Stephen Gordon has some interesting predictions about the future of manufacturing.

That article was thought-provoking.

But do we really need more stuff? No, really, think about it.

All the time? A printer at home that spews more of that plastic junk my home is already full of. Maybe. Maybe not.

If you think about how you spend your day, you just don't have enough time to use all kinds of different gadgets all the time. Be it the ipod, digicam or cellphone. It's just very much an image that has been advertised to us: we need this and that - but in reality it won't change anything in our lives if we have some gadget or not. We're already near the saturation point.

But yeah, I've heard cell phones already cost only a few bucks to manufacture currently, so the most expensive thing in it might actually be the software nowadays. This puts the original question in a different light - we're already almost at the point where hardware is ubiquitous, but it's being held back by the manufacturers - for a good reason. If phone software would be open source, maybe all the companies would lose almost all of their revenue and might cease to exist, and you would buy no-name-brand bulk cellphones for 5 bucks and load the sw yourself - a variation of the described situation requiring not much new technology. (Except the chinese would still make them.)

I also claim we use much more of our time in services than with accessories. And they can't be "printed". A service, like being in a warmed public building or going to a restaurant or riding a bike on a public road.

(Also, you still need some separate parts like at least the lenses for the digital camera or magnets for speakers if you have a reasonable home fabrication machine which does plastics & semiconductors. So you have to buy separate stuff anyway.)

But a good link anyway.

Yeah, just like inkjet printers have put print shops out of business. That is to say, they haven't.

In fact, the print shops are the ones running the bigger, more complex inkjets, because suprise, they are big and complex and expensive, and thus lend themselves to being operated by experts who can take the time to master their complexity, and will run enough business through them to pay off the investment.

And the inkjet printers haven't killed off the big, old, milllion dollar offset presses, either. Mainly because an offset press beats inkjet on price, time and quality when you're dealing in volume.

I see a direct correlation between the print industry and the manufacturing industry here.

3D printers aren't going to supplant conventional manufacturing, because an injection moulding system can crank out 100 units in the time it takes a 3D printer to assemble one. And time is money. OTOH, there is a good chance that the injection mould itself will be made on a metal-sintering-type 3D printer. There will be some exotic luxury goods that could only be made on a 3D process. And you'll probably find a 3D Printing kiosk in the mall next to the guy who prints stuff on mugs. To say nothing of the rapid prototyping stuff which is already done on 3D printers right now.

btw, your content scanner won't let me say "s p e c i a l i s t"

> But do we really need more stuff? No, really, think about it.

One reason that we accumulate is that it is cheaper to keep and reuse than it is to acquire on each demand and dispose of between.

Things will get very interesting when the home manufacturing devices become capable of making things out of strong metals.

Note that the 3-d printer approach isn't the only way to go. The key part is "computer controlled with programs that can be run by the unwashed", not the fabrication mechanism.

First off the bat will be prescription drugs. Imagine Merck encrypting their molecules to make them harder to copy.

Just hope that no one buys the patent and locks it away for the rest of our lives.

I am imagining a fab not at home but the office. I need part X for Giant Assembly Machine Y. Instead of having GAM's service tech haul the part from Chicago, and having to wait for next day service, I walk over the fab and have it _make_ the part and we're back in service by the end of the shift.

I would not expect this to be free, but it _should_ cost less than having to haul and install, and less than having to keep a thousand parts ready to hand 'just in case'. The amount of money saved is not chicken feed.

I'm surprised that nobody's mentioned the space applications yet.

What's the one environment where you have limited mass/cube to store spare parts, may need a critical part on hand within minutes to hours, and can't get one shipped for years?

IMHO, this technology offers huge risk reductions for long-duration missions. Pack a couple of tons of goop in a tank and use it to build anything from pipes to emergency rover wheels. If the technology advances to allow printing of electronics, then new instruments developed and tested here can be built on site without having to be shipped across the solar system.

One note... while I doubt this technology will move as fast as its supporters hope, I think that the economic implications will be larger than expected. The Industrial Revolution took people off farms and put them in unimagineable sweatshops--and later gave rise to today's society (which is a heck of a lot better than subsistence farming). This could do something similarly drastic.

Whether or not Drexler-like nanotechnology comes about this century, I am convinced we will see self replicating machines, whether it be nano, micro, or macro. Long before that we will see sophisticated personal fabbing equipment. Given an energy source (solar or nuclear) and space resources, many things become possible. You can then seriously consider truly terraforming Mars and advanced habitats throughout the solar system.

It'd also be quite interesting if in the future these fab plants were also able to economically recycle/reuse the objects it produces. Instead of throwing away unused items or putting them into storage, one could just toss them back into the fab plant.

This would be particularly nice for items which are only used intermittently, like party decorations or winter clothing.

Neil, I would have thought that would be the default -- considering how quickly my printers go through ink, I'd love to be able to recycle old printouts that I don't need to save, if the dang machine could extract the used ink and reuse it.

The materials for a fab would be more expensive and quicker to deplete, so it would only make good economic sense.

For something like this to happen would really hinge on the ability of processors to be able to self test themselves. There are a great many people who's sole purpose in a semi-conductor business is to sit around and run testers on chips rolling out of the fabs. Dependant upon how many tests a processor passes determines the various levels of capability for a given processor. Lots of reduncancy is built into the design of a chip and if one area of the processor doesn't survive the manufacturing process then it can still find a useful role in a less capable function. This is most evident in graphics processors for computers where chips are built up in stages and blocks and if any one block fails then they just market the chip at a lower price point with a slower performance margin. Processors that would be fabricated within ones home would need the same type of self testing functions where if part of the processor failed then other parts could attempt to kick in redunancy or perhaps even just completely change the behavior of the item to cope with the lost of functionality. Perhaps a toy cow that would sing ABC's and make moo sounds when you pet his tummy. But when the processor failed then it would no longer sing ABC's but would instead make mooing sounds in the cadence of the ABC song.

The big problem with 'fabs' is that the parts they make are *more* expensive than mass produced items.

This is true, because of the machines themselves are typically more expensive to run, but even if they weren't there's the second problem:

-> In order to build something you need the instructions of how to build the thing you want to build!

In other words, you typically have to pay somebody to come up with a design, to test it, and rework it until it's working well enough to be useable. That's the real problem. I mean marginal cost of a car is probably only a few hundred dollars or so, but add on the wages of people and R&D and so forth it becomes much more costly.

Now, whilst the idea of open source plans might well work; that's going to take a long time to sort itself out. So for the foreseeable future fabbed equipment is unlikely to be cheaper and probably will continue to be a spec.alist sideline thing (but with occasional spectacular and/or gimicky mainstream breakthroughs).

So fabbing doesn't actually damage the economy, indeed it may enable new economical activity.