Transformative Technologies Wednesday, Sep 26 2007
futurism 1:17 am
Three technologies I predict will be the origin of transformative technological changes in the coming century are:
- molecular manufacturing
- artificial intelligence or brain-computer interfaces, but not both
- solar or nuclear energy, but not both
By origin, I mean that a particularly large breakthrough in one of those areas will have effects that ripple into every other field of technology and science as well as politics. For example, molecular manufacturing will make it easier to build anything, which will have completely transformative effects in building construction, aerospace and space travel, cybernetics, etc. I wouldn’t list “cybernetics” as an item on this list, because it’s an example of a technology that will accelerate rapidly when enabled by advances in another area — in this case, the quick fabrication of nanostructured materials, actuators, and sensors. Molecular manufacturing will open the floodgates of cybernetics, not the other way around.
In the same way, I believe that brain-computer interfaces will either progress to the point where creating a theory of intelligence for AI programming becomes a lot easier, or we create advanced AI before sophisticated brain-computer interfaces (BCI) and use it to make critical breakthroughs in analyzing the huge amounts of biological data necessary to implement BCI.
My model of future technological advances is not like a line of marching soldiers, where each field progresses steadily along, but more like a fireworks factory going up in flames: one spark ignites numerous powder kegs and the whole thing goes up in a great explosion. My current idea is that the spark will be something involving intelligence enhancement technology, either brain-computer interfaces using arrays of billions of nanowires as interface mechanisms, or artificial intelligence based on some scheme of merging together statistical inference with sequential decision theory. Although, in retrospect, it is possible that molecular manufacturing may be the key spark, as when this is introduced, it will be a tremendous enabler for both AI and BCI research. This is primarily through the nine-or-more orders of magnitude computing power increase it will offer, unless the availability of such resources are restricted through regulation.
For solar vs. nuclear energy, I do think that one will dominate the 21st century, but can’t tell which yet. Anyone who says one or the other will definitely dominate is intellectually biased. Fossil fuels are obviously on their way out. Solar can be distributed everywhere, which is good, but it produces negligible power relative to harnessing the nuclear strong force (the strongest force in the universe), and consumes real estate. When we can manufacture carbon nanotubes in ton quantities, they will be available for covering the walls of reactor cores, as well as pipes able to hold extremely high-pressure steam, which will miniaturize nuclear power plants while maintaining energy output. With these infrastructural improvements, the large, clumsy plants of the past will seem like ancient history. Irrational environmentalist hangups over nuclear power could evaporate, as we see them doing so today, and the world could embrace the power of the atom. On the other hand, high-efficiency solar cells laid down in the world’s deserts by autonomous robots could power all of human civilization, if deployed aggressively enough, making nuclear somewhat unnecessary for the near term. Only the future will tell.

September 26th, 2007 at 8:05 am
On solar power, I’m a little surprised to see the conjunction of “autonomous robots” and the usual desert-real-estate scenario. There’s a desert (or, at least, no cloud cover, no storms) just a few miles from you right now: it’s right over your head. As you probably know, what we need for our current total power needs is 0.01% of that energy, so even with rather inefficient solar collection we’re not talking about blocking the sky. I sketched an oversimplified version a year or so ago. (I would at least mention direct decomposition if I were doing that now, and the fact that the generators — of whatever form — need not ever come down, because our autonomous robots will include some very big transporters that can trade water for compressed/adsorbed/whatever hydrogen.)
I realize it’s not central to your point, but I do think that there’s a quite high probability that solar will dominate the latter parts of the 21st century: not the solar I just described, but solar collected in orbit by billions of square miles of reflectors, built by self-reproducing robots which work with millions of cubic miles of material from the moon and miscellaneous other sources.
On the other hand, maybe not.
September 26th, 2007 at 8:19 am
As an extension to Tom’s thoughts on solar power in context of power densities — Assuming that someone else starts up where the last private company left off and manages to manufacture a carbon nanotube ribbon “star ladder”, the power utility curve of solar power as direct electrical input into the grid has phenomenal potential. Just because we can’t go all Dyson-Spherical doesn’t mean we can’t utilize solar power that doesn’t reach earth directly. When you consider that there is work being done right now to maintain self-assembling photovoltaic chlorophyll-polymers, (and they’ve had some success) the picture gets a little more… replete.
But, of course, as technology progresses the answers to any given problem become more diverse. The main argument against the orbital photovoltaic ’solution’ is power transmission. IF inertial-electric boron-hydrogen fusion generation (perhaps the inclusion of pyroelectric crystals might help cross the threshold?) could be miniaturized sufficiently to permit high-density energy for transit, there’s room for “all of the above” there.
So I, personally, don’t think either solar or nuclear generation will be in and of itself transformative; but energy generation will be.
September 26th, 2007 at 10:04 am
I agree with technology related to fire model, but also feel the pile of people racing to a goal like a mass start of a marathon or the start of an avalanche models can also be useful.
In terms of the fire model definitely molecular manufacturing and some method of cognitive boosting are critical. Although they will delivery useful technology by themselves, it is their role as fire (technology) accelerants (as in arson) where they are most important. MM and AGI are explosives attached to a metal tank of explosive liquid/gas inside the fireworks factory.
Taking a closer look at the starting points for MM and AGI/brain interfaces, we can see that there are enabling technologies and pathways for them as well. For MM, quantum computers could radically alter the modelling and design of molecular systems. Metamaterials making superlenses or other technologies which could provide the improved microscope and mechanosynthesis capability to enable MM. Sparks near the MM firework fuse.
There are a combination of alternative technologies which could provide some of what early full blown MM can provide. Smouldering dry paper can be made into a big fire if properly tended. It takes longer and requires more work and some skill, but you can still get the big fire. As noted by Michael, the MM and AGI fireworks are still in the same building so if you get a big fire going through other means you will probably trigger the MM and the AGI.
-advanced chemistry, near MM nano and industrial efficiency in large scale carbon nanotube factories. We could make ten of thousands of tons of carbon nanotubes each year without MM.
-Efficient processes for rapidly incorporating new materials (like carbon nanotubes and improved nanostructured metals etc…) into products like cars, planes, bridges, buildings and other products
-using advanced self assembly, lithography or polymer manipulation and redesign of the electronic elements (like swithing to ovonic quantum control devices instead of transistors) and enable rapid reel to reel production of computers and solar cells and other electronics.
-optical computers, spintronics, universal memory etc…
-synthetic biology, gene therapy, gene editing, RNA interference, RNA activation
Where we might have expected a 10 times boost over 6 years in computers and a 30% growth in the overall world economy, we could see 1000 times boost in computers and a 10 times price reduction in solar electronics and a 50-60% growth in the overall world economy.
Other fairly powerful technology fireworks could go off by themselves, but everything is slower and harder and more dependent upon bottlenecks and hurdles not stopping or slowing down things without MM and AGI. The next firework does not go off or the fuse fizzles.
September 26th, 2007 at 1:19 pm
“As you probably know, what we need for our current total power needs is 0.01% of that energy, so even with rather inefficient solar collection we’re not talking about blocking the sky.”
The problem isn’t the difficulty of engineering building a solar system; it’s the cost-benefit analysis. Building enough of these balloons to power the Earth is going to be too expensive to implement compared to nuclear and space-based solar.
“The main argument against the orbital photovoltaic ’solution’ is power transmission.”
Microwaves are around 50% efficient, not bad at all.
September 26th, 2007 at 1:48 pm
Persistent programs with space based solar power has been the rectenna receiver taking up square miles of land on the ground and not scaling down with lower power. So it had to be multi-billions for multi-GW or nothing. Plus there have been the launch cost issues. Also, stabilizing and pointing the orbital systems had fuel consuming systems to counter gravitional gradients.
Also half the energy is wasted in losses
New proposals deal with those issues.
Japan has a direct solar to laser conversion system that is fairly efficient. The laser receivers can scale to smaller size. Thus current launch systems can handle it.
40% efficient conversion
http://advancednano.blogspot.com/2007/09/key-part-of-space-based-solar-40.html
The microwave power transmission details and efficiencies (which would be avoided in early systems)
http://en.wikipedia.org/wiki/Microwave_power_transmission
43 page pdf on a lot of details. 5.8 Ghz is better.
http://www.sspi.gatech.edu/wptshinohara.pdf
http://en.wikipedia.org/wiki/Wireless_energy_transfer
space based solar is a lot more expensive. Initially we need customers that will pay a lot more for the advantages offered by space based solar. Costs of military supply chains for forward bases is one customer
http://spacesolarpower.wordpress.com/
The transforming the world with energy is to get ten times or more power generated and supplied for a lot less than the 1.5-2 cents per kwh that our best systems have now. Take the current 17 Terawatts from all sources and make a lot more of it.
World 2005 about 450 quadrillion BTU
http://www.eia.doe.gov/oiaf/ieo/world.html
http://en.wikipedia.org/wiki/World_energy_resources_and_consumption
September 27th, 2007 at 11:18 am
In terms of energy transmission, that’s horrifically wasteful. Especially when we consider the geometric curve for radiated energy over distance. That’s basic physics, Tom.
A microwave transmission is less than 1/10th as powerful at 10 meters than it is at 1 meter.
Another problem with solar/nuclear power is of course energy storage. With increasing power-densities of commercial-grade batteries seemingly around the corner thanks to self-replication nanoscale technologies, this might become a somewhat moot point.
But I for one won’t be holding my breath.
September 27th, 2007 at 2:55 pm
“That’s basic physics, Tom.”
No, it’s basic engineering, and by engineering standards, 50% efficiency is very good; our very best combined-cycle gas-fired power plants are around 50% efficient. Recharging a battery-powered electric car is around 50% efficient. A standard car-sized hydrogen fuel cell is around 50% efficient.
“A microwave transmission is less than 1/10th as powerful at 10 meters than it is at 1 meter.”
Microwaves can be generated and focused rather easily compared to visible or infrared.
“With increasing power-densities of commercial-grade batteries”
It is chemically impossible to produce a battery which can store much more than a standard Li-Ion. We’ve simply run out of elements; there’s nothing which is better than lithium at storing energy (lithium/fluorine is the best possible standard chemical rocket fuel). Self-replicating factories might make the batteries much cheaper, though.
September 27th, 2007 at 3:18 pm
It might be possible to make new batteries based on different principles, such as flywheels. MNT would at least let us try.
Solar satellites will eventually be a big deal. They would be enabled by MNT factories. Nuclear may still be desired for certain applications.
September 28th, 2007 at 10:12 am
Tom Wrote:
Really. Angela Belcher of MIT disagrees with you. (Note; your wording was “standard” Li-Ion.)
Stay on topic, man. Energy conversion and storage has nothing to do with transmission efficiency. 50% energy loss from transmission is ruinous. Some basic reading for you.
This is irrelevant. Please note where I stated that the element which would make orbital photovoltaic platforms cost-effective was carbon nanotube star-ladders. As you should already know, one individual nanotube has been measured to have the electrical conductivity of a standard copper house-wire. Do the math from there, man.
September 29th, 2007 at 12:21 pm
Another benefit of solar is that it can be used in a decentralized, close to the user way. It does use a lot of real-estate, but it can be real-estate that we’re already using.
Just covering most rooftops with cheap, efficient solar panels would make a huge difference. With some parallel improvements in efficiency (we’re already seeing the beginning of this, but the higher energy costs, the faster it will happen), this could be huge.
September 30th, 2007 at 7:37 am
Leveraging Michael G.R.’s comment, I believe the “decentralized” attribute of solar could allow for a combination both nuclear + solar technologies to work together.
I am not convinced that it’s an either-or proposition between solar and nuclear; they are both sufficiently distinct (and mature) that a hybrid solution is quite possible.
September 30th, 2007 at 9:39 am
Well, of course some solar advocates say they are really nuclear advocates; they just want the reactor to stay about a hundred million miles off.
I would like to respond to Tom McCabe’s “cost-benefit analysis” remark on solar balloons, though: I absolutely agree, as long as the basis is similar to contemporary manufacturing. I just don’t happen to think that contemporary manufacturing will stay put for very long, and I’m not necessarily talking about nanofactories or even the effects of Gordon Moore’s famous Law. Edward Moore (the creator of “Moore machines”, which I’ve lectured on to many computer science classes) described “Artificial Plants” in the October 1956 Scientific American, with the basic idea being that self-reproducing machines will transform cost-benefit calculations in fundamental ways.
To generate product X, you need a collection of factories (F1,F2,..FN) which are capable of manufacturing X, but also capable of manufacturing and assembling copies of one another. We’ve been moving towards this for a couple of hundred years, and it looks to me as if robotic manufacturing, still very limited, is now improving at roughly Moore’s-Law rates. It doesn’t require a Singularity. (But it may accelerate one, and be accelerated by one.) I guess I’d claim that self-reproducing machinery (of the non-biological sort) is itself a transformative technology, which will in some sense dominate the coming century.
(I did submit a version of this with links, and was told it was apparently spam; I’m trying again, a day or two later.)
October 1st, 2007 at 9:25 am
“It is chemically impossible to produce a battery which can store much more than a standard Li-Ion.”
But there’s nothing to say that all future battery technology have to employ chemical processes. Personaly I like the idea of carbon nanotube Flywheels although admittedly that’s purely based on me thinking the idea seems really cool.
October 1st, 2007 at 9:37 am
“and it looks to me as if robotic manufacturing, still very limited, is now improving at roughly Moore’s-Law rates.”
I don’t know why Digital Fabrication/3D printing doesn’t pop up in these kinds of discussions more. The price/performance ratio on these devices is improving at an exponential rate that exceeds Moore’s Law. The Fab@Home guys use a 600$ assemble it yourself kit and they’re hoping to announce self fabrication in 2008 (well self-fabrication of all it’s own parts a human would still have to put it together).
Not as powerful as a Drexler machine to be sure but it has the benefit of already being here and advancing at an exponential rate. Molecular Manufacturing may or may not arrive when people think it will but I’m pretty certain we’re going to have a Fabbing revolution first.
October 1st, 2007 at 1:56 pm
Tom Myers: What you’re looking for there is better discussed as “Von Neumann machines“. Has more pull due to the ‘infamy’ of Von Neumann probes and their impact on the Fermi Paradox.
Namedropping for teh win, apparently.
July 26th, 2008 at 10:14 am
Space-Based Microwave Power
http://www.p2pnet.net/story/16477
If not the US, I’m sure other contries will develop it.