The Speculist

Very exciting developments. What do you make of this Wohler individual who commented (in the New Scientist article) that this apporoach won't catch on, because all you're doing is making parts, which could be made more quickly and less expensivley by standard means?

Bower seems to think that people can "clip the parts together." That might be a show-stopper for me. I assembled our new dining room table last week, but I think that's about as far as I can go with that kind of thing.

Maybe a bicycle. I could probably do a bicycle. Could a fab lab output a mountain bike?

But I draw the line at a digital camera. No way.

To be capable of producing items of the camera level of precision is going to need extremely precise, sophiticated but very flexible robotic manipulators, moulds, jigs etc.
Very tricky.

Highly engineered jigs and moulds are one of the key factors in manufacture, and a major part of production costs.
You'll also likely need furnaces and ovens, sprayers, weavers, electro-deposition gear etc, etc. The sort of gear able to work with hard metals, ceramics, glass, various interisting organics.

Spray deposition of low-temperature resins is not going to get you far.

If it can be done, this will make all the difference between producing non-complex goods, prototypes, models or at best, in Bruce Sterling terms, products "shoddier than the cheapest postwar products of Japan" and one able to produce high-grade consumer goods; including reproducing itself.

And there is no reason why a fab could not be so capable in principle. It's all a matter of size, complexity and initial cost. And of course, if self-reproduction can be achieved, or even almost achieved, subsequent costs will rapidly reduce to inputs costs.

The first of these self-replicating fab labs will be slow, ugly, and very limited in what they can fabricate.

But from day 1 on there will be steady (and very cheap) improvements. People will not hesitate to upgrade because the upgrades will be mostly free - only the cost of energy and the reusable raw materials will be involved.

I agree with Wohler to some extent. The first products from these fab labs will be more expensive than standard products and subpar. Manufacturors will laugh...nervously.

But Wohler has forgotten the power of the nerd. Nerds will get a charge out of tinkering with these things in their spare time.

The first generation of fabbed products will require assembly - even then they won't be much more than naked circuit boards. Probably a popular first generation fab product will be A.M. radios that will remind people of those kits you buy at Radio Shack. No danger to Sony there!

On down the road whole products will be produced. At first we'll have the Wright Flyer, later comes the 747.

I can see only three major flaws with these schemes. First, as the self replicating factories themselves begin to propagate they themselves will decrease in value relative to their raw materials, and people will soon enough realize that the raw materials are where the money is.

Second, once you have the factory, unless you want to invent everything they make, you will need intellectual property to run on them, and again, people will realize that this is where the money is.

Finally, one can assume that these devices will be regulated in the near term after their development, since they do consume resources and presumably do produce waste, and if one is a conspiracy theorist, they might be regulated to preserve existing economies. One could certainly also make the case for regulating them before they "get loose" in whatever fashion is feasable and undergo runaway reproduction.

Still, it is interesting to consider how this technological evolution will go.

-Jim Strickland

Jim Strickland:

Intellectual property issues are going to arise. But what about copying old but still serviceable designs? If the quality is good enough, an older design might be acceptable.

Thinking about materials inputs for (sophisticated) fabs: this would be an ideal fit with some sort of limited "wet-nano" or "bio-nano" system, that was still way short of the nanobot or replicator or "dry" molecular manufacture stages.

Something capable of extracting a lot of pure materials from wastes, for instance.

If able to produce organic compounds of wide variety, plus semi-structured, but not complexly "machined" bio-materials - such as homologues of wood, ivory, natural fibres, leather, nacre, natural etc. - it would be an ideal "add-on" for a fab working at the micro to macro scale.
Especially if the fab included spinning, dyeing, weaving, cutting and stitching capabilities = instant clothes.

Another possibility:
A chemicals/materials production capability like that above - or even one not up to producing structured material - would also capable of producing virtually any organic compound. Hook that up to a good analysis system and molecular design system. Watch the existing pharama business model go belly up.
New drug designs can be sold as software; but how to get a reasonable return if all a pirate need do is buy a pill, drop it in an analysis system, hit copy and sit back?

Oh, just thought of this: wave goobye to any control of "recreational" drugs, when this sort of unit could churn out THC or LSD or MDMA or diamorphine or cocaine etc. How's that for a killer app for sales to local youth?

"Johnny's in the basement, mixin' up the medicine..." Indeed.

And one more really scary prospect comes to mind: it would also be a potential portable nerve gas and biotoxin factory.

John F wrote:

"You'll also likely need furnaces and ovens, sprayers, weavers, electro-deposition gear etc, etc. The sort of gear able to work with hard metals, ceramics, glass, various interisting organics."

On his Listserv, Andrian Bower is having a conversation about exactly this:

Vik Olliver wrote:

Has any work been done on the FDM printing of moulds for ceramic or plaster components? I wondered if this might provide a way of creating the heat-resistant items required in the RepRap? Heater bodies, ready to be wrapped in nichrome; nozzles perhaps, made as they used to make clay pipes? If necessary, the RepRap would even be capable of printing the moulds for its own oven!

Andrian Bower wrote:

we've already thought of that. The primary idea was to have RepRap make a mould in thermoplastic and use it to cast a thermoset part that would go up to high temperature. But we could also do plaster or ceramic slip.

John F wrote:

if self-reproduction can be achieved, or even almost achieved, subsequent costs will rapidly reduce to inputs costs.

I think this is a key point that we've not discussed here yet. Self-replication doesn't have to be perfect to have a very important exponential effect. If a RepRap could fabricate 90% of itself, the cost of new units would drop close to the cost of the unfabricatable 10%. Imagine if cars could reproduce except for wheels and brake pads. Pretty soon the cost of cars would be mostly the cost of wheels, brake pads, and the raw materials.

Jim Strickland wrote:

as the self replicating factories themselves begin to propagate they themselves will decrease in value relative to their raw materials, and people will soon enough realize that the raw materials are where the money is.

I agree in part, disagree in part. Replicating machines will decrease in value relative to their raw materials, but being a commodity broker is not the way to get rich. Information is where the value is today and increasingly will be in the future. I think the future for investment dollars is in service industries and in information businesses.

Information businesses consist of companies you may not currently think of as information brokers. Drug companies are in the information business. A world away the film and music industries are in the information business too.

once you have the factory, unless you want to invent everything they make, you will need intellectual property to run on them, and again, people will realize that this is where the money is.

Right, information is the key. I think there will be significant file sharing on some of this stuff too. Various enthusiasts will share CAD plans just for prestige. We see this sort of thing with freeware today. There may also be some shareware business model that can be developed for these CAD plans.

One could certainly also make the case for regulating them before they "get loose" in whatever fashion is feasable and undergo runaway reproduction.

Like Napster, I think this will be one of those things that hit before the world understands what's coming. I'll bet that not a congressman in Washington is even aware of this as a possibility. If any are aware, they probably either discount it as something that's years away or a crackpot theory. The public will have these devices before Congress can get mobilized to prohibit them.

John F wrote:

A chemicals/materials production capability like that above - or even one not up to producing structured material - would also capable of producing virtually any organic compound. Hook that up to a good analysis system and molecular design system. Watch the existing pharama business model go belly up.

New drug designs can be sold as software; but how to get a reasonable return if all a pirate need do is buy a pill, drop it in an analysis system, hit copy and sit back?

Oh, just thought of this: wave goobye to any control of "recreational" drugs...

And one more really scary prospect comes to mind: it would also be a potential portable nerve gas and biotoxin factory.

What you are considering is beyond the Fab Lab concept. This would be a nanotech-enabled replicator. Such devices are years away. Once such devices become possible they will have to be closely regulated for the reasons you gave. That's the sort of device that would allow a single nut to unleash a world-wide plague.

Stephen Gordon:

I grant my "universal organic chemistry set" is way beyond the Fab or RepRap, so I'm a bit OT. I wasn't thinking in terms of replicating technology (at least initially) at all, more an abstract "black box" module for chemicals as an "add on" to an otherwise micro-to-macro scale Fab or RepRap system.
Interesting how radical things could get even if replicators, nanobots, even any "dry" nanotech at all, are excluded by fiat from a scenario. Because we know "wet" nanotech - i.e. biology - can produce these organic chemicals and materials.

Add such capability at reasonable cost to a Fab setup, and you get a really useful system. And if I can't have diamondoid I like the idea of a proliferation of stuff made out of synthetic wood, ivory and mother-of-pearl as well as cheap plastics. :)

Jim Strickland already noted my main original observation.

The point about IP is interesting.  What happens if all the ESR's of the world start hacking interesting things, and release their designs under the GPL?  People who really need money could use something like the "street performer protocol"; enough people kick in a buck each, and the design gets released to the world.  I can see how that might cut the cost of finished devices to the raw materials plus whatever energy and fabricator time costs.

Processing is going to be an issue.  Heat-treated alloys and hard ceramics are going to be tough to do even if you can supply the powdered raw materials; reproducing a laser to sinter them is going to be the Achilles' heel of any such scheme.  The easy things to do are going to be chemically cured polymers of some kind.  If you can make actuators out of polymers, so much the better.  Better yet if the polymers are designed such that off-the-shelf reagents, such as enzymes, can break the bonds and recycle the materials to monomers.

After a few generations of improvement to make more and more of the fabricator reproducible by the fabricator, wouldn't this start resembling living organisms?  Organisms which can reproduce and "eat" their "dead"?  The parallels become almost eerie.