Brian Wang is a long-time futurist listed as a Big Thinker on the Kurzweil AI website. A member of the CRN Task Force and an advisor to the Nanoethics Group and the Lifeboat Foundation, he has a column on the Nanotechnology Now website and his own blog Advanced Nanotechnolgy. He has a degree in computer science and an MBA and has worked in the IT industry for twenty years. He created and ran his own professional computer consulting company with offices in Canada and the U.S. and clients in the U.S. and Europe. For the last eleven years he has lived in the Bay Area, where he has been in touch with the technological changes in computer science and nanotechnology. His talk at the CRN Future of Nano & Bio Conference was entitled “Economics in a New Era.”
The following transcript of Brian Wang’s CRN presentation “Economics in a New Era” has been reviewed and approved by the author.
Economics in a New Era
I will be talking about economics, to a certain extent, and to the lead-up to molecular manufacturing, as well as what I see happening afterwards, talking about business as well as economics. I have some slides which I haven’t included in the main presentation. If people are interested I can show about certain aspects of economics in an advanced technological society. So, I will be going through several slides in terms of what I see as the current technology, and the stuff that I see as coming to fruition by 2015-2020. Then, leading into the nanofactory impact, which I see happening around that time, plus or minus five to ten years. Then, looking at what that means, in terms of the primary impact of fast production as well as the enabling aspect of it. There is a lot of science, a lot of breakthroughs, that we could achieve, but cannot, because of limitations in our current technology. I see molecular manufacturing making it possible to change. I will discuss also that choices matter in all this. I can describe something that is very great and very possible technically, but the choices that society and individuals make affects this a great deal.
Some of this technology may have been gone over in the first few days of the conference. My hit list over the next three slides: metamaterials, superlenses, also new states of matter are enabling things that were thought to be impossible just a few years ago. Being able to focus light, beyond what the diffraction limit thought possible, is going to enable viewing of things down to like 1 nanometer, which would be just a few atoms big. It’s probably about 100 times better than what was able to be done just a few years ago in terms of direct visible light. Quantum computers, there is still some controversy around this area in terms of whether this company in Vancouver, D-Wave Systems, has actually achieved a quantum computer. But we will pretty much know over the next two years because they are going to be rolling out some more advanced systems late this year and early next year, up to about 1000 qubits next year, which would pretty definitively determine how useful this is going to be. We don’t need to debate whether it’s going to work or not. Either they make it work or they don’t. But then there are ten other different ways to do quantum computers that are trailing behind this, which may end up working as well. So, it’s going to happen, it’s just a matter of when. It could be very soon.
Superthread is something that’s made by Los Alamos. Superthread is made from carbon nanotubes that are 4 centimeters long. It seems to have nearly the full strength of a single-thread carbon nanotube brought up to macroscale. If they could make this in high-volume production, it would be something that would enable the high-strength material predictions that were made in Engines of Creation and other previous nanotechnology work. Zyvex talked about their NanoSolve additive, which makes things four times stronger. You can also tune properties. Ned Seemen‘s work on DNA robotic arm arrays, the work on DNA origami. Basically, I see DNA nanotechnology really getting very powerful over the next 7 or 8 years. The virus assembled batteries by Professor Becker at MIT, where she’s hijacked viruses and used those to produce structures powered with structures of 6 nanometers by 100 nanometers, something like that. Then there’s the combination of several technologies around gene therapy, RNA interference. It seems to me that this will allow pretty much full control of our genome as a growing capability, as well as being able to make synthetic life.
Some other recent work just over the last few days, IBM has this nanogravure printing process where basically you are able to place small objects within 2 nanometers. Currently they’re at larger sizes, but they think they can get down to 2 nanometers. Nanopantography, they split an ion beam into a billion smaller ion beams. It is a way to get highly parallel in terms of your nanostrcuturing creation. Thermochemical nanolithography, you heat up an AFM tip and then you are able to write 10,000 times faster than with dip pen nanolithography. So, instead of .0001 millimeters per second, you go at several millimeters per second. The dimensions scale down to about 12 nanometers in width. Fracture induced structuring, where basically you create certain small fractures, and then you can break apart and get 60 nanometer lines, which is useful for computer chips.
So, you have all this capability and basically I view them as all these little fires. This is how I view convergence. If you have all these little fires of technology developing, and then technological convergence comes along, you have this far greater change. So, it could be good or bad. It could be a bad fire or a good fire. You could get a lot of power. Also, convergence does not necessarily mean that you have to have full-blown molecular manufacturing. Things could converge in a more difficult way, in my opinion, without it. But again, once you get molecular manufacturing, whatever was there just gets bigger. Just one last bit on current technology that I think is underestimated. Labs on a chip and bubble logic. Bubble logic is where you take microscopic bubbles and you run them around on a chip. Labs on a chip is where you have a bunch of microchannels, a bunch of small reservoirs of chemicals, and basically you are able to do the work of a bio lab on a chip. Bubble logic is a way to improve upon that, make it a thousand times faster. It lets you have this systemic capability, which if you were to combine it with some of these other atomic force microscope techniques, along with self-assembly, could get you a kluged capability. Again, convergence. You can start doing some useful work.
It’s not just the fact that you can get this far greater efficiency in your engines and faster computers, but that superconductors can also enable certain other science that you could not have otherwise. I think that it really busts out once you get room temperature stuff, which they’ve got some theoretical stuff around. This I think molecular manufacturing could help enable. Nanomembranes, I think that nanopores have far more power and utility than people realize. One example is that if you thread DNA through nanopores, this is one way to get to very cheap gene sequencing. It might cost $1000 for a genome, instead of $1 million now or $1 billion a few years ago. If you get really good membranes, you could then drop DNA sequencing down to pennies per genome. Lasers have all kinds of capabilities, not just for reading DVDs and CDs, but military and space applications, which I don’t think is commonly appreciated. Lasers could also be used for communications. You could get really high-speed communications, like multiple gigabits per second, even terabits per second. Rapid prototyping and rapid manufacturing of electronics. These are actually industries now, where people are making hundreds of millions of dollars, where they are making parts from machines that cost in the range of $5000 to $100,000, where you can extrude out pieces. It’s not as good as having a factory dedicated to it. But if you are doing small runs, it’s more economic to do it.
Claytronics is an idea from Intel that they are working on to make small units of devices which will be generally spherical, which would be magnetically positioned relative to each other. An end-state goal might be to get it down to about a millimeter in size, where if you had enough of them so that they would equal my weight, if they were cheap enough to make, you could add them up so that they were like a granular version of me working remotely in some other location. It could also be viewed as a reusable version of a rapid prototyping thing, kind of a weak version of a utility fog. For certain things like superconductors or molecular manufacturing, if it turns out that you need conditions where things are a lot cooler, getting cheap cooling could be a very important aspect of that. If you can’t solve the problem by getting things up to room temperature, if you can bring the temperatures down cheaply then that could enable things.
Social change leading up to 2015. A lot of people aren’t aware of how many rich people there are in the world. There are about 7 million millionaires currently. I expect that number to double by 2015. You can just see it in the wealth reports, increasing about 10% per year. There are also studies about how many technology angel investors there are. 234,000 individuals. If the number of rich people doubles, I’m thinking that the number of technology investors will more than double, just because I think the opportunity will be there for them to invest in it. If you project into the future you have to look at the rise of China and India as economic forces. There is also a long-term acceleration of economic growth. Robin Hanson, the economist, looked at historical economics and saw that if you look at the various ages, from primitive man, to hunters, to farmers and so on, there is this constant acceleration of economic growth. You see that now in recent decades with the rapid growth of Japan and then China, these are examples of countries industrializing faster than the United States.
So, more countries, companies, and people are able to fund technology and change. It’s like you have more participants in a big race. That’s why the change can accelerate. You have more people doing more projects, smaller groups able to achieve a disruption. A few Stanford students making Google work and then gathering resources to get more people behind them to effect more change. There are also things happening faster than Moore’s law, and Moore’s law is spreading outside information technology. This is something documented by Ray Kurzweil. Flash memory from Samsung is more than doubling every year, and that’s been happening for several years. It looks like it will happen for several more. System integration, not the main process but all the subcomponent chips are improving faster than Moore’s law. Graphics chips, the NVIDIA general purpose graphical processing unit, they’re able to use that for more than just graphics now. That also has been improving faster than Moore’s law for the Intel AMD chips. The falling pricing of DNA sequencing as well.
These are 2015 projections to summarize some quick points about where I think we’re at, on the edge of molecular manufacturing. 200,000 – 1 million qubit quantum computers. That depends upon D-Wave not being totally full of it. Billions of artificial or simulated neurons. There’s some work right now at Stanford where they are making systems with a million simulated neurons. I think IBM has simulated neurons on their supercomputers. We have a few million now. I’m thinking with better than Moore’s law you can get to a billion by 2015. That is relevant in that 100,000 billion neurons for a human brain, 6 million for a mouse. That kind of gets you up the old artificial intelligence food chain there.
Also, better computer memory. NRAM from Nanotero is supposed to be out this year. I haven’t heard any announcement. If it does come out, it would be a big deal. But there are a bunch of other improved memory options in the race, where basically you can get hard drive memory that is persistent, it stays there, it’s fast and safe. Again, it would help accelerate the computing power situation. Ovonic quantum control devices, which would enable flat computers. Stanley Ovshinsky has a close-to billion dollar company. He helped invent the nickel metal hydride battery, and he has this new thing, quantum control devices, which are a more neuron-like version of a transistor. It would have the advantage of being, one, like a neuron. Two, it would also enable all parts of the computer to be printed reel-to-reel. It would make things cheaper and faster.
Projecting out that flash memory stuff, about 64 terabyte flash drives by that point. Gene therapy and gene doping I think will be making big medical impacts around that time. Maybe also some areas around human enhancement. Superthread and carbon nanotubes will be common at that point. I’ll have a slide on the production side of that. The wireless communication trend I see continuing to improve roughly about 100 megabits per second to 5 gigabits per second. Some legislation will also impact what happens there. Gigapixel cameras. Currently we have common cameras up to about 10-20 megapixels. There are higher-end cameras up to about 100 megapixels. Then you have backplane cameras which take slow shots, but are about a gigapixel now. I’m just saying that you get them cheaper at that point. Ubiquitous computing, just because there are small computers everywhere. Wireless power and communication. I also project that China’s economy passes the United States around 2018 plus or minus three years.
Nanomaterials were talked about by Jim Von Ehr the other day. There are some companies saying they will make a 3,000 ton factory by 2011. Currently production is about 60 tons per year. I think that with that increase in volume that the price will drop. Jim was saying it will be $50 per pound and I think we could even do a little better. After that, if production goes even higher, the 40,000 to 100,000 ton level is about the current level of carbon fiber. Then you get to around that price of about $1-5 per pound. Also, there are the different types of carbon nanotubes, single or multi-wall purity levels. I think the high-end stuff falls even more because they get better at making it.
Graphene, 2-dimensional carbon, could have some pretty big impacts in computers and electronic. There is a lot of money going into it and some interesting lab work. Then there is high-precision parallel AFMs and STMs will bootstrap to some interesting nanotechnology capabilities. The price of DNA sequencing is dropping very fast. Currently we can make about 45,000 base pairs. They’re talking about making up to a million sequences for some artificial bacteria. That’s what Craig Venter and some others are working on. It would pass 3 million base pairs for a ribosome around 2010-12, and then go beyond that, starting to head towards higher life forms, 3 billion base pairs for a human cell.
In the pre-nanofactory world, China is the world’s production factory. This Atlantic article “China Makes, The World Takes” discusses how it’s not just that, it’s about speed. China can respond more quickly. They have these cities where there are hundreds of makers of the same product, and they can quickly scale up and make a prototype. The bulk human version of the industrial nanofactory is what’s there. You can get them to adapt and start making something. That’s something to appreciate in terms of what a nanofactory has to replace. That’s the bulk version, what we currently have instead of a nanofactory.
About 2015, full-blown molecular manufacturing. It will help break the barriers to what’s been holding back a lot of technological change. I focus a lot on different things that could be technological solutions to problems like how to make computers a lot faster. There are certain economic production barriers, like why have silicon computers dominated for all this time? It’s not as if no one has thought of stuff that’s hundreds of times better. The problem is getting the alternatives scaled up fast enough before the doubling every 18 months or so of silicon catches up with it. Or you have hundreds of thousands of people that are skilled in the one technology and cannot go off into this niche thing, which could have a lot of interesting, useful value. They don’t want to take the risk of trying to develop that alternative, as in magnetic bubble memory, spending a bunch of money and getting nothing out of it. So, if you can accelerate from lab to deployment to making money, then a lot of the things that have been held back suddenly start to come into play. This could be a big boost, because you could get computers a million times more powerful than we have now if you had several technologies come out of the lab. We know how to do it, but we don’t do it.
Also, in terms of solving the world’s problems, changing the infrastructure has been hard. Energy, they talk about taking years to make a plant and get all this funding organized. If you bring the cost down and the production up, infrastructure roadblocks could go away. Again, this is dependent upon societal and individual choices. Even if you could do it, people could stop you from doing it. Bad choices in governance, I expect that to persist. You have to fix the world in spite of that.
The rough projection of nanofactory capabilities is a kilogram of production in one hour. That means you can make 4000 tons of nanofactories and 8,000 tons of products in a single day. You could then replace all your current production capabilities in the world in weeks to months. Of course there is more complexity to this, but theoretically. When you talk about what you can make in terms of this world-shaping capability, this is theoretically possible, but you would have to do a lot. Sections of the world infrastructure could be switched out and it would be faster and more flexible than our current production machine.
Once you have the nanofactory that is theoretically possible. You can break down these technological barriers. You have the physical barrier of the wall and then you tear it down, but you still have mental restrictions. Consistently, people make the wrong choices and that prevents the full capability from being developed. You could do something, if you were not tied up and prevented from doing it. Right now we develop coal as a primary power source. It causes air pollution and kill millions of people each year. I view that as a bad choice. But the incentive feedback structure is such that even though the coal companies in the U.S. is a $50 billion industry, they don’t have to pay for making people sick. So the incentive feedback structure is, just give us the cheap kilowatt. I actually think that is a fundamental flaw of societal policy when people talk about Medicare, but 30% of the costs are coming from air pollution problems.
An example of a past technology that has been underdeveloped at only a small fraction of its potential is nuclear power and propulsion. In my view, people have been unnecessarily fearful of this thing for somewhat justifiable reasons of the fact that you have nuclear weapons. However, if you dig down into this a bit you can see that even if you have less of the good uses, you have not reduced the bad uses. We still have thousands of nuclear weapons. The fact that you have fewer or more nuclear reactors, or the fact that you did not develop nuclear propulsion, did not change that. People do not understand the details of it, which would require them to study it and say, what is the fuel cycle? What are the different types of nuclear reactors? And to figure this all out instead of going with a knee jerk reaction. The same thing for nanotechnology: to really understand and make good choices you have to study it, but the policy makers and everyone else have no time. So you end up with these half-assed choices. If we had not stopped making nuclear reactors, we could have gone the French road and gone up to 80% nuclear power instead of the current about 20% for electrical power. That would have saved 20,000 lives per year. Twenty times more than we lose right now in the Iraq War. Every three years, it’s a little Hiroshima. People are concerned about Chernobyl, a kind of reactor that there were only eight of in Russia and not anywhere else. It didn’t have the proper containment dome. And Three Mile Island, no one died. That’s a matter of people dying versus the fear of people dying, something I talk about a lot in my website.
Any delays or production limitations for molecular manufacturing? We don’t know. It’s a prospective technology. It could take longer than 2025, although I think that this pseudo-molecular manufacturing from these other klugey ways could get a lot of the capabilities that we have projected out. Having carbon nanotubes out in bulk production is not dependent upon molecular manufacturing happening. A lot of what was written about in Engines of Creation could happen without it. Use the advanced nanometer manipulation systems, only use molecular manufacturing for critical components, if you have a limited capability. I’m not sure how you could have a limited capability that you could not then use to bootstrap into a really good capability. It seems like if you make a little bit of stuff, you could make assemblers, and that gets you to full-blown molecular manufacturing. That’s where it seems like it should be inevitable to me.
We need to maintain momentum once you overcome the current challenges. Once we get to molecular manufacturing, I still see a lot more that needs to happen. You know Maslow’s Hierarchy of Needs? I feel that there’s a civilizational analogue to that. People talk about utopia being possible and that kind of stuff. We want to get as close as we can. In terms of applying molecular manufacturing, I would look at the best non-molecular manufacturing plans and try to enable and enhance them. For transportation, there’s this thing called dual mode transportation, where you have a mix of car and rail. You have cars that follow guide rails to get electrical power, and also because of the linear induction, you can keep them at a very close spacing. You can increase the traffic from one lane on a highway ten times, and you can also increase the efficiency because you have the air flow of one car closely following another. Also, you can get rid of accidents, because if they are all having linear induction then they can’t run into each other. They are all forced to go at certain speeds and keep to particular spaces. You cannot do this now because it would cost a trillion dollars, plus you have to swap out and rebuild your road infrastructure. But with molecular manufacturing you could start considering that. A step before that would be just getting out the robotic driving capability to take over cruise control, getting rid of accidents and allowing for the close platooning of cars. If you get rid of the need to travel with the use of Halo room virtual meetings, it’s currently helping to save money for executives and business people from flying on a plane to go meet someone. It currently costs about $300,000. If you can bring that down to $1000 or less, every person could work from home and that makes moving less polluting.
Solar power in space and space colonization. I think the technology is getting close on that. Mirror and laser arrays, you may not be familiar with the work by Robert Forward, where you could have a laser-powered sail. You could get up to a significant fraction of the speed of light if you had a really big laser. You could then push the sail over to another solar system. You can get around the problem of the really big laser, which we don’t have, by making an array of smaller lasers. Then you could reduce the power requirements if you could reflect the lasers between mirrors, which they have done in some lab work. There are ways to get up to a “science fiction” level capability of space travel using molecular manufacturing. The bottom line is that molecular manufacturing makes space travel super easy.
Food production, there’s some stuff around making stem cell factories, where basically you immortalize your stem cells and grow your own meat. This could be a cheaper, more environmentally benign way of getting rid of agriculture that messes up the environment. You could drill down on each one of the aspects of your economy to apply and accelerate technologies, while getting rid of the need for things that are not useful. You could use new technologies to substitute for social security. If you could provide free and guarenteed access to a nanofactory, there would not be the need to save up the money for social security or retirement. You could provide for people far better than with our current subsistence level healthcare in terms of food, clothing, housing, and so on. You could then get rid of certain aspects of financial services by having a technology replacement for a social need. These things would not cost more, in terms of fraction of the economy, as welfare or international aid. There should not be any tolerance for having failed states, failed regions, failed individuals, because they end up being disgruntled and maybe try blowing things up.
Productivity growth, depending on how clever and bold we are,we can collectively increase GDP growth to 20% to 50% per year as full-blown molecular manufacturing kicks in. This figure is somewhat supported by the economist Robin Hansen, who did an equation of the basic economic equation and said, What if you had AI that was as good as people? Then that changes the whole labor component to get your GDP growth rate calculated out. Just allowing for AI you get from 4.3% to 45% growth. In 20 years at the 50% annual rate, we would be past Kardashev level 1. If you can do that for 60 years, you hit to K2, where you are using all the solar power of the sun. It would mean big things if you actually could do it.
You have to restructure for hypergrowth. You have to have a system that is enabling faster growth. I think there will be the need for really good simulations. To simulate the effects of the introduction of new technology, you need to have sensors to know what the current state of the world is. Molecular manufacturing I think means improved superconductors and enabling wave after wave of new technologies. Nuclear fusion, there are several ideas now. Z-Pinch is a magnetic way to get fusion they are using at Sandia National Labs to model atomic bombs. It can also develop into a fusion power source. They have something now where they can repeatedly fire it off every few minutes. If you develop that technology over the next 20-30 years, they have a roadmap for getting to fusion. Molecular manufacturing could help that technology, too. If you have a lot of energy you can do a lot of things, like solve water problems. Fission can be done a lot better with molecular manufacturing mass production. Solar cell production would be another of the capabilities. Health, life extension, enhancing transhuman capabilities.
When you are running a business and thinking about technological change, you want to shift more of your focus to game-changing. You want to leapfrog and have aggressive goals. Before I talked about a marathon, where you have a lot of people running. There can also be a marathon where the conditions are changing, so they are not just running in clear weather but in scorching heat, in rain, snow, over obstacles. You have to have something that is robust and adaptable to different situations in order to prevent the existential and other risks from happening. I’m for moving to the world of molecular manufacturing, space colonization, and that kind of thing. I think that the current status quo of pollution, war, disease, famine, etc. is not something to solve before we do anything else. We can move past it. Maybe the problems don’t go away, but what would have been a fatal stab to the heart becomes this little pinprick. Because your civilization is so big and robust, you can handle it.