The "scary" technologies/breakthroughs

[Gilthan]

I don't put much stock in grey goo nanoswarms, simply because unicellular life has had aeons to dominate every single energy-using niche on the globe. Sometimes you just need a bigger machine to do a job effectively.

Any entity approaching making "grey goo" would be greatly to be feared or respected, as self-replicating technology in general is not to be underestimated.

Self-replicating technology is of astronomical potential power if avoiding the limits of biological life, such as an artificial mixed ecology including some bigger units. Even earth's crust contains on average 110 times its equivalent energy in TNT explosive due to its thorium content, meaning sufficiently advanced technology could in theory sustain itself eating rock. (Around 1 W/m^2 inefficient biological primary-producer solar power utilization is weak compared to 2.7 GW-years/m^2 available thorium on land or 7 GW-years/m^2 available ocean deuterium, let alone extraterrestrial resources).

Constructively (such as barely comprehensible quadrillions of kilos of wealth) or destructively (such as trillions of insect-size smart missiles), having the ratio of productive output to human labor input approach infinity is a potential change impossible to overstate.

On the other hand, AI superintelligences are liable to be developed first, being indeed the entities likely to accomplish the vast challenge of creating self-replicating technology. After all, on a logarithmic brainpower and intelligence scale where a fly would be 1, a mouse 3, a dog 5, and a human 6, having AI progress from 1 to 6 while just hovering there for long would be a doubtful coincidence compared to instead further orders of magnitude in both speed and complexity of thought soon after. And that, superintelligent AI, is the fundamental ultimate threat or hope and promise.

Still, a lot of the applications of Nanomachines we see are more or less bullshit (Rapdid fabrication, most "nanoweapons" and by extension Grey Goo).

Why? Simple, these tasks require enegy, and the machines are individually quite vulnerable.
How are they going to get all this energy to assemble/disassemble anything of any notable size (say, a human or even a wrench)?
And if they are used in a uncontrolled enviorment - what is going to protect them from wind, rain and all the other nasty stuff going on on this planet? Even if it does not destroy them, it will easily split up the swarm and destroy its ability to work together.

That'd be partially true if (1) all bots are universally the same minuscule size, none bigger than a few microns or whatever (2) the designers are morons.

Neither assumption has to be valid for an artificial ecology of self-replicating machines.

Larger bots can be nuclear-powered, like the earlier example of each ton of the average rock in earth's crust having trace thorium with equivalent energy potential to 110 tons of TNT explosive, or the deuterium in seawater. Barring unknown breakthroughs or micro-scale aneutronic fusion (secondary reactions probably still leading to too much penetrating neutron or high-energy gamma radiation even for rad-hardened hardware more resistant than biologicals), smaller bots utilize energy from the bigger entities in the artificial ecology.

One example would be the larger bots flooding areas within meters or kilometers with inductive energy transfer, using coupled magnetic resonances. Alternatively, even the bigger nuclear bots resupplying the smaller ones with synthetic chemical fuel made from carbon/hydrogen in air and water/rock would work, as such can be 10x the energy density of the poorer-quality water-diluted food consumed by biological insects. For instance, a 100kg mass of a hundred million 1-milligram bots of 1mm dimension each with micro jet engines could most certainly go anywhere insects can fly or anywhere a 100kg human can go. If used as a weapon, only 1 of those 100 million needs to deliver a poison payload, since the world's most toxic poison is lethal in microgram payloads. Coordinated AI swarms with explosive charges could defeat barriers or armored suits, and, as long as they're made sufficiently EMP hardened, the result can be more than effective.

As for the very smallest bots, pure nanobots alone, how do you think the death toll from the 1918 pandemic would be different if it had instead done what no mere natural biological pathogen could, by instead spreading dormant to about the entire populace without causing any noticeable symptoms in anyone before, finally, by remote-control or a preset atomic clock timer, suddenly releasing a few micrograms of toxin in everyone in the same instant?

Of course, the potential for self-replicating technology to be used constructively is greater still than mere destructive capabilities, and, as argued earlier, the challenges to development of such appear so high that general AI first is more likely.

[Molyneux]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses. The critical point will most likely be sometime shortly after a fabricator becomes capable of producing another fabricator.

[Sky Captain]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses. The critical point will most likely be sometime shortly after a fabricator becomes capable of producing another fabricator.

Is this even remotely possible in near term future (10 - 30 years)? I can see home fabrikators producing something simple like teacup or simple parts for something else, but being able to build another fabricator is a whole new level. After all such machine will contain complex electronics that currently are possible to produce in big dedicated factory and require complex manufacturing procedures and all sorts of different raw materials combined in elaborate ways.

[Zixinus]

Perhaps we won't get something as complex as home producing similar-level home fabs, but a large variety of other thing: simple electronics that were previously unimaginable to be produced at home, such as video game controllers or even mid-quality (for the time) TV screens.

I get that the general rule is that you use simple tools to make complex tools, and with complex tools you make precision tools and with that, you can make anything. So there likely still be a knowledge-gap.

[Starglider]

Is this even remotely possible in near term future (10 - 30 years)? I can see home fabrikators producing something simple like teacup or simple parts for something else, but being able to build another fabricator is a whole new level.

Well, exactly how good 3D printing can get, without requiring full scale nanorobotics, is an open debate. However as long as you are able to accept some materials constraints (e.g. everything made out of plastic), the answer may well be 'good enough to be seriously disruptive'.

I wouldn't classify it as 'scary' though.

[Singular Intellect]

I wouldn't classify it as 'scary' though.

I daresay more intelligent and educated individuals don't fear technology, they tend to embrace it. This doesn't mean there isn't any understanding of risks or respect for it's potential power, however.

I personally don't subscribe to the concept of 'bad' technology. Only bad and stupid applications/uses of it.

[Starglider]

I wouldn't classify it as 'scary' though.

I daresay more intelligent and educated individuals don't fear technology, they tend to embrace it. This doesn't mean there isn't any understanding of risks or respect for it's potential power, however.

That's not what I mean. Stupid people evaluate technology in terms of direct threat, whether real or imagined. Nuclear bombs blow you up, nuclear waste gives you cancer, AI makes robots that kill you, genetic engineering makes you slaves of the master race, whatever. Replicators don't pose an obvious physical threat to the average person. If they work in manufacturing then they'll fear for their job, but most first world citizens work in service industries, and would probably perceive it merely as a cheaper and more convenient way to satisfy their consumer cravings.

Unless it can replicate people of course, in which case there's some body horror and the terror most people experience when you threaten their notion of 'i am a unique indivisible being'.

[Simon_Jester]

The thing is? Most of the fears about new technology aren't coming from intelligent people. If you only focus on the "realistic" fears you aren't going to be left with very many. I mean fuck, at the time the atom bomb was being developed people were afraid that once they detonated it the chain reaction would never stop. Which really isn't that much more radical than the LHC making a super black hole. (Considering the LHC really could make singularities that last for a minuscule amount of time).

I'd debate that premise. For example, nuclear bombs present the very realistic fear of civilization being destroyed in an avalanche of huge fiery explosions, because they can actually do that. Genetic engineering presents the very realistic fear of terrible plagues, because (to be frank) genetic engineering could do that. Maybe not by accident, or easily, but it could be done. And so it goes.

Obviously, even these realistic fears can be ludicrously overplayed by people who aren't good at risk assessment, but that doesn't mean they aren't there or aren't relevant.

So there does need to be some limit to our technophilia, because it's entirely possible for us to find a tool for doing something that, on inspection, we don't want done. It shouldn't surprise anyone that this happens. After all, the set of all possible events includes many things no sane person would want to bring about. Why shouldn't that be equally true of the set of all possible technologies?

[General Zod]

So there does need to be some limit to our technophilia, because it's entirely possible for us to find a tool for doing something that, on inspection, we don't want done.

I'd say that's rather up to the OP, as opposed to anyone else suggesting that <fear x> is more valid than <fear y>.

[Ariphaos]

Is this even remotely possible in near term future (10 - 30 years)? I can see home fabrikators producing something simple like teacup or simple parts for something else, but being able to build another fabricator is a whole new level. After all such machine will contain complex electronics that currently are possible to produce in big dedicated factory and require complex manufacturing procedures and all sorts of different raw materials combined in elaborate ways.

You are quite wrong. Printing traces on a substrate is not a magical process. No, they're not able to do their own submicron lithography (yet), but they don't have to do much to start having an impact. There is also the possibility of printing (and recycling) rubbers, ceramics, etc.

[Ryan Thunder]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses.

How do you figure?

[Molyneux]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses.

How do you figure?

How long do you think car companies will survive once it's feasible to create a car from raw materials, in your own home, with virtually no labor costs involved? The first person to think of licensing automobile designs would clean up, but any companies that tried to avoid the paradigm shift - like we're seeing with music now - would lose a hell of a lot of money.

Now if you can make your own household appliances, and televisions, and computers...I don't expect this to actually happen, but I could buy people being frightened that it would.

<edit> corrected "will" to "would". Damn tenses...</edit>

[Gilthan]

Any home fabricator actually affordable to have in the average house when used a few times annually would have a far lesser ratio of capital cost to production if utilized in a factory operating constantly most of the 8000 hours in a year.

As for myself at least, since I only buy a handful of consumer goods per few years (electric shaver, cellphone, laptop computer, etc), I want each one to be top performance, in every way down to the strength of their case materials. For instance, I could see a relatively near-term fabricator making modified simple low-resolution circuitry allowing it to technically make a simple computer or calculator, yet that wouldn't compete with the newest model of CPU from Intel.

Home fabricators making custom plastic toy figures or models would be easier, though, already starting to be available.

Rapid-prototyping and 3D printing does have its uses, especially in more expensive industrial models, for everything from speeding up a major corporation's product development time to putting one in the machine shop of a large spacecraft on an interplanetary mission to help make replacement parts for repair.

To the degree that technology increases economic productivity, hopefully eventually we see a reduction in the standard workweek, down to 25 hours and below. There's not been much progress there in recent history, but we have already at least dropped well below what was expected of 19th-century factory workers.

Home fabricators able to truly produce almost anything would be the domain of almost arbitrarily advanced technology like vast numbers of inexpensive nanorobots, way out of the league of current development.

[Gilthan]

I would be curious, though, if we might see home 3D fabricators not so much for consumer goods as for food. Many of us buy only a few kilograms of plastic and metal consumer goods a year, but we eat a lot more food. I don't mean Star Trek instant food replicator magic but just something that had many stored feedstocks of everything from flour to frozen meats and could produce complex meals as good as the world's best cooks.

Speaking of other technologies which would be disruptive yet not scary:

Most of the retail industry is begging to be disrupted, all the inefficiencies like so many human cashiers and long lines. Walmart has automated checkout machines, but only in a few lanes so far, while there's a big mismatch between its system based on weight sensors versus a huge percentage of customers not being smart enough to figure out how it works. (Typically, out of several lanes, half of them soon get held up for minutes because a customer put an item straight into their cart, causing it to lock up until a store employee enters their code). However, given some redesign, there's no reason many stores couldn't operate completely automated and unattended most of the time, aside from the ocassional security guard. Even janitorial work is mostly suitable for near-term robots.

[Sarevok]

Just look at the immense line of industries working in concert to bring you a simple ballpoint pen. How do you shrink down an entire nation or severals industrial base into a "home fabricator" to build a modern automobile ?

[Starglider]

How do you shrink down an entire nation or severals industrial base into a "home fabricator" to build a modern automobile ?

The same way that a common houseplant produces a stunningly sophisticated minature solar power plant/hydralic pumping system using nothing but air and mud. Huge specialised plants expoliting massive economies of scale only make sense when your techniques for rearranging atoms are at an intermediate stage of development. Primitive techniques only worked at small scales, and supported only cottage industry. Future techniques will mirror what growing living organisms do, but with vastly more speed and flexibility - likely prompting a return to small-scale localised production. Having something in every dwelling capable of building a vehicle may never make sense, but having several in every small town will.

[Gilthan]

How do you shrink down an entire nation or severals industrial base into a "home fabricator" to build a modern automobile ?

The same way that a common houseplant produces a stunningly sophisticated minature solar power plant/hydralic pumping system using nothing but air and mud. Huge specialised plants expoliting massive economies of scale only make sense when your techniques for rearranging atoms are at an intermediate stage of development. Primitive techniques only worked at small scales, and supported only cottage industry. Future techniques will mirror what growing living organisms do, but with vastly more speed and flexibility - likely prompting a return to small-scale localised production. Having something in every dwelling capable of building a vehicle may never make sense, but having several in every small town will.

However, if you're referring to the capabilities of fully advanced nanorobotics for "rearranging atoms," wouldn't you agree that being able to produce near-infinite trillions of complex micromachines for tiny cost would be a still greater R&D challenge and investment than developing general AI and hence likely to occur in a later year than when humans become no longer the most intelligent entities on this planet?

Otherwise, without that extreme level of technology, you'd be talking about something that could optimistically technically make automobiles of a sort perhaps (depending upon many things like how good it is at making rotating bearings of high-strength high-melting point metals!) but would still fall under the domain of normal economics, meaning economies of scale apply, with having such in a factory operating 8000 hours a year being better if trying to minimize the cost of those cars produced.

[Starglider]

However, if you're referring to the capabilities of fully advanced nanorobotics for "rearranging atoms,"

I'm not. Plants aren't built out of nanorobotics. When the phrase 'nanotechnology' first started being thrown around, people automatically equated it with super-advanced self-replicating general assemblers. Now that it's been used as a commercial buzzword, it's either that or near term nanostructured materials stuff like nanoparticles and tubes, seemingly with no intermediate. Actual develoment of the technology will not jump magically from the later to the former (not if it follows the normal human R&D process). There is a huge spectrum of capability and sophistication between where we are now and optimal general assemblers. In particular, construction with a fixed planar array of microtools (extruders, manipulators, cutters etc) and a reasonable set of refined feedstock materials is a much easier proposition than growing an active structure in place the way a plant does, much less controlling and powering billions of freely moving robots.

would be a still greater R&D challenge and investment than developing general AI and hence likely to occur in a later year than when humans become no longer the most intelligent entities on this planet?

Well sure, but AI has already been mentioned as 'dangerous technology' candidate.

meaning economies of scale apply, with having such in a factory operating 8000 hours a year being better if trying to minimize the cost of those cars produced.

Economy due to better utilisation (duty cycle) is something entirely different from economies of scale due to plant size and centralisation.

[Gilthan]

However, if you're referring to the capabilities of fully advanced nanorobotics for "rearranging atoms,"

I'm not. Plants aren't built out of nanorobotics. When the phrase 'nanotechnology' first started being thrown around, people automatically equated it with super-advanced self-replicating general assemblers. Now that it's been used as a commercial buzzword, it's either that or near term nanostructured materials stuff like nanoparticles and tubes, seemingly with no intermediate. Actual develoment of the technology will not jump magically from the later to the former (not if it follows the normal human R&D process). There is a huge spectrum of capability and sophistication between where we are now and optimal general assemblers. In particular, construction with a fixed planar array of microtools (extruders, manipulators, cutters etc) and a reasonable set of refined feedstock materials is a much easier proposition than growing an active structure in place the way a plant does, much less controlling and powering billions of freely moving robots.

Plants are fundamentally what I'd consider natural nanorobotics, formed of self-replicating extremely complex micron-scale machines (cells), protein-based instead of inorganic but still machines really, even if it is against traditional semantics to call anything evolved in nature a machine.

Aside from that, there's indeed some validity in your points there.

However, ordinarily microelectromechanical systems have a very substantial economic cost relative to their volume and size, as we see with CPUs that cost a significant portion of the average individual's income for circuitry within a roughly square centimeter area of some microns thickness, being many dollars per cubic millimeter let alone per ton of CPUs. The cost wouldn't be a problem if production instead approached infinity due to self-replicating technology, but, of course, developing self-replicating technology is a rather hard problem where so far nobody has thought they could just put a few billion (or trillion) dollars R&D onto the problem and get unlimited trillions of dollars benefit back.

If you don't try to go for micro-scale parts for an assembler of macro-scale objects but just operate on a larger scale like computerized milling machines do, then the economics are better with near-term technology. At some point, though, we'd have to ask whether we were talking more about "nanotech home fabricator" or large "increasingly automated factory."

Simply doing 3D printing of shapes is relatively simpler, yet, as you yourself pointed out earlier, has materials constraints (among other limitations), with current 3D printers working with plastics and some particularly low melting temperature metal alloys. If trying to make parts like axles and ball bearings for a car, that'd take more a miniature (or giant) automated factory than just 3D printing.

would be a still greater R&D challenge and investment than developing general AI and hence likely to occur in a later year than when humans become no longer the most intelligent entities on this planet?

Well sure, but AI has already been mentioned as 'dangerous technology' candidate.

Essentially saying "after we and/or our servants attain superintelligent AI godhood, we'll be able to develop supertech home fabricators able to displace most conventional manufacturing" is a lot different from worrying about or looking forward to a more near-term technology developed before then, though.

It's like the difference between saying "your house will be repainted" versus "your house will be repainted after we have starships."

The near-term versions of household fabricators, are, after all, closer to an improved successor to this:

http://www.dimensionprinting.com/3d-pri ... print.aspx

meaning economies of scale apply, with having such in a factory operating 8000 hours a year being better if trying to minimize the cost of those cars produced.

Economy due to better utilisation (duty cycle) is something entirely different from economies of scale due to plant size and centralisation.

A good semantic, technical distinction in usage of terms can be made there. However, this is a case where duty cycle is highly related to centralization. Put a given piece of manufacturing equipment in someone's home, and the mass of production they need per year isn't high enough to keep it in constant operation. Put it in a more centralized location serving a whole state or metropolis, in a big factory, probably making its design of bigger size in the process too, and maximum duty cycle can better get obtained.

Besides, among many other factors, a really advanced fabricator, like your hypothetical example of one making whole vehicles, would need a lot of different feedstocks and regular resupply of everything from perhaps aluminum to tungsten, even varying how much of each had to be resupplied each day if the items it made varied. Those kind of resupply needs are easier met with centralization, much like it is economically cheaper to have a big coal power plant for a city than a little engine-generator in each house each needing its own frequent coal resupply.

Resupply complexity is less, of course, for a more near-term fabricator just making simple parts, toys, and models mostly of the same thermoplastic resin.

The main disadvantage of centralization is increased transport time, but an automated transport system rapidly sending goods from manufacturers to individual residences throughout a city or nation is comparably low-tech versus much of what we've been talking about.

When it comes down to it, the biggest expense we have that most needs reduction isn't household consumer goods. It is houses. Residence costs are a far higher portion of income than toothbrushes. Equipment big enough to make whole houses in hours, by any mixture of methods including casting, would greatly beat taking months for a team of laborers to do it. Thomas Edison tried that in the early 20th century with complex concrete castings including even concrete furniture inside, but his particular equipment had some practical issues, while concrete alone isn't very aesthetic.

[Coalition]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses. The critical point will most likely be sometime shortly after a fabricator becomes capable of producing another fabricator.

One use I can think of for a home fabricator would be building plastic shapes. Ranging from cups to plates to replacement plastic shells for a game controller, the fabricator takes in basic feedstock, and the user purchases the designs from an online source. The other option would be someone using a 3-D CAD program, and using that to make what the owner desires. Need a bunch of forks for a picnic dinner? Just fab them up and cut them off the sprue.

Make too much spaghetti (got noodles and sauce on sale), and you need a plastic container to store the extra, in the right size for dinners? Download the design for a plastic container of the volume you select, and have the fabricator churn a few out. Bought 50 pounds of rice, and don't want bugs eating it? More containers.

Kid has a project for school and you need an ant farm? Fab up the plastic holders, and buy a pair of clear plastic windows (unless the plastic is transparent).

Models needed for the gaming group? Fab them up. Games Workshop will scream at this point though.

If the material isn't microwave safe (or the form becomes permanent in a microwave), then you still need your regular plates, containers, etc.

When the containers are empty, you wash them, and use the chemical stick/funnel to render them down to raw materials for later use. The chemicalstick lets you 'cut' a piece of recycloplastic to a size where it fits in the funnel, and the funnel then dissolves the outside of the plastic, so it drips down into the feedstock container. The chemical slowly gets used up in the process, requiring eventual replacement.

Chemical sticks and funnels sold separately (at a high markup, plus different companies will have their own chemicals for plastic and sticks).


Edit - I forgot the main one: home-made custom LEGOs!

[salm]

My money for the next "scary" tech is home fabricators. Expect people to be scared as hell about the "end of property", and possible economic collapses. The critical point will most likely be sometime shortly after a fabricator becomes capable of producing another fabricator.

One use I can think of for a home fabricator would be building plastic shapes. Ranging from cups to plates to replacement plastic shells for a game controller, the fabricator takes in basic feedstock, and the user purchases the designs from an online source. The other option would be someone using a 3-D CAD program, and using that to make what the owner desires. Need a bunch of forks for a picnic dinner? Just fab them up and cut them off the sprue.

Make too much spaghetti (got noodles and sauce on sale), and you need a plastic container to store the extra, in the right size for dinners? Download the design for a plastic container of the volume you select, and have the fabricator churn a few out. Bought 50 pounds of rice, and don't want bugs eating it? More containers.

Kid has a project for school and you need an ant farm? Fab up the plastic holders, and buy a pair of clear plastic windows (unless the plastic is transparent).

Models needed for the gaming group? Fab them up. Games Workshop will scream at this point though.

If the material isn't microwave safe (or the form becomes permanent in a microwave), then you still need your regular plates, containers, etc.

When the containers are empty, you wash them, and use the chemical stick/funnel to render them down to raw materials for later use. The chemicalstick lets you 'cut' a piece of recycloplastic to a size where it fits in the funnel, and the funnel then dissolves the outside of the plastic, so it drips down into the feedstock container. The chemical slowly gets used up in the process, requiring eventual replacement.

Chemical sticks and funnels sold separately (at a high markup, plus different companies will have their own chemicals for plastic and sticks).


Edit - I forgot the main one: home-made custom LEGOs!

This is something that could allready be done with current 3D plotters. They´re a bit too expensive now for the average household but sooner or later the prices will drop and everybody will have such a thing at home.
It´s going to be a horror scenario for large companies but a great chance for small companies and freelancing product designers who now can dirrectly sell their blueprints (or more specific, a 3D Model) to the consumer.
Pirating objects is going to be a problem, just like pirating software and music. Perhaps there´s even going to be a modding scene.

However, 3D plotters that can create plastic objects are a far cry from machines that can make objects out of different types of materials, let alone complex eletronic devices.

[Starglider]

Plants are fundamentally what I'd consider natural nanorobotics, formed of self-replicating extremely complex micron-scale machines (cells)

The components in a cell do not proceed under positive control. They are produced, travel randomly around by diffusion, and engage in local interactions. A robot has a control system that co-ordinates the actions of its parts, almost always a central CPU. Organic life does not implement this until the scale of relatively complex multicellular organisms is reached. Furthermore, designed industrial systems almost always use channeled transport (e.g. pipes, power cables), segregation of processes (e.g. specific reaction vessles for each step of the process) and relatively rigid structures. Nanorobotics designs generally do too. By contrast cells are a flexible bag of solution with a huge number of reactions taking place in the same vessel, with almost exclusively diffuse rather than chanelled transport and control via independent feedback loops rather than a central processor. This structure exists because it was (relatively) easy to evolve, not because it's particularly efficient or a good idea (if designing from scratch).

as we see with CPUs that cost a significant portion of the average individual's income for circuitry within a roughly square centimeter area of some microns thickness,

What? The CPU in your washing machine costs about ten cents. Even the CPU in your phone costs less than $10. Oh sure, if you really want you can spend thousands of dollars on an enterprise server class CPU, but why bother? Even cheap-and-nasty cellphones now have CPUs that would power a state of the art workstation in 1992. Washing machines have CPUs that would be state of the art in 1986. See where this is going? At current rates of litho progress, anything that initially costs $1000 to produce will drop to $1 over the next 20 years.

Well sure, but AI has already been mentioned as 'dangerous technology' candidate.

Essentially saying "after we and/or our servants attain superintelligent AI godhood, we'll be able to develop supertech home fabricators able to displace most conventional manufacturing"

This is for a play, for presumably for non-technically literate audiences. It's going to be hard enough to deal effectively with a single issue. The subject of how technology as a whole will develop in the next fifty years is way beyond the scope of this thread.

Put it in a more centralized location serving a whole state or metropolis, in a big factory,

You are missing the point. Increasing generality means that you no longer have to care about the number of a specific product required, you simply have to care about the combined mass/volume of demand for all the products that type of replicator can produce. Centralised factories are needed because current factories work effectively when set up to produce large runs of one specific thing, or perhaps a few variations thereof (e.g. in a car factory). When a single device can produce a large range of products, a much smaller region can produce enough aggregate demand to keep it occupied nearly continuously.

Besides, among many other factors, a really advanced fabricator, like your hypothetical example of one making whole vehicles, would need a lot of different feedstocks and regular resupply of everything from perhaps aluminum to tungsten, even varying how much of each had to be resupplied each day if the items it made varied.

An argument for having them about as common as existing supermarkets, not as common as existing large-scale manufacturing plants.

Those kind of resupply needs are easier met with centralization

Shipping the feedstock to distributed manufacturing sites is a simpler and cheaper proposition than shipping all the different products to distributed collection sites, even moreso because these fabricators are likely to be co-located with nanotech recycling stations that supply much of the feedstock need.

[salm]

Not sure if this really counts but small portable video projectors could be seen as the next "evil" such as computer games, vhs recorders, televisions and movie theaters were said to turn our youth into mindless zombies.

Imagine everybody having a projector in his cell phone. Walls and floors in public places will be plastered with porn and gore. Esspecially on the weekends when people are drunk.

[Gilthan]

What? The CPU in your washing machine costs about ten cents. Even the CPU in your phone costs less than $10.

Oh, you're referring to an integrated circuit chip with much larger and orders of magnitude less complex circuitry than a modern PC CPU. That's pretty far from nanotech. But okay. Well, for instance, taking the example of going to a major electronics component supplier:

http://www.mouser.com/Semiconductors/In ... =Pricing|0

... the cheapest 4-bit microcontroller they have costs $1.80 each, which includes a very simple 4-bit CPU. (They also sell cheaper models of discrete ALUs for about the same price).

What is the cost per unit volume? Its total SSO 20 package amounts to 0.039 cubic centimeters volume, but so much of that would be just packaging as opposed to anything with near-micron-scale complexity.

So, even there, the result would still be at least on the order of $100 per cubic centimeter or proportionally $100000 per cubic liter of near-micron-scale complex circuitry.

At current rates of litho progress, anything that initially costs $1000 to produce will drop to $1 over the next 20 years.

But cost per unit area / volume isn't decreasing according to Moore's Law, rather cost relative to the number of transistors within a given area. It is true that there are around $1 IC chips, yet such was also so a decade or two ago, even though the performance of those IC chips has increased. There's no reason to conclude we'll soon see IC chips of the same size that cost only $0.001 each or IC chips of 1000 times the volume that cost $1 each.

Admittedly, this IC comparison isn't entirely fair. If one looks at how how cheaply a high-resolution printer working with paper and ink (instead of semiconductors) can print at 1000 dpi, 0.025mm resolution, we can see by analogy that theoretically some form of non-semiconductor-based micron-scale IC could be made a lot cheaper per unit area of complexity. Still, even there, per volume the cost is high, not to mention that 0.025mm-scale isn't nanotech.

Different products have different appropriate manufacturing methods. Even you, I'm sure, wouldn't suggest we replace injection molds able to mass produce household plastic bottles of multi-liter size for ~ $0.01-$0.10 each with instead slower, more complex nanotech scale assembly lacking the economics of self-replicating tech?

When a single device can produce a large range of products, a much smaller region can produce enough aggregate demand to keep it occupied nearly continuously.

Maybe we should better specify this. What range of products is your hypothetical replicator envisioned to produce at a cost and performance competitive with more specialized, centralized manufacturing factories?

Electric shavers? Cell phones? Laptops? Water bottles? Trucks?

And what approximate time frame and technological level are you envisioning for when it exists?

Shipping the feedstock to distributed manufacturing sites is a simpler and cheaper proposition than shipping all the different products to distributed collection sites, even moreso because these fabricators are likely to be co-located with nanotech recycling stations that supply much of the feedstock need.

You're arguing for hypothetical replicators about as common as supermarkets, right? Either way, whether that or my more centralized production, the product has to be loaded and shipped from the manufacturing site to a house then. The difference is just such as whether it spends 0.5 miles and maybe 5 minutes in transit versus whether it spends 50 miles and maybe up to a hour in transit depending on transportation technology.

A given individual like me purchases only a few times a year a new consumer product like a new calculator, laptop, shaver, etc. So not more than a small number of dollars annual expense are involved in shipping those consumer goods to my residence.

Also, why assume nanotech recycling necessarily? Especially if this is to be pre-self replicating technology, more bulk methods could more simply recycle all waste, assuming abundant nuclear or solar energy. Just subject everything to bulk processing using temperature, etc. to separate out hydrocarbons, molten metal, metal oxides (to refine into metal versus waste oxygen to return to the atmosphere), etc. If trying to process many tons a day, that requires simpler, cheaper equipment than trying to have countless complex nanotech manipulators breaking down everything on a micro scale.

Here's one partial method even now:

The PWC process doesn't combust waste like an incinerator. Instead, the plasma, an electrified gas, is discharged within the PWC chamber. The continuous arc produces temperatures as high as 30,000 degrees Fahrenheit. The energy within the plasma excites the waste material's molecular bonds so that the material separates into its atomic components.

Once this dissociation occurs, the atomic components can reform into recoverable, saleable, nonhazardous commodities. This process can be used for solid, liquid and gaseous wastes, including hazardous and nonhazardous wastes, organic andinorganic solids, bases, and aqueous and non-aqueous liquids.

The PWC processes wastes so that they can be recovered in phases. Outside of producing hydrogen, the PWC produces PCG, which can be used as chemical feedstock to produce polymers and other chemical products. PCG also can be used as clean fuel to produce electricity, desalinate sea water or heat buildings.

Depending on the waste feedstock, other commodity products can be processed. For example, if the wastes contain sufficient metallic materials, these will collect in liquid form. Inorganic, glass-like silicates form a separate layer with small quantities of metal encapsulated in the silicate stone. These ceramic-like silicates can be used as quality aggregate material, and in the abrasives industry.

Startech's system does require electricity to create the plasma. However, when feedstocks rich in energy are processed, the PWC is a net energy producer. Each unit of electrical energy used to process wastes yields approximately four units of energy recovered with synthetic PCG gas.

PWCs also reduce the waste's volume by an approximately a 300 to 1 ratio.

Startech says the PWC can process waste at 8 cents to 11 cents per pound, including capital and financing costs plus other operating expenses such as consumables, labor, spare parts and electricity.

Startech's first commercial system is operating at the U.S. Army's Aberdeen Proving Ground, Md., processing up to 7 tons per hour.

http://wasteage.com/mag/waste_technology_waste_meets/

A system with more refinement would be more complex, such as not just getting a generic liquid metal stream out but separating back into iron versus nickel, et cetera, or making specific alloys. However, it doesn't need nanotech levels of complexity.

[Starglider]

Oh, you're referring to an integrated circuit chip with much larger and orders of magnitude less complex circuitry than a modern PC CPU. That's pretty far from nanotech.

A typical phone CPU has one to two orders of magnitude fewer transistors than a desktop core. This is a small gap relative to the difference in component counts between a processor and a nanofabricator.

... the cheapest 4-bit microcontroller they have costs $1.80 each, which includes a very simple 4-bit CPU. (They also sell cheaper models of discrete ALUs for about the same price).

Oh come on, a trivial search would turn up something like this, an 8-bit microcontroller that costs 34 cents each. That's a sticker price for 100 units, I guarantee you that a manufacturer negotiating a contract for 100,000 chips will be paying half that.

But cost per unit area / volume isn't decreasing according to Moore's Law, rather cost relative to the number of transistors within a given area.

Of course the cost per area isn't decreasing (no, actually it is; modern CPUs have larger dies despite being cheaper) when the feature size keeps decreasing. If the feature size is held constant, then the cost per area will decrease. You can see this quite clearly in that older semiconductor processes are used for progressively cheaper chips.

Even you, I'm sure, wouldn't suggest we replace injection molds able to mass produce household plastic bottles of multi-liter size for ~ $0.01-$0.10 each with instead slower, more complex nanotech scale assembly lacking the economics of self-replicating tech?

That primarily depends on the cost of transportation. Right now, transportation is cheap relative to production, a fact that has dictated global trade patterns. If transportation becomes more expensive (which is a frankly likely outcome of oil depletion), localising production becomes more attractive. You are also completely ignoring the inherent customisation advantages of local direct assembly, relative to mass production where varying the output even slightly is difficult and costly. Correctly marketed, this is an extremely attractive advantage.

Maybe we should better specify this. What range of products is your hypothetical replicator envisioned to produce at a cost and performance competitive with more specialized, centralized manufacturing factories?

I am not hijacking this thread any further, start a new one.