Automatic fabrication ("Autofab") is the ability of a machine to manufacture a wide variety of product designs, including, most importantly, a copy of itself. All that is required is electrical power, relatively low-cost raw materials and components, and a design to produce in digital format.

Automatic fabricators are often shortened to "autofab". While the "auto" prefix does indeed mean "automatic", indicating that the device can operate largely independently of human oversight, it also indicates the recursive, self-referential nature of the device - it can manufacture a copy of itself. "Recursive fabricat/or/ion" might also be a good term to use.

Self-assembly of the product is not required; it is acceptable - in fact, it's preferable, for safety reasons - that a human operator be required to put together the end product into it's final form, in order to maintain a certain degree of human control of the technology.

The most dramatic thing about autofabs is the tremendous spped with which they can saturate the market. If the first autofab built can turn out a copy of itself in a matter of hours, then we have a class of device that is capable of doubling it's population in that same period, limited only by the availability of raw materials and components. You can start with one the first day; on the second you have two. Then both of those can reproduce. Each subsequent day could see, 4, 8, 16, 32, 64, 128, etc. Presuming that the requirements were available, every human being on the planet could have one within a month.

The first autofabs will take the world mostly by surprise. Existing sources of components and raw materials for them which are built on pre-autofab levels of demand will sell out their inventory quickly, and scramble to feed an explosion in demand. At the same time, the first autofabs will be quite primitive, like the University of Bath's RepRap project. They'll be able to make simple, solid-state products made solely of plastic, and that's it (RepRap actually intends to include primitive printed circuit boards, as well. We'll see). The cost of the first autofabs and their raw materials and components will still put them out of the reach of the average household in developing nations. Nevertheless, progress will be made in improving the technology in terms of price (cost of raw materials & components), and capability (types of products that can be built - electronics, base material, mechanical pre-assembly).

The thing to note about this shock factor is that the first, primitive autofabs are the ones that can sucessfully multiply exponentially are the ones that will take the planet by surprise. After that, debugging and upgrades to machines will spread along the same lines as before, with improved autofabs replacing old ones as they go. Unlike the initial deployment, however, improvements will all be incremental in nature. The initial autofab explosion will be a revolution, followed by the steady and eagerly anticipated evolution of the technology. The initial deployment will be the "fool me once" moment. Once people see what the basic idea is, all the following implications will be pretty obvious, and will not be nearly so shocking.

The idea for autofabs is mainly inspired by nanotechnology. Computer Scientist John von Neumann envisioned "universal constructors". Eric Drexler later envisioned these in more detail in his book Engines of Creation, drawing on the ideas of Physicist Richard Feynman. Chris Phoenix of the Center for Responsible Nanotechnology recently published a rudimentary design for a nanoscale "desktop nanofactory". (EXPAND)

However, it's apparent that the critical criteria for the dramatic impact autofabs can have on manufacturing have no bearing on the scale at which the device works. There is no need to invent molecular-scale manufacturing processes in order to produce a machine that can produce copies of itself and other products. Gershenfeld Fab (EXPAND)

Furthermore, similar devices don't have to be able to make themselves right off the bat in order to be tremendously useful. Today's existing rapid-prototyping machines do great work for engineers designing new products. RP machines are still hovering at the US$50,000 mark, however - mostly because they lack the critical criteria for an autofab: the ability to manufacture copies of themselves.

Autofabs will take the manufacturing industry and put them in a position similar to the RIAA and MPAA and their client companies today.

The Patent Commons project is analogous to Creative Commons for the age of autofab, and applying to patents. IBM, Novell, Sun, Red Hat, OSDL, and other large corporate interests who have recognized the business value of the intellectual commons are collaborating to create it.

(aside: the intellectual commons provides a way for industry to "collude" with competitors in a way that benefits consumers. This has implications for my feeling that competition is primitve and wasteful, whether its market driven or not. expand this somewhere else)

The Free Software Foundation is currently in the process of revising the GNU General Public License. One of the primary concerns is how to protect GPL'ed software against patent claims.

Getting a patent is expensive, in comparison to a copyright, which is automatic. However, there is a concept in patent law called prior_art, which enshrines the requirement that patented inventions be novel - that is, they have to do something no one else has done before. In practice, A patent examiner is not expected to know everything, and it is permissible (and common) for a patent to be granted despite existing prior art. The onus is then upon others to prove in civil lawsuits that the prior art existed prior to the patent-holders implementation, and is applicable the patent in question.

In principle, this provides a way to protect unpatented inventions from being patented by others. Unfortunately, it still requires copyleft inventors, who may be operating without the benefit of generous legal defense funding, to defend themselves in a lawsuit agsint well-heeled commercial interests. This problem may be being addressed by the Patent Commons project, or even the forthcoming GPL v3.

One of the most difficult parts of patent litigation today is the ability to defend suits on the basis of prior art. Finding relevant inventions, and even prior patents, that implement relevant inventions is hard enough for a patent attorney, never mind a patent examiner.

This is a problem technology can help solve, however. Once product designs are standardized in some digital file format, they can simply be published on the web. Copyleft and open source designs will be passed around just like software today. A search engine could easily find these files and index their contents, much like web pages and the myriad other document formats available online today. Once a method for searching product designs is available, the ability to defend inventions on grounds of prior will be greatly improved, making frivolous lawsuits against unpatented inventions much less likely to succeed, and thus, to be launched. This will be a great advantage for the commons.

Proprietary designs, on the other hand, couldn't be indexed by the search engines as easily, making them difficult to find. This opaqueness conceals the threat of the submarine patent, which no one knows about until the patent holder, having no intent to commercialize the patent himself, contacts you to demand royalty fees or stop distributing your design.

The reluctance of proprietary inventors to publish their fab designs free of charge may very well cripple them in the patent marketplace, as well. For reasons similar to those in the field of copyright, creators who tend not to be stingy with releasing their works free as in beer actually sell more copies, so long as they're not among the most popular of their category.

Furthermore, grass-roots innovators newly enabled by the advent of autofabs will explode along with the machines themselves. There will simply be far more people who want to use an autofab to make themselves something that THEY can use, and, much like software, they won't mind sharing and improving on the design with others, free of charge. The probability that autofabs will eventually saturate human society much more fully than relatively expensive computers have in our age will unleash a flash-flood of copyleft and open source inventors such that even modern CLOS has not realized. (the fact that autofab is likely to drive the cost of computers down even further means that a similar explosion will also occur in software, as well).

Speed

manufacturing unemployment

• auto - taxis vs buses vs private cars

if cars get cheap (even as gas goes up - see solar panels), more people will jam the streets. perhaps taxis will become popular. combinations of cheap vehicles and autopilot systems may seriously slash costs.

• construction - contour crafting at USC. automatic construction systems.

service unemployment in fields relating to manufacturing:

• delivery
• energy - solar panels can make electrical utility companies obsolete.

Political lobbying; which leads to...

Manufacturing interests lobbying the NSA/FBI/legislature. Autofabs are not software. products have real danger potential. As a result, I suspect they will succeed at convincing governments to outlaw private fabs in order to keep them locked up in local fab facilities they own. This way, they can be made accountable to government (and possibly the public) for what they produce, and who they produce it for.

They can also:

• prevent public access to unsafe products (both unsafe products, and weaponry).

hmm. the practice of disclaiming liability in software designs is very customary, and uncontroversial. will courts allow this for product designs? This is a problem for Commons designs.

• enforce the collection of patent and copyright royalites.

When you order a CD or DVD, a player could be embedded directly in the thing for trivial cost. Of course, the analog hole still exists.

• refuse to produce competing products, including any design released under a copyleft or open source license. Or charge exorbitant fees for producing them, making them effectively non-free. Either legislation will need to set stautory rates for production of commons products, or the barriers of entry into the fabrication market will need to be regulated. eg, fab license applicants will need infrastructure fro reporting and sruveillance of their oeprations, and, of course, they'll need access to an otherwise illegal fab for a reasonable price ("reasonable" being nothing, unless you're talking to the entrenched fabrication cartel. How do you evaluate the damage a competitor can do to your monopoly business?).

Trade prevents war. In All of human history, it seems we don't really need a good reason to go to war; we need a good reason to stay out of war. Trade seems to be a good reason. In the 1980's, The manufacturing sector of the US economy was steadily "outsourced" to China - although the term never really surfaced until the 1990's, when we began outsourcing services like computer programming.

It was always curious to me why we thought the Cold War was over when the USSR collapsed around 1990. The biggest communist country of them all was still around: China. They're still communist today, although economic reforms are slowly occurring. Despite the remaining human rights issues and one-party dictatorship, we don't pick on China much. We don't seem to mind political dicatorships, so long as they allow us to take advantage of their social control to extract cheap labor.

The reason is pretty obvious: look under half of the stuff in any American home and you'll find "made in China". If we got into a fight, they'd lose a massive chunk of their economy, and we'd lose an equivalent chunk of access to consumer goods. Trade dependencies encourage peace.

Add the recursive nanofactory into this equation, and that state of affairs collapses. If goods can be made just as cheaply with fabs over here as in China, then Americans can keep all the money they send to China for theirs. Presumably, China would also benefit from better, cheaper products, but then they wouldn't have this revenue stream they've gotten used to. The collapse of such a trade dependency is one less good reason to avoid war, if not a reason in itself.

A synonym for automatic fabrication, emphasizing the self-duplicating nature of the device.

Existing industrial appliances that can automatically forge products from digital designs (short of copies of themselves) are know as Rapid_prototyping machines.

John_Von_Neumann's Universal_Constructor

K._Eric_Drexler's Molecular_assembler

Chris Phoenix, Center for Responsible Nanotechnology Constructors are not free-floating, self-reproducing units, but nanoscale construction platforms that are anchored in a grid to form a macro-scale device. This has great safety benefits. Continuing the scaling up, and security benefits continue to accrue!

Neil_Gershenfeld, author of Fab: The Coming Revolution on Your Desktop - From Personal Computers to Personal Fabrication (ISBN 0465027458)

In his book, Gershenfeld reports on the results of taking a package of hi-tech industrial fabrication machines to various international venues and teaching locals to use them. A fab lab is only about US$20,000 worth of equipment that can make "just about anything", when you provide some basic training to the lab's operators. Not surprisingly, deployments in Africa and rural India received a trememndously positive response, whereas in Massachusetts, people were stymied by concerns over patent infringment.

This term is also coined by Gershenfeld, a "personal fabricator", or "PF", is similar in usage to "personal computer", or "PC". See fab lab

Star Trek

Neil Stephenson, The Diamond Age Stephenson's "MC's" are hooked to what amounts to a universal utility line to every home which provides all the device's raw material, design and control data from a centralized source. "The Feed" serves as a form of social control as well, since it can be turned on and off much like modern utilities like electricity or water. It is unclear whether or not MC's can create copies of themsleves, but it is clear that they cannot operate without access to a Feed line. The plot of The Diamond Age revolves around the development of "The Seed", which is a new form of MC which can operate without the Feed, and thus represents a dire threat to established social institutions which formed around the power of the Feed.

http://www.RepRap.org/ Not only is the RepRap project (short for Replicating Rapid-prototyper) of the UK's University of Bath focusing on creating the first simple autofab, but they're aiming at relasing it's design under the GNU General Public License as a gift to the world. All the reprap is designed to be able to build is plastics and limited, large-scale electronic circuits (something you can plug chips into later).