Slashdot Mirror
Envisioning the Desktop Fabricator
mkl writes "Yesterday I fantasized about a generator of matter. Not a laser plotter for carving 3d objects, but a device that will assemble any given object from its base, out of atoms. I was thinking about a device that can find its place under the roofs of all the people working on PCs all over the world. So I fantasize about it at work and what do I see in the Wired News newsletter? 'Any product, any shape, any size -- manufactured on your desktop! The future is the fabricator.' Heh."
First post to mention The Diamond Age : Or, a Young Lady's Illustrated Primer (Bantam Spectra Book) by NEAL STEPHENSON. All about nano-tech and fabricators and stuff.
Test 1 2 3 4
I saw a prototyping machine at a recent trade show, that could lay down ABS plastic. For a six cubic inch toy wheel, it was an overnight job. It wasn't neceessarily a desktop unit, it was still considerably larger in footprint than an HP LaserJet 4, and is floor standing, I think.
It also costed $25,000.
The machine type described are good for prototyping and custom parts, but there are usually better mass production methods. Laying down atom-by-atom will be slow for a loooong time and at best be of most consequence to nanomachines for that time.
"Heh," indeed. Actually, I think Freeman Dyson was speaking about nanoassemblers in the 50s... Drexler was the first to assemble a full theoretical defense of the idea. (Though some still think it's fantasy regardless)
I can't imagine how it feels to publicly embarass yourself by posting a 'new idea' that is 60 years old.
PS. The entire text of ENGINES OF CREATION can be read at foresight.org!
See this link on Slashdot trolling phenomena.
Be prepared for even more "eeeeww"s!
Go read Drexler's "Engines of Creation" for the classic "nanoassembler" hype. The idea of pushing atoms together is neat, but it's hard to do. Free-floating nanoassemblers are still a fantasy. I expect to see nanoassemblers, but they'll probably be more like scanning tunneling microscopes made on an IC substrate and used to read and write DNA. Making big hunks of solid materials that way is too slow. Look at how long it takes to make a tree, or a coral reef, or a pearl. (Admittedly biology is power-limited. In a manufacturing environment, you can run external power into the nanomachines and remove that limitation. But that won't work for the free-floating nanomachine concept.)
If you have a good milling machine, you can make almost any solid shape you want. I know four people with milling machines at home. Two of them have good computer-controlled mills with all the necessary software. Yet they don't actually make all that much. One of them is building a steam engine, and he's been at that for years.
Then there are stereolithography machines. The newer ones work fine. You can now make things out of ABS and nylon, which are tough enough to be useful. This is a big improvement over the early models, which made only soft wax models nice to look at but useless.
It's a very slow way to make stuff. In the real world, almost all consumer products, with the notable exception of wood and fabric products, are made by some kind of cheap moulding process. There are dozens of such processes, from die casting to injection moulding to progressive stamping, but they all involve forcing material into a mould. This is an incredibly cheap process in quantity, and is why manufactured goods are so cheap. Very few consumer items are made by machining down a solid hunk of material.
Even ICs aren't made by direct writing. It's quite possible to make ICs with direct-writing electron beam machines. This eliminates the need for masks, and every IC can be different. Works fine. Useful ICs have been prototyped that way for years. Too slow to be commercially feasible.
> fantasized about a generator of matter,
> one that was able to generate [image of beauty]
> right in front of me complete with a handbag
> full of a strange gritty substance...
this fantasizing of procuring women from stone has persisted
thousands of years in the greek legend of 'pygmalion galatea'
Pygmalion and Galatea in Greek Mythology
Pygmalion saw so much to blame in women that he came at last to abhor the sex, and resolved to live unmarried. He was a sculptor, and had made with wonderful skill a statue of ivory, so beautiful that no living woman came anywhere near it. It was indeed the perfect semblance of a maiden that seemed to be alive, and only prevented from moving by modesty. His art was so perfect that it concealed itself and its product looked like the workmanship of nature. Pygmalion admired his own work, and at last fell in love with the counterfeit creation. Oftentimes he laid his hand upon it as if to assure himself whether it were living or not, and could not even then believe that it was only ivory. He caressed it, and gave it presents such as young girls love, - bright shells and polished stones, little birds and flowers of various hues, beads and amber. He put rainment on its limbs, and jewels on its fingers, and a necklace about its neck. To the ears he hung earrings and strings of pearls upon the breast. Her dress became her, and she looked not less charming than when unattired. He laid her on a couch spread with cloths of Tyrian dye, and called her his wife, and put her head upon a pillow of the softest feathers, as if she could enjoy their softness.
The festival of Aphrodite was at hand - a festival celebrated with great pomp at Cyprus. Victims were offered, the altars smoked, and the odor of incense filled the air. When Pygmalion had performed his part in the solemnities, he stood before the altar and timidly said, "Ye gods, who can do all things, give me, I pray you, for my wife" - he dared not say "my ivory virgin," but said instead - "one like my ivory virgin."...
--
sometimes you don't get what you want, but you get what you need.
go for the real thing, reciprocal exchange is so much better...
best regards,
j
Then again, this all sounds way too good to be true.
Probably because it is too good to be true. They're just so many flawed assumptions behind the idea of the desktop replicator that puts it on the same level as warp drive, a literary device that is good for those moments when you need the hero to create an object on the fly, but really bad when talking about future economics.
Such a device would make physical goods value-less. The only things of value any longer would be services and artistic creations.
Whoa! Hold on here just a minute! There are a large number of costs involved in producing a physical good. While your machine might be able to eliminate the labor costs, you still have to deal with the costs of raw materials, the costs of time, the costs of design, and the cost of energy.
Just in terms of energy costs, it is quite possible that it is cheaper to create a toaster using traditional metallurgy than to try to assemble a toaster atom by atom. The big problem with assembling metal objects atom by atom is how you deal with oxygen. When you are working with ingots, sheets, and wires, the fact that contact with oxygen is limited to the surface area helps quite a bit. The smaller the particles that you try to push around, the worse the oxygen problem becomes. Once oxidized, separating oxygen away from what you want to work with is a hugely expensive process in terms of energy. The reason why living organisms prefer to work with carbon is because carbon is one of the few elements for which the energy trade-offs are reasonable for both oxidation and reduction.
The second problem is time. It is not necessarily the case that a fabricator will be able to produce a widget atom by atom in a time frame that is competitive with an assembly line. This also adds value to products.
I think that there is also a basic misunderstanding of economics expressed here on this topic. Economics is not just about our current fuzzy version of capitalism. Economics applies to just about any situation in which you have local surpluses and local scarcities. Even with a desktop fabricator, there are still surpluses and scarcities that do not spell an end to economics.
A revolution of affordable open source desktop manufacturing is on the way. There is already an alternative approach to rapid prototyping and manufacturing using inkjet technology.
Well before we are building things atom by atom, desktop manufacturing will be producing some stunning and swift changes in what we can produce for ourselves. The humble Inkjet, in a jag of hardware hacking is already spitting out solar cells, batteries, complete working gadgets, human tissue and computer circuitry,. A computer printing computer circuits simply from software instructions. That's only a stone's throw away from self replication.
There's more.
Researchers Hod Lipson and Jordan B. Pollack at Brandeis University have coupled inkjet technology and software to autonomously design and fabricate robots without human intervention.
or
Google Search
The software simulates a variety of rudimentary virtual robots. In an accelerated Darwinian contest of survival over hundreds of generations, the most successful robotic designs are then physically prototyped. Robots autonomously designing, testing and manufacturing robots.
The implications of open source desktop manufacturing are perhaps more in the questions inspired than in what is produced.
What will be the effect of open source hardware? What happens when a desktop peripheral as economical as your printer manufactures custom computer circuitry, solar cells and batteries as cheap as wallpaper? Or when distributors ship a product as software, with the end user supplying the raw material. No distribution costs and instant delivery of a physical item. Or when autonomous robots fitted with accelerating computational intelligence design and manufacture their own next generation?
And now another working approach to desktop manufacturing pops up. I say 3 years will see the revolution spill out of the manufacturing sector onto our desktops.
Thoughts on the Emergence of Computing Intelligence
In 1959 Feynman described what it would take to build arbitrarily anything at the atomic level: http://www.its.caltech.edu/~feynman/plenty.html