nanometer scale features.
The industry for these applications has hardly even begun.
The achievements of the lithography industry are absolutely stunning. And if you want to call them a branch of nanotechnology, that's fine too.
But they have not achieved the holy grail of nanotechnology, and the tricks of lithography never will. The holy grail is atomic-level precision; not just in restricted circumstances (e.g. single atomic layers under some constraints), but in the general case. As in, you draw in some CAD program an arbitrary (within physical law) device wherein each atom is specified... and then you get it built. Lithography cannot do this. Synthetic chemistry can do this for a subset of chemical compounds, but can't tackle the general case and certainly can't currently make arbitrary nano-devices with atomic-level precision. You're right that bottom-up approaches like self-assembly also can't currently do this (they are more of a way to assembly precise sub-units into larger assemblies).
This final "true" nanotechnology (Drexler now calls it "molecular nanotechnology" to differentiate it) won't be easy, and may very well require a delicate combination of everything we've learned from of top-down techniques (e.g. lithography) and bottom-up techniques (e.g. synthetic chemistry, self-assembly). Or maybe it require radically new thinking. The point is we don't yet know, so to say that "top down beat bottom up years ago" really misses the point: molecular nanotechnology has not yet been acheived.
In the meantime, our current tricks all have their uses (lithography is great for, e.g. making microchips... whereas self-assembly is great for making, e.g. coatings for pharmaceuticals and fuel-cell membranes).
Science is always changing, always evolving, because we are always learning (some of us anyway). If you expect science to be exactly the same in 10 years, you have a fundamental misconception of how science works. In fact, science not only does change, it must, otherwise it becomes more dogmatic garbage the world has too much of anyway. Don't mistake your lack of comprehension for the fallibility of science.
"I disapprove of what you say, but I will defend to the death your right to say it." - Evelyn Beatrice Hall, re Voltaire
Well, you're clearly drawing parallels with claims with AI. I think the difference between Molecular Nanotechnology and Artificial Intelligence is that 40 years ago, people didn't really have an understanding of artificial intelligence or how to achieve it. Or perhaps it's more fair to say that what people really meant by artificial intelligence is artificial consciousness and that the underlying mechanism for consciousness is still not understood. However I think it's getting clearer that it's an emergent property of certain types of extremely complex neural networks. If the electrical/nanotechnological engineers can keep Moore's law going, then we should be able to simulate that level of complexity in another 30-40 years, which would mean that artificial consciousness should be technically feasible by that time, even if it's likely to take some more time to actually achieve it. That's something that was far from true 40 years ago. In addition, fMRI's are now helping us get significant macro-structural insights into the functioning of the human brain. so that we can work both top down and bottom-up to a solution.
Now the thing about molecular nanotechnology is that, unlike with artificial consciousness 40 years ago, we do actually have some idea of the types of mechanisms that we want to achieve. While we do need to develop a lot of tools, and there are some questions about the feasibility of certain activities, there appears to be far fewer major unknowns in the roadmap than there were in AI/AC 40 years ago.
What I'd really like to see is PV efficiency in the 20-30% range. Then not only would it be viable for home use, but it would mean the solar panel roof of the new Prius could drive the whole car, instead of just the AC.
Er, well, no, not really.
Cars use tens of kilowatt hours to get around over the course of a few days. Even with 100% efficiency, the upper surface area of a (normal) (street legal) (meets all US federal safety standards) car isn't big enough to soak up that much sunlight. Now what high(er) efficiency does allow is for the roof of the typical suburban home to power both the entire house underneath it (including overnight, with batteries) plus a car. That's something worth wishing for.
But while we're wishing, you're aiming too low. Let's wish for the 40% to 70% efficient cells we've heard about in labs but that never saw the light of day afterwards. Whatever happened to multi-bandgap or quantum dot solar cells, anyway?