An Exponentially Expanding Future From Exponentially Shrinking Technology
5:30-6:00 p.m. Reception
6:00-7:15 p.m. Meeting
Council on Foreign Relations
New York, NY
RICHARD FOSTER: Good evening. Welcome to the Council on Foreign Relations. My name is Dick Foster, and I have the pleasure of hosting Dr. Ray Kurzweil tonight.
Before we get started, I’d like to ask you to take a moment to turn your cell phones off, your BlackBerries and any other wireless devices that you have.
We do, with close cooperation with the Department of Defense, have monitoring devices. We know if you have it on, and we will deal with you in an appropriate way.
I’d would also like to remind you that tonight’s meeting is on the record. Anything that you can say will and probably should be used against you. So when you get to the questions, we’ll have that.
I will also say that tonight we’re going to go 15 minutes longer than the normal session. This session will go to 7:15 to allow you all to get your questions in to our very interesting guest.
This lecture tonight with Dr. Kurzweil is a part of the CFR’s ongoing series on science and foreign policy. The purpose of this initiative is to engage council members, all of you, in talking about the foreign policy implications of changes in science and technology.
The series aims to draw attention to those critical areas where science and foreign policy intersect and explore the implications of emerging technologies on trade, security, U.S. competitiveness and global health. And we will have a direct opportunity to do that tonight.
Ray Kurzweil has been described by The Wall Street Journal as a restless genius and by Forbes as the ultimate thinking machine. Forbes also called him the rightful heir to Thomas Edison. PBS included Ray as one of 16 revolutionaries over the past 200 years that have affected the culture and direction of the United States.
As one of the leading inventors of our time, Ray was the principal developer of the first omnifont optical character recognition system; the first print-to-speech reading machine for the blind; the first charge-coupled flatbed scanner device; the first text-to-speech synthesizer; the first music synthesizer — a name you may associate with his name — capable of recreating the grand piano and other orchestral instruments; and the first commercially marketed, large-vocabulary speech recognition system.
Among Ray’s many honors, he is the recipient of a half million dollar Lemelson Prize for innovation at MIT. It’s the world’s largest prize for innovation. And in 1999, he received the National Medal of Technology, the nation’s highest honor in technology, from President Clinton in a White House ceremony. In 2002, he was inducted into the National Inventors Hall of Fame, a hall of fame established by the U.S. Patent Office for our most distinguished inventors. He’s received 12 honorary doctorates, as well as honors from three U.S. presidents. He’s written five books, four of which have been national best sellers. “The Age of Spiritual Machines” has been translated into nine languages and was the number one best selling book on Amazon in science. Ray’s latest book, which he’s going to discuss with us tonight, “The Singularity Is Near,” is now in its fourth printing after two months and has been number one on Amazon in both and science and philosophy.
RAY KURZWEIL: Well, thanks. It’s a pleasure to be here. I was here about six years ago when I talked about my last book. And I’ll try to comment on implications for foreign policy, the basic message being that we are achieving a deeply interconnected world. Increasingly commerce dialogue, political processes transcend national boundaries. And I’ll talk more about that.
The key message that I want to share in the sort of brief time I have to give an introduction before we have a dialogue is that information technology in all of its manifestations is accelerating. It’s not growing linearly. It’s basically doubling its power. It’s measured in price performance capacity, and with every year, which is pretty phenomenal growth. I mean, it’s already deeply influential, and it’s going to expand its capabilities by a factor of a billion in 25 years, which is quite daunting if you think about how influential it is already, how it’s already transforming business models and deeply affecting the world.
The face of change is accelerating. I mean, we didn’t hear the word blog more than two years ago. We didn’t use search engines five years ago and so on.
And I’ve been tracking technology trends for 30 years because of my interest in being an inventor. I realized that my technologies had to make sense when I finished a project, and invariably the world was a different place two or three years later.
So I began to build mathematical models of how technology evolves. This has taken on a life of its own. I now have a team of 10 people that gathers key measures of technology in different areas and projects how it will progress as we go forward.
And with these models, we are able to anticipate not just two or three years out, but 10 years out, 20 years out.
That brings up an interesting issue. Can we tell the future? The common wisdom is that we cannot, but as I’ll show you — and I’ll give you a few brief examples among hundreds that we have — of just how predictable the progression of information technology in many different areas is.
And you might wonder, how could that be? I mean, how could we reliably predict what’s going to happen in the future? And, in fact, we cannot predict reliably what will happen to a specific project. If you asked me will Google’s stock be higher or lower than it is today three years from now that’s difficult to say. What will the next wireless standard be? Will it be Ymax, CDMA G-3? That’s hard to say. But if you ask me how much will it cost to buy a (mix?) of computing in 2010 or to sequence a base pair of DNA in 2012 or the spatial and temporal resolution of brain scanning in 2014, I can give you a figure, and it’s likely to be quite accurate.
And I say this not just backdating to past data, then making these forward-looking predictions for 25 years and it’s been quite accurate, like the emergence of a worldwide communication network emerging in the mid 1990s, which I made 10 years earlier because I saw the ARPANET doubling in size every year. But it was only used by a few thousand scientists. When you double from 10,000 to 20,000 nobody notices it. But it was clear that there would be 10 million to 20 million to 40 million 10 years later and would be a worldwide phenomenon.
How can this be when each specific project is unreliable? We see other examples of that in science.
For example, thermodynamics — the path of any specific particle is completely unpredictable, yet the overall system has very predictable properties according to the laws of thermodynamics to a very high degree of precision, even though every one of the particles is random and chaotic.
And the evolution of technology, particularly when we can measure it in precise information terms, is similarly a complex, chaotic system, where each specific project, each individual, is very unpredictable. Yet, the overall flow of these technologies is remarkably predictable.
Despite the fact that we have wars and recessions and IPOs and countries dumping products on each other and bankruptcies and so on, as you’ll see, the results are remarkably predictable.
And the other important point is that information technology is not just little devices we put in our pocket and iPods and so on. It’s deeply influencing everything that’s important. It already underlies a very substantial fraction of the world economy. It will be a majority of the economy by the 2020s.
We’re understanding our biology as information processes. That’s quite a different — that’s quite a paradigm shift from the old paradigm of biology, which was hit or miss. We just find something that happened to work — oh, here’s something that lowers blood pressure; we have no idea how it works; there is no theory of operation. Now we’re actually able to understand the information processes underlying these disease processes — like cancer, atherosclerosis, and (then ?) intervene very selectively. We’re getting the tools to reprogram our biology.
I mean, how much software do you use that you haven’t changed in 30 months? We have these 23,000 software programs called genes inside us we haven’t changed in 30,000 years. And there’s some genes we’d like to change — the fat-insulin recepter gene basically says hold on to every calorie because the next hunting season may not work out so well.That was a very good strategy 30,000 years ago. (Laughter.) And we’d like to change that strategy now. It underlines an epidemic of obesity. What if we turn that gene off? Well, that was done with mice — these fat insulin-receptor knockout experiments. These mice then ate ravenously and remained slim and they got the health benefits of being slim. They didn’t get diabetes. They didn’t get heart disease. They lived 20 percent longer. They got the benefits of caloric restriction while doing the opposite. (Five?) pharmaceutical companies have noticed that might be a good market for the — product for the human market, and they’re rushing to bring fat-insulin receptor inhibitors to the human market.
There are a lot of other genes they’d like to turn off that encourage, or that are required, for cancer, insulin resistance underlying diabetes and so on to progress. And we’re gaining the tools to turn genes off. There are new means of actually precisely turning genes on or adding new genes, turning off enzymes. (Inaudible) — for example, targets one specific enzyme needed and one specific stage of atherosclerosis, and if you turn that off it prevents atherosclerosis from progressing. At least that’s what appeared to be the case in the phase-two trials. Pfizer’s spending a record $1 billion on the phase-three FDA trials.
And that’s just one example of thousands of this rational drug design — understanding and reprogramming the information processes underlying biology with the recognition that biology is basically an information process.
And then the most important phenomenon in the world — intelligence, as represented by human intelligence, which is the best example we have of an intelligent process — we are — those are also information processes. We’re also making exponential gains in understanding that.
So let me take you quickly through some of these trends and show you how predictable they are. And then we can use a little bit of imagination to contemplate some scenarios of what the world would be like based on the evolution of information technology. And if anything, the future will be more remarkable than what we can anticipate today because we’ll have all of you and your colleagues and millions of other people applying their creativity to these exponentially expanding powers of information technology.
Probably the most important trend is that the rate of progress itself is not a constant. Hundreds of years ago, people thought that there was no change. And then we realized that change was pervasive, and so it became an axiom that change is a constant.
But in fact, change is accelerating. The whole 20th century was not a hundred years of progress at today’s rate of progress. We’ve been speeding up to this rate. According to my models, it was 20 years of progress at today’s rate. We’ll make another 20 years of progress at today’s rate in 14 years. We’ll do that again in seven years. Because of the sort of explosive nature of exponential progressions, we’ll make 20,000 years of progress in the 21st century — a thousand times more than the 20th century.
There’s many examples of this, but these are all logarithmic graphs, meaning if you go up the graph, it represents multiplying some key feature by the powers of 10.
So this is the adoption of the telephone. It took 50 years to adopt this (as a ?) communications technology. It was actually the first virtual reality technology. You could be with someone else, at least as far as one human sense is concerned, even though you were hundreds of miles apart.
Recent technologies, like the cell phone, did the same thing in seven years. And if we look at different communication technologies — television, radio, telephone — it took decades to be adopted by a quarter of the U.S. population. The web, cell phones, PCs did that in a matter of years. We have actually better than exponential growth — a straight line on a logarithmic graph is exponential growth.
This is an interesting graph. This is a double logarithmic graph — on the X axis how many years ago the event took place in powers of 10; and on the Y axis how long it took for that paradigm shift to occur until the next paradigm shift. And you have both biological evolution and technologic evolution.
And so the first paradigm shift — the evolution of an information backbone to biology, DNA — actually, RNA came first — took billions of years. But evolution works through indirection. It creates a capability and then it uses that capability to bring on the next stage. And that’s why an evolutionary process inherently accelerates.
So the next stage — the Cambrian explosion, when all the body — (inaudible) — of the animals evolved, went a hundred times faster. It took 10 million years.
In biological evolutions (it kept?) accelerating. Homo sapiens evolved in only a few hundred thousand years. There were only three small genetic changes that distinguish us from other — our primate ancestor. One was a larger skull at the expense of a weaker jaw, so don’t get into a biting contest with another primate. More of our brain was devoted to the cerebral cortex so we could do more abstract reasoning. And the pivot point of the thumb was moved up one inch. If you look at a chimpanzee’s hand, it looks similar to our hand, but it doesn’t work as well. They don’t have a power grip. They don’t have fine motor coordination. They really can’t manipulate the environment.
So we have this abstract reasoning ability to imagine changes and then we could actually carry them out. And those — that actually only comprises a few tens of thousands of bytes of genetic information.
And now — these were the enabling changes that brought on the next stage, which is technological evolution. And that, again, went a little bit faster. The first stages — fire, stone tools, the wheel — took tens of thousands of years.
And then we always use the latest stage of technology to create the next stage. Half a millenium ago, the printing press took a century to be adopted. Half a century ago, the first computers were actually designed pen on paper and wired with screwdrivers. Now we can design a new computer system in a matter of weeks with computer system design software working out twelve layers of intermediate design.
And you see that this is a straight line showing this continual acceleration of both biological and technological evolution, with technological evolution emerging from the biological evolution that created the first technology-creating species.
And some people said, okay, Kurzweil only put points on his graph to fit on the straight line. So I took 14 different lists — Carl Sagan’s cosmic calendar, Encyclopedia Britannica, American Museum of Natural History — 14 different thinkers and reference works. And these were people not trying to make my point or disprove it. This is just what they thought the key events were in biological and technological evolution.You do see some spreading of the points, but you see a very pervasive and clear trend line showing this acceleration. If you look at this on a linear graph where the X axis is linear, it looks like everything has just happened. That’s the nature of exponential growth.
Now, if we compare — this is a linear graph, but it’s showing you an exponential trend and the linear one. And this is really a key point. Most people intuitively assume that the current pace of progress will continue at the current rate; we’ll have the current tools. This is very pervasive, intuitive view of otherwise sophisticated thinkers.
I had a debate with someone doing brain reverse engineering recently who was just thinking linearly, okay, it’s going to take me 18 months to finish modeling this one ion channel and there’s five other ion channels. So that’s five times 18. And then there’s these other details and this other dendrite and six other ion channels. He’s adding it all up and says, oh, it will be a century before we reverse engineer the human brain without taking into consideration that it’s going to accelerate.
The genome project was similarly controversial. Mainstream critics said, there’s no way you’re going to sequence the genome in 15 years. We just had our most advanced equipments used by — (inaudible) — students around the world and in 1989 we succeeded in sequencing one-ten-thousandth of a genome. It’s going to take, you know, hundreds of years to finish this project. But the amount of genetic data that we sequenced doubled every year, actually very smoothly, as I’ll show you. And the project was completed on time.
Ten years later — 10 years after we started the project, it still looked controversial because we had only finished 2 percent of the project. But it’s the last seven doublings that get you from one percent to 100 percent.
Now, information technologies are doubling their power every year, which is pretty phenomenal. It’s 50 percent deflation, which has very substantial economic impact.
A personal experience — when I was at MIT, a computer that took up twice the size of this room (was about? ) a thousand times less powerful than the computer in your cell phone today.
If we look at computers going back a century to the computing technology used in the 1890 census, then the relay-based computers that Alan Turning cracked the German enigma code, then vacuum-tube computers — CBS predicted the election of Eisenhower. The first time the networks did that in 1952 using vacuum tubes.
They were then shrinking vacuum tubes, making them smaller and smaller. People say, well, exponential can’t go on forever. They must hit a wall like rabbits in Australia. They eat up all the vegetation and then the exponential growth stops. Well, the exponential growth of price performance of computing due to shrinking vacuum tubes stopped. And that was the end of shrinking vacuum tubes, but it was not the end of the exponential growth of computer price performance. It led to another paradigm.
Every time one paradigm runs out of steam, it creates research pressure to create the next one. So then we had transistors used in the first NASA space flights, and we’ve had 40 years of — (inaudible) — shrinking transistors on an integrated circuit.
And we have this very smooth exponential progression going back a century. And you might notice that’s not a straight line. I mentioned a straight line on a logarithmic graph is exponential growth. We have exponential growth on the rate of exponential growth. It took us three years to double the price performance of computing in 1900, two years in the middle of the century. We’re now doubling it every one year.
And same thing with super computers — I mean, any type of computer or any type of information technology, every type of electronics, is progressing at this rate. Super computers will hit the 10 to the 16th calculations per second — 10,000 trillion CPS — that I estimate is necessary to emulate the whole human brain.
Actually, when my book came out — I mean, generally my predictions, even though they’re considered radically optimistic, turn out to be pessimistic because I’m purposely conservative. I’ve projected here 2013 to achieve that milestone in the super computer. Just a month ago, Japan announced two super computer projects to achieve that level by 2010.
But you see smooth exponential progression. Processor performance — this is the price of a transistor. You could buy one transistor for a dollar in 1968. You could buy 10 million in 2002. When I was a high school student, I used to hang out around the surplus electronics shops on Canal Street and buy something about this big, which was equivalent to about one transistor relay with support circuitry, a million times slower, for about $40.
But look how smoooth this progression is. It looks like it’s the output of some tabletop experiment. But this is the measure of millions of people’s activities — designers and engineers and marketing people anc competing marketing programs and through recessions and boom times. You have this very smooth, very predictable, progression, despite the fact that the activity of all these companies involved is highly unpredictable.
And we see that time and time again as we measure these overall results of information technology in a broad variety of fields. As we make the transistors cheaper, they’re actually better, because they’re smaller. The electrons have less distance to travel, and we have exponential growth in the speed of transistors.
If we put those factors together, the cost of a transistor cycle has been coming down by half every 1.1 year, and that’s — and if you add other levels of innovation, we get actually a doubling of price-performance in every type of information technology every year.
That is 50 percent deflation, and depending on what week it is, the economists will worry about inflation nor deflation. They’ll say deflation’s just as bad; it can lead to a shrinking of the economy. We had deflation and the Depression. That was a different phenomena, though. That was a collapse of consumer confidence, a collapse of the money supply. This is due to increasing price-performance, increasing productivity.
But they’ll say it’s still a bad thing because it is going to shrink the economy. I mean people aren’t going to buy twice as much capability each year and keep up with this doubling of price-performance. So they will buy a little bit more, but the overall economy, especially at least in dollars, will shrink.
But that is not what we’ve seen. The actual doubling of the consumption of electronics — and there’s two of every type of information technology — has more than doubled every year. We have 18 percent per year growth for the last 50 years in the consumption of information technology in constant dollars, despite the fact that you can get twice as much capability each year.
And this is very pervasive. I just — (inaudible) — it’s a different technology problem — different engineers, different companies, same progression. It’s an inherent feature of this evolutionary process of technology evolution.
And I mentioned the — we’re now understanding our biology. The old method of drug development, it was called drug discovery, which literally was that, just discovering something that happened to work, but since they didn’t really have a model of how it worked, these were and are crude tools, and 99 percent of the drugs on the market today were done this way.
The new paradigm is to really very selectively, you know, intervene with one specific process and understand and model of these information processes underlying biology. And biology is an information process. Genes are an encoding of information.
DNA sequencing came down from $10 in 1990 to 2 cents in 2004. It’s now about a penny. The amount of genetic data — this is logarithmic graph — and this slope represents doubling every year the amount of genetic data.
It took us 15 years to sequence HIV. We sequenced SARS in 31 days. It was only two years that we finished the genome project. We now already have the HapMap, where we have sequenced all the genetic variability among humans, and we are now sequencing and understanding the genetic basis of different diseases and gaining the tools to really reprogram our biology.
Communication technology — this is one that has particular foreign policy implications, because we have a real interconnected world. We didn’t, you know, abolish customs officers and national boundaries, but we have this new economy that just kind of got layered on top of everything else that really ignores national boundaries, and it’s a very substantial portion of the economy. E-commerce, just measured in this country, is already a trillion dollars. And this — it’s really one integrated world economy. And that part of the economy is going to grow.
And many different ways of measuring this. I don’t want to dwell on this, since I don’t have a lot of time. But this is this graph I mentioned that I saw in the mid-1980s. It was called the ARPANET — then the Advanced Research Project Agency. Only a few thousand scientists were using it, but it was doubling every year. It’s clear to me that doubling every year was going to multiply by a thousand. So instead of being 10,000 scientists it would be 10 million people and then 20 million the next year and then 40 million; it would be a worldwide phenomenon.
So I made this projection in the mid-1980s in my first book, “The Age of Intelligent Machines,” — this is what the same data looks like on a linear graph. And we live in a linear world. This is how we experience the world. So it looked like the Internet came out of nowhere in the mid-1990s. But you could see it coming if you looked at the exponential progression on a logarithmic graph.
But then people in the 1990s said, well, you know, Kurzweil made this radical prediction in 1985, but I guess it’s wrong, I mean, because in 1990, nothing was happening. You know, when you’re doubling these small numbers, nobody notices it. But then when you reach the — (inaudible) — of the curve, there is sort of an explosive phenomena. And that’s really where we are at in terms of the impact of these technologies in many different arenas.
Miniaturization is another exponential trend, both electronic and mechanical. These are some illustrations — (inaudible) — 1986 book, which have been simulated, and some of them have been built. I talk about in “The Singularity is Near” many have many examples now where we can actually now build things at the molecular level. One scientist just built a little robot that walks with a human-like gait built at a molecular level.
The real sort of killer app of this — of nanotechnology will be small devices we can place inside the bloodstream that will keep us healthy from the inside. If that sounds very futuristic, I’d point out that we have already demonstrated that in animals with a lot of different applications.
One scientist cured Type I diabetes in rats with an nano-engineered device — 7 nanometer pores lets insulin out in a controlled fashion, blocks antibodies, because Type I diabetes is an autoimmune disease.
And this is today. If we contemplate the kind of trends I’m talking about multiplying over the next 25 years the power of electronics, communications, by a factor of a billion — (inaudible) technology by a factor of over a hundred, a (3-D ?) volume every decade. These devices will be very sophisticated in the 2020s.
One scientist has already designed a robotic red blood cell that basically does what our red blood cells do. We have reverse engineered red blood cells.
They’re — and it does actually show another key observation about biology, which is, while it’s intricate, it’s also very sub-optimal compared to what we can engineer. And these devices are actually 1,000 times more effective than our biological red blood cells. And analysis shows if you replace 10 percent of your red blood cells with these robotic respirocytes you could do an olympic sprint for 15 minutes without taking a breath or sit at the bottom of your pool for four hours. Honey, I’m in the pool will take on a whole new meaning. (Laughter.)
These designer robotic white blood cells, I can download software from the Internet to combat specific pathogens. If that sounds very futuristic, I’d point out that we already have devices that we can download software into. There’s, for example, neural implants that have replaced portions of the brain that are diseased. There is an FDA-approved neural implant for Parkinson’s that replaces the biological neurons. The biological neurons in the vicinity get signals from the computer. They are perfectly happy to get signals from the computer, whereas they used to be getting signals from the biological neurons. And this hybrid of biological-non-biological intelligence works just fine. And the latest generation of this FDA-approved neural implant allows you to download new software to your neural implant from outside the patient.
So we have today neural implants and other devices that are placed inside the body that can download software from outside the patient.
We have already robotic devices that are blood-cell sized that are at least being experimented with in animals. If you apply, you know, these very predictable exponential trends in hardware and software, communications technologies, to what’s already feasible today, this will be very pervasive and influential technology in really extending human capability as we go forward.
If we could expand this exponential progression of computing to the 21st century, $1,000 in computation will equal the human brain by 2020, at least as far as the hardware’s concerned.
I said that in 1999 in “The Age of Spiritual Machines.” It was a controversial notion then. It’s really a mainstream version — a mainstream position today. You can read Intel’s roadmap or the ITRS roadmap for the semiconductor industry, and they are projecting chips by 2020 that will equal the capability of the human brain with five nanometer features — that’s the width of 25 carbon atoms. And this is part of the semiconductor’s roadmap — semiconductor industry’s roadmap — which has been followed very closely for the last 25 years.
But if it’s a mainstream view that we will have the hardware to emulate human intelligence, it’s not yet a mainstream view that we will have the software. But I make the argument that we will, because another grand project that we are in the early stages of, sort of comparable to where the genome project was earlier in its progression, is understanding the human brain itself.It’s not hidden from us. And we’re making exponential gains in brain scanning. We’re doubling the spatial resolution of brain scanning every year.
It was only recently that we could actually see inside the brain with sufficient resolution; fMRI can only see clusters of neurons. There’s a new scanning technology, for example, from the University of Pennsylvania, that can see for the first time individual interneuronal connections and seeing them signal in real time. And we’re getting the data to actually see how the brain creates our thoughts and how our thoughts create our brain.
But then it brings up a question, okay, we can get this data, but can we make any sense of it? Maybe it is just too complex for us to understand. Doug Hofstadter uses that, well, maybe our brain is just below that threshold needed to understand our brain. And if we were smarter and able to understand it, then necessarily our brain would be that much more complicated and we’d kind of never catch up with it. Maybe that is an inherent property of complex systems. They can be so complex as to understand their own complexity.
Turns out, that’s not the case. In the regions that we’ve gotten data — for example, 15 regions of the auditory cortex — there are models of simulations running on software that perform very well. We can apply sophisticated psychoacoustic tests in simulation and get very similar results when we apply the same tests to human auditory perception.
There’s a similar simulation of the cerebellum, which is where we do our skill formation. This comprises more than half the neurons in the brain. And the simulation works quite well.
We are gathering more and more data, and these simulations are scaling up. It’s a conservative projection to say we will have detailed models and simulations of all several hundred regions of the brain by the 2020s.
And it brings up another interesting issue, which is, how complicated is the brain? Well, if you take a mature brain, and you really analyze and modeling of all the nonlinearities and all the trillions of ion channels and dendrites and so on, it looks very complicated. I estimate it’s thousands of trillions of bytes to capture this data of one human brain.
But the design of the brain is a billion times simpler, and we can see that because the design of the brain is in the genome, and the genome has the design of the human body and the brain. How much data is in the genome? Actually, not that much. There is 800 million bytes uncompressed. It’s replete with redundancies. One sequence called ALU is repeated 300,000 times. If you take out the redundancy with lossless compression, you get 30 to 100 million bytes. I make that analysis in the book. That’s less that Microsoft Word. And that is a billion times less than the apparent complexity of the brain.
Now, you might say, how could that be? I mean how could something in 30 to 100 billion bytes, which is a small fraction of a CD, capture the complexity of a brain, which is a billion times more complicated?
Well, we see that all the time in computer science. We can take, for example, genetic algorithms, where we simulate evolution. We start with a simple solution. We have it actually evolve in a simulated evolutionary environment. And it actually evolves in solution by interacting with a complex environment that is millions of times more complicated than itself.
And that’s actually exactly how the genome relates to the brain.
With regard to the cerebellum, there’s actually only a few tens of thousands of bytes that describe the wiring of the cerebellum that comprises half of the neurons in the brain. It basically says there’s these four different types of neurons. They’re wired like this in one cell; now repeat 10 billion times; add a little bit of random variation within the — (inaudible) — with each repetition. And that’s a summary of what the genome says about the cerebellum.
And you have this largely randomly wired cerebellum that has the ability to self organize in respond to a complex environment. So if a child grows up, he or she learns how to walk, to talk, to catch a fly ball, and it gets filled up with meaningful information. But the design is actually relatively simple.
Now, I’m not saying the design of the brain is simple. But I’m saying it’s a level of complexity that is less than it appears by looking at the — an actual mature brain, and it is a level of complexity that is manageable with today’s technology.
And all this is driving economic gains. GDP is growing exponentially, even on a per-capita basis. Private manufacturing — well, manufacturing with the value of an hour of labor has gone from $30 to $130 in the last 45 years. It’s been a very smooth progression.
KURZWEIL: — in artificial intelligence in the 1980s. It’s probably happening now in nanotechnology and even happened with the railroads in the 19th century — a boom and a bust. But the railroads were ultimately a true revolution.
And information technology is growing. It’s a share of the economy. It will be a majority of the economy as we reach the 2020s. At that point it really will be a deeply interconnected world. There will be one world economy. This idea of trying to stop outsourcing is like trying to sweep back the ocean.
And let me show you briefly one technology that we put together. We created the first — (inaudible) — speech recognition, the first speech synthesis. We put modern contemporary versions of those together with language translation.
Language translation has actually come a long way if you can use these Rosetta Stone texts. We have the same texts in two different languages and actually use pattern recognition, which is my field of study, to find the the patterns.
I was at Google a couple of weeks ago, and they actually created an English to Farsi and Farsi to English translator when nobody on the team spoke a word of Farsi. But the pattern recognition system were able to track relevant, you know, translation rules. And that system actually compared equally to human translators.
So this will be — I’ve actually used this system to converse with people in Europe. I speak English; they hear me in German; they speak German; I hear them in English. And this will be a routine feature of your cell phone no later than the next decade.
(Off-mike) — This is a demonstration, comma —
TRANSLATING TELEPHONE: Dies ist eine Demonstration.
KURZWEIL: — of a prototype of a, quote, translating, end quote, telephone, period.
TRANSLATING TELEPHONE: Von einem Prototyp eines ubersetzenden Telefon.
KURZWEIL: Within a few years, comma —
TRANSLATING TELEPHONE: Innerhalb einige Jahre —
KURZWEIL: — we will be able to talk to anyone, —
TRANSLATING TELEPHONE: Wir werden fahig (sein ?) fur jemandem zu reden.
KURZWEIL: — regardless of their language.
TRANSLATING TELEPHONE: Ohne Rucksicht auf ihrer Sprache.
KURZWEIL: The rain in Spain, comma —
TRANSLATING TELEPHONE: (French spoken.)
KURZWEIL: — stays mainly in the plain, period.
TRANSLATING TELEPHONE: (French spoken.)
KURZWEIL: (French spoken) — period.
TRANSLATING TELEPHONE: Thank you for your attention. (Laughter.)
KURZWEIL: So there is — that is actually synthetic speech, even though it sounds recorded.
So let’s put a few scenarios together, and then we’ll have some dialogue about this. But I’ve tried to make the point that this progression — this exponential progression — of information technology is quite inexorable; it’s quite predictable. I’ve made — actually, “The Age of Intelligent Machines,” which I wrote in mid-1980s, had hundreds of predictions about the 1990s and early 2000 years which tracked very accurately.
And it’s very pervasive. It’s not just computer devices. It’s not just i-Pods. As new — as price performance reaches certain levels, new applications open up. We didn’t buy i-Pods for $10,000 10 years ago. So we’re constantly creating new opportunities.
And we’re going to make very revolutionary gains in understanding our biology and reprogramming it to really overcome disease. I really think we will overcome cancer and heart disease and diabetes and all the diseases that kill 95 percent of us — the degenerate diseases — over the next 15 years.
And it’s very pervasive in its influence. And particularly capturing non-biological intelligence will be very formidable.
By 2010 computers will start to disappear. They won’t be these small objects that we put in our pockets. They will be in our clothing.
There’s this basic dilemma. We want to make the devices smaller and smaller. But we also like to have very large screens. People like, you know, large high-resolution, high-definition screens, but you can’t put that in a tiny little device. And people have been complaining that the Nano i-Pod is so small they lose it.
So the answer is actually to create a display that’s really tiny but projects images into your retina and create a sort of a full immersion virtual reality environment that will be, you know, very large, as large as the world.
And this technology exists. I’m on the Army science advisory board; I advise the Army scientific technology. And this — they have these types of technologies to put soldiers in virtual-reality environments, so they can take the soldiers out of the weapon, which is very often not a safe place to be. So the armed Predator is an early harbinger of this trend. Even if you’re inside a weapon like an Abrams tank, which actually is a safe place to be, probably safer than walking around New York. There’s only been three combat casualties in 20 years inside an Abrams tank.But they don’t want the soldiers just looking outside the window to see what is going on, so they put the soldier in a virtual reality environment.
So these are extensive today, but these are early adoption applications. Surgeons will use them so they can do surgery on, let’s say, the eye. And they’ll be in a virtual reality environment where the eye is this big, and they can make, then, precise movements, and then translate into even finer movements by a robotic surgeon.
But these will be ubiquitous technologies early in the next decade. We’ll be online all the time. The electronics will be woven in our clothing. We’ll be interacting with virtual personalities that will have these (sort of ?) pop-up displays. As we look at people — you look at someone who gives you information about them, like remind you their name. That — even that would be very helpful. (Laughter.)
If we go to 2029, it’s really where these technologies will be quite dramatic. Twenty-five years from now these technologies will be fully a billion times more capable than they are today. We’ll have reverse engineered the human brain. We’ll have non-biological systems that combine the subtlety and suppleness of human intelligence, the real strength of which is pattern recognition — we’re actually not very good at logical analysis; computers can already outperform us in that.
Ways in which machines are already superior — I mean, this device can remember billions of things accurately. People are already relying on Google for their memory; you don’t have to remember things anymore.
And machines can share their knowledge. We’ve spent years training this one computer to understand human speech. We trained him like a child, corrected its errors and readjusted its self-organizing neural nets and other paradigms.
And now, if you want your personal computer to understand human speech, you don’t have to go through those years of human training like we did with our research computer. You can just load the evolved patterns that the one computer learned — it’s called loading the software. Machines can share their knowledge. We don’t have quick downloading ports on our neurotransmitter concentrations. Human language is a million times slower than electronic sharing of information.
But this is not an alien invasion of intelligent machines come from over the horizon to compete with us. It’s emerging from within our civilization.
AI’s already much more influential than people realize. We have hundreds of examples deeply embedded in our economic infrastructure.Every time you send a message, connect your cell phone call, get an electrocardiogram, it comes back with an automated diagnosis from your doctor, same thing with blood cell images. Intelligent algorithms guide intelligent weapons, fly airplanes, land airplanes, automatically detect credit card fraud, make billions of dollars of automated financial investment decisions, control just-in-time inventory levels, design products, make them in automated factories.
If all the AI programs stopped tomorrow, you couldn’t get money from your bank, transportation and communications would stop, civilization would grind to a halt. That was not true as recently as 25 years ago. These were all research projects then.
So very often people say, well, what happened to AI? It’s already deeply embedded in our economic infrastructure. But when it really achieves human levels of intelligence, and by 2030, $1,000 of computation will be 1,000 times more powerful than the human brain, it will be quite transformative.
But it is going to — what it’s going to transform what is the nature of what it is to be human. These nanobots will keep us healthy from inside. They’ll go inside our brains. (For instance ?), one application will be to provide full-immersion virtual reality from within the nervous system. You want to go to virtual reality; the nanobot (shut down ?); the signal’s coming from your — (inaudible) — senses. Replace them with the signals you would be receiving if you were in the virtual environment. And then your brain feels like it’s in the virtual environment, and you can go there by yourself or with someone else. (Laughter.)
Design of virtual reality environments will be a new art form. But most importantly, it will extend human intelligence — our memory, our cognitive ability, our pattern recognition facilities.
And I’ll leave you with one last thought. All of this has already — this is not a new story. We’re already gone beyond human limitations. We are the species that seeks to go beyond our limitations. We didn’t stay on the ground; we didn’t stay on the planet; and we’re not staying with the limitations of our biology. And when our genes evolved, it was not in the interests of the species for people to live past child-rearing; that only meant like 28. And human life expectancy was in the 20s 10,000 years ago. It was only 37 in 1800. Sanitation, antibiotics, a few other things have pushed it to 80.
When we have this sort of full mastery of biotechnology, we will extend it dramatically. And it’s only 10 or 15 years from now. That will be a bridge to the full blossoming of the nanotechnology revolution, where we can go beyond biology, have these nanobots keep us healthy from inside.
So if you can stay healthy the old fashioned way for a few more years, we may get to experience the remarkable century ahead.
Thank you very much.
FOSTER: Thank you very much, Ray.
I’m going to try and get us out of here so we can both go home and take our vitamins. I’m sure we all want to make it to 2020 — that’s for sure.
Ray, listening to this very optimistic view, is it fair to characterize you as the anti-Fukuyama?
KURZWEIL: Well, the thing I most disagree with Fukuyama is the definition of what it is to be human.
I just mentioned that my view of being human is we’re the species that seeks to go beyond our limitations, and he prefers to define human in terms of our limitations and that if we overcome our limitations we’ll no longer be human.
I don’t like the word transhumanist, which is commonly used to describe some these ideas because it means going beyond being human. I think we’ll go beyond biology, but I think it’s inherent in the nature of being human to seek to go beyond our limitations.
And he says it’s immoral to do that moreover.
FOSTER: Maybe your next book will be the beginning of history.
KURZWEIL: I think it’s the nature of these exponential trends that even in the next 10 or 15 years we’ll see more change than we’ve seen over the last thousand because it’s the nature of exponential growth.
FOSTER: You’ve certainly brought us a very enlivening picture of what the future could be like, but in your book you’ve also talked about these technologies — some of these technologies — being double-edged swords. You’ve talked about dangerous types of knowledge.
Tell us about the risks and the most dangerous aspects of these technologies?
KURZWEIL: Well, my vision is actually not a Utopian vision. I mean, I think we will be able to solve problems like poverty and environmental degradation. We’ll have, for example, clean energy using nanotechnology and nano-engineered fuel cells and so on. But these technologies, aside from empowering our creative side, also empower our destructive side.
We have a new existential risk today — and actually as we talked about earlier — recognize this when I wrote my first book, “The Age of Intelligent Machines” that the opportunity to bioengineer biological viruses, which could be done for destructive purposes. And I didn’t write about it back then because it was not — it was not out in the public domain and I didn’t want to turn on the TV and have somebody say, oh, I got this destructive idea from Ray Kurzweil’s book after some disaster.
But I did write about it in “The Age of Spiritual Machines” in 1999 because the idea was out there, and this is actually what turned Bill Joy on to the downsides of these technologies.
And this is really an existential risk we have today. You can send in a genome to a mail-order house and get it made for you. And it’s not necessarily easy, but it’s not that hard either to create a new bioengineered biological virus, which we would not have a defense against.
The good news, though, is we actually have some new technologies that can defend us. I mentioned RNA interference that can block genes and turn them off. We send in little pieces of messenger RNA that latch onto the messenger RNA expressing a gene and destroy it. That actually works for biological viruses because viruses are genes.
And I gave some testimony to Congress and also advised the Army, because I’m on the Army science advisory board, as I mentioned, about creating a rapid-response system that could use RNA interference and new vaccine technologies to combat biological viruses.
And some elements of that proposal were in President Bush’s $7 billion program, which is a good start, but it’s too small by a factor of 10. This is a existential risk we face. And it’s really a race. We need to have these defensive technologies in place.
There’s been a big worldwide debate, where some people, for example — Bill McKibben, who was the environmentalist who brought global warming to our attention — I have a lot of respect for him. But he’s recently said, we should relinquish all these technologies. He wrote a book “Enough,” saying technology’s been pretty good, but enough is enough and we should stop these technologies before they get dangerous.
The problem is that would really just drive these technologies underground where we would actually have less opportunity to defend ourselves. That was the moral of the novel “Brave New World.” And it would deprive us of the benefits, as well.
And even though Bill Joy has been associated with this idea of relinquishments, we actually agree on both the promise and the peril of these technologies.
Nanotechnology will have a peril if self-replicating nanotechnology gets out of control. That would be a new danger. And artificial intelligence, if it’s malevolent and more intelligent than us, is obviously a danger.
There are strategies we can use to defend ourselves. It’s a complicated subject, but, you know, in summary I would say this is really, I think, the most important issue facing human civilization. We have tremendous promise. We can really overcome age-old problems with these emerging technologies, but defending ourselves against the peril — and there are strategies we can deploy is — really should be our top priority. And we have to put more stones on the defensive side of the scale by consciously investing in defensive technologies.
One last point on that. We can take some comfort in how well we have dealt with a new self-replicating danger. All of these dangers have to do with self-replication. I mean, human disease is almost all self-replication — bacteria, viruses, cancer cells replicating out of control. The concern that nanotechnology — the gray goo problem — is nanotechnology self-replicating out of control.
We have actually a new pathogen in our civilization that didn’t exist 30 years ago that is self-replicating — the software virus — and it replicates in computer networks. And when these first emerged, observers said, okay, this first generation is crude, but eventually these are going to be so sophisticated they’re going to destroy computer networks, and we won’t be able to use them anymore.
And they have become very sophisticated, but they really remained on a nuisance level because we have this emerging immune system — technological immune system. So a new virus and very clever new attack emerges within hours; we create a defense and distribute it. And it’s been actually quite effective.
If we can do half as well as we’ve done with software viruses in say the biological arena or nanotechnology, we’ll be doing well.
FOSTER: One listens to you talk and immediately one thinks back to nuclear power and the threat of nuclear power and the success eventually of mutually assured destruction as a foreign policy.
What are the policy implications that we should be considering in state of self-replication gone amuck?
KURZWEIL: Well, you know, the existential risk, as I mentioned, that we face right now is bioengineered biological viruses. And one thing that’s required — I mean, not only do we have an interconnected civilization where information can spread around the world in minutes, but biological viruses don’t respect national boundaries, either.
And what China does with its chickens is of, you know, grave concern to us. It’s actually, I think, quite promising that they’re trying to inoculate, you know, six billion chickens. It’s pretty impressive.
But it really requires an integrated, coordinated world response. At first, there was concern about how China was responding to SARS. But finally, actually, they did respond very effectively. We used some old technologies like quarantine and some new technologies like the internet to spread information around.
And also we’ve sequenced SARS in 31 days, which is part of our — (inaudible) — to actually get this new threat that came out and we actually got it under control. I think that’s encouraging. But it requires international cooperation at a very intimate level to actually get in what families are doing with their pets and things like that.
FOSTER: I think it’s time to — let’s move to some questions. I see there’s a number of hands in the audience.
The gentleman back there in the sixth or seventh row?
QUESTIONER: These technological advances are made by brilliant individuals like you and are brought to market in the context of a social order and a rule of law.
Would you comment on the implications for what we need in the way of educating people in this country and protecting the rule of law on a worldwide basis, remembering that a majority of Americans do not believe that evolution is a valid scientific theory and remembering that there are serious threats to the rule of law at home and abroad?
KURZWEIL: Well, that’s a, you know, a pretty broad subject. I’ll comment on a couple of implications of that.
We’re actually not keeping up in terms of education in science and technology. This is a scientific age. I actually did some charts recently comparing American to Asian education in science and technology. And America’s pretty flat. We had 60,000 engineers graduating a year 15 years ago. It’s still — well, it’s actually gone down to 53,000. China was like 15,000. It looks like one of my exponential graphs — it’s up to 300,000 and continuing to grow.
It’s also the same kind of phenomenon — same comparison at the Ph.D. level in every scientific area and the same comparison if you take Japan, Korea and India.
The other side of the story, though, is that we do — we are very good at applying technology. I speak to a lot of different groups, and every group I speak to it feels like a computer conference. I talked to a music conference, but they’re involved in very sophisticated sound processing equipment and sequencers and it read like a computer conference. I spoke at the American Library Association, and that read like a computer conference with data-mining tools and data search and so on.
But there is — but fundamentally we’re not keeping pace in terms of science and technology.
In terms of the rule of law, there’s been — I mean, that touches on a lot of topics. I’ll just mention one, which is intellectual property, and there’s been a lot of concern that some countries, like China, have not sufficiently respected intellectual property. And intellectual property of knowledge — proprietary knowledge — is sort of fundamental to this emerging economy. Everything is becoming information.
The good news, though, is that they’re now actually filing lots of patents and creating a lot of intellectual property, and so they’re probably going to have an interest now in protecting it since they’re going to have a lot of it. So that’s the good news.
But again, we need international cooperation. I mean, the patent system is not really set for this modern world where we have this very fragmented patent system with hundreds of different jurisdictions and that these technologies are instantly available worldwide. There’s this fundamental problem of protecting, you know, the copyright of information. It’s not just music or movies — it’s, I mean — designer products and software and almost everything of value ultimately is going to be these information processes. And if we destroy the business models that allows for the capital formation to create them then there won’t be the intellectual property to distribute.
So I think the answer in terms of foreign policy considerations is an unprecedented pressure for, you know, very deeply intertwined international cooperation and really creation of worldwide institutions that can deal with these worldwide issues.
FOSTER: The gentleman on the aisle here?
QUESTIONER: Thank you very much. I enjoyed your “Spiritual Machines” book, but I have not read the first one, and I look forward to reading this one. And you were just as inspirational as I expected you to be.
I wonder about — to put this in terms of foreign policy, it is really not clear to me that the advances in technology have had much of any effects at all on the thinking that goes into foreign policy decisions.
For instance, in terms of how we got into Iraq. You make a very good point, you know, that it’s really important for us to try to stay alive for a good longer because that may turn out to be very important in terms of where we go from here. Well, we have leaders now who are supposed to keep us alive and functioning well over that period of time.
And I wonder if you really, in the context of your working with the Army and others, have thought seriously about the interface, not just between the use of high-tech tools, but in terms of cogitation, of really coming up with more sophisticated results than we would have in the past.
And I really don’t see that at this stage.
KURZWEIL: Well, — (inaudible) — assess in a contemporary political issues, I will say that there’s a lot of influences now that are affected by the things I’m talking about on the political process.
I mean, take blogs, for example, which really harnesses the wisdom of crowds. I mean, any one blogger may be very unreliable, but the truth of the situation actually can emerge now in a matter of hours with this sort of exchange of 20 million blogs doubling every five months. And that has a real profound impact.
And I do think this decentralized technology, aside from the wisdom of any one leader or one administration, is deeply democratizing. In my first book, “The Age of Intelligent Machines,” I predicted that the Soviet Union would be destroyed by this emerging decentralized technology that said they would either adopt these very powerful workstations, which were much more powerful than the copiers they had been banning, which would destroy centralized control of information, or they would try to ban them, and that would destroy their economy. And I actually felt they would do a little bit of both and both would do them in.
And in that coup against Gorbachev in 1991, although the photo-op of Yeltsin bravely standing on a tank, it was really this emerging clandestine network of fax machines, early e-mail, with (teletype ?) machines, where everybody kept in the know and the authorities grabbing the centralized TV and radio station, which had worked in the past, no longer worked because everybody knew what was going on.
I actually mentioned this at a luncheon I had the privilege of sitting with Gorbachev and mentioned this theory. And he readily agreed with that because anything that put Yeltsin down — (laughter) — he was going to agree with.
But I think the whole movement we’ve seen towards democracy in the 1990s at the political level has been fueled by this decentralized electronic communication. And it’s becoming quite intensive. And it’s not just at the political level. You know, a patient going into a doctors office now is armed with information that they have gathered. If they have a chronic condition, they’re in touch with everybody around the world. They’ll know more than their doctor because the doctor’s keeping track of a lot of things.
So I think it is a deeply influential technology, even though this censorship of the blogs in China — there’s millions of blogs in China. The censorship is actually dealing with — in a very specific narrow issues. I obviously don’t support that. But it is actually a very democratizing force. It will be interesting to see how that plays out.
FOSTER: And another question. And when you rise, please state your name and affiliation so Ray has some sense of you who’re — young lady in the front row here.
QUESTIONER: Sharrid Lizob (ph) of Lehman Brothers (ph).
My question’s actually about dissemination of all this technology. It seems that if there is a 10-year gap between the haves and the have-nots at the beginning of any new technology, that doesn’t matter so much. But as you really start to get up to the exponential part of the curve, a 10-year gap can be huge.
And I’m wondering if you found that the increase in dissemination of this technology is also increasing exponentially.
KURZWEIL: Yes, it is. And you’re right about the 10-year gap. I mean, right now it’s a 10-year gap from early adoption to late adoption.
Early adoption is when the technology is unaffordable except by the rich. But, of course — but at that point it doesn’t actually work very well. A few years later, it works better, but — and it’s merely expensive. And then it becomes inexpensive and it works quite well. And then ultimately it becomes almost free, and actually, it’s quite perfected. And that is a 10-year progression.
But in keeping with this doubling of the paradigm shift rate every decade, that 10-year progression will be a five-year progression in 10 years. And it will be a two or three-year progression in 20 years. And these technologies ultimately will be very inexpensive.
But we can already see the impact. The World Bank released figures recently showing a reduction in poverty in Asia by 50 percent, and at current rates will be down by 90 percent in another 10 years. Asia has exceeded the pace of the rest of the world, but all areas of the world have benefited except sub-Saharan Africa. But even there we will see benefit now because — I mean AIDS drugs, for example, is an information technology and followed this paradigm. Ten or 15 years ago it cost $20,000 per patient per year for drugs that didn’t work very well. They’re now down to about $100 per patient per year, at least in these poor countries. Of course, it’s still too expensive for the individuals, but it’s now affordable by NGOs and foundations and governments that actually provide that. And I think we will see progress in the AIDS problem which will enable even sub-Saharan Africa to benefit from this progression of these information technologies.
I’ve talked to some foundations who are actually planning to give — web-enabled communicators to everybody — or to every family in certain African nations to jump-start an educational system and access to health information and so on.
So ultimately these technologies actually will be enabling and that this lag will be shortened. But even with a 10-year lag, I mean ultimately these technologies do reach everybody. I mean, people say, well, isn’t this increasing the have-have not gap? It’s not. I mean the have-not gap is shrinking as a result of ultimately the dissemination of very inexpensive information technology. And the pace of that will continue to accelerate.
FOSTER: Gentleman on the aisle, yes.
QUESTIONER: Steve Hellman (sp).
Can you comment on the nature and capabilities of quantum computing and the implications on quantum computing on our understanding of the universe and time itself?
KURZWEIL: Well, it’s not clear that quantum computing is feasible on the scale at which it would be useful. Quantum computing has been scaling up very slowly. And as you add another Q-bit, the power of the quantum computer increases exponentially. But if the engineering difficulty of adding another Q-bit increases exponentially, then you are not getting any scale. And there is concern that quantum computing could break encryption codes. But then there was the concept of quantum encryption, which would reestablish unbreakable codes. And we actually have quantum encryption working today, but we still don’t have quantum computing.
Quantum computing at, any rate, is not general-purpose computing. It’s only applicable to certain special problems. So it’s always going to be a niche application if it ever works at all. It’s really not part of sort of our future concept. We don’t appear to need quantum computing to emulate human intelligence.
FOSTER: Let’s see here, front row, this gentleman here. Wait for the microphone, please.
QUESTIONER: I’m Bob Waggoner (sp).
I’m curious about technology for generation of electric power and powering automobiles and so forth. With the — such a long time using the internal combustion engine, and fuel cells have been around for a long time, what kind of breakthroughs will it take to get wider application of the most modern generation of nuclear plants and fuel cells, which would be non-polluting essentially, in transportation?
KURZWEIL: Well, fuel cells is a way of storing energy. But I talk about this extensively in the book. We basically have sort of old-fashioned, first-generation industrial technology still dominating energy.But ultimately we’ll be able to sort of overcome the energy problem using renewable energy through nanotechnology.
I will give you a couple of examples. If we captured 1 percent of 1 percent of the sunlight that falls on the earth, we could meet 100 percent of our energy needs, and that will grow to 3 percent of 1 percent by 2025.
We can’t do that today because we still have these old solar panel technologies that are heavy, inefficient, hard to install, expensive and so on. There is a new generation of nano-engineered solar panels that are better. I think you’ll start to see some impact over the next five to eight years. But if you go out 20 years, you really will be able to create extremely inexpensive, highly efficient solar panels that could be integrated with common building materials and capture that what would then be 3 percent of 1 percent of the sunlight and meet all our energy needs, and then store them in nano-engineered fuel cells that will highly distributed at the opposite end of the spectrum from the highly centralized energy facilities that we have now.
That’s another major trend I didn’t get to comment on is moving from centralized facilities like liquid natural gas tankers and nuclear power plants to highly decentralized ones that are highly stable and invulnerable to disruption. I mean, the Internet is the — sort of the classic example of a decentralized system. Nobody has taken the Internet down for even one second or even a portion of it.
We ultimately will have an energy system with, you know, billions or trillions of these tiny nano-engineered fuel cells getting energy from a number of renewable sources. We could do it entirely with solar panels, but there are other promising ideas, as well.
So that’s one problem I think we’ll actually be able to get under control within 20 years. But it does require sort of full nanotechnology, which we won’t see for another decade and a half.
FOSTER: Gentleman in the third or fourth row here.
QUESTIONER: I’m Bruce Schearer.
Normally, I’m more interested in the consequences, but you got me stimulated about the causes. Those are really amazing graphs and charts, and they’re all based on deep mathematic correlation with power functions.
And that one that goes from, you know, life all the way down — you have this straight line. What’s going on? Is there some sort of underlying force here, a reverse entropy function built into nature’s laws? Is it intelligence design? I mean, how can it be that we keep going down this progression? (Laughter.)
KURZWEIL: It is a reverse entropy function. That’s the nature of an evolutionary process. If you have this increase in complexity — I just had a debate on this subject. I’m not saying that every step in an evolutionary process goes toward increased complexity. It goes through — every step leads to greater adaptation, and some of those steps lead to greater simplification. Some don’t increase complexity; they just — complexity stays the some. But some do go towards increased complexity.
So if you look at the complexity of the most complex entities, as you go forward in an evolutionary process, it definitely increases. So billions of years ago we had these sort of simple one-celled creatures. Then we had multicellular creatures, and the complexity of the most complex entities has increased.
We still have the simple, one-celled creatures running around. And that’s an inherent — and I talk about this in Chapter 2 of the book. The law of accelerating returns is really a theory of evolution, both biological and technological. The competitive landscape of technology is sufficiently complex to be an evolutionary process and as I pointed out, emerged from biological evolution.
And that is the nature of evolution. It creates increasing complexity at a predictable way. It creates a capability and then it uses that capability to evolve the next stage.
So it has — evolution essentially has more powerful tools to use for the next stage of evolution. And that’s why the Cambrian explosion went 100 times faster than the stage that created DNA, for example. And we see that clearly in technology. And that is really the reason that this accelerates, and I have a mathematical treatment of that in the book as well.
FOSTER: We’re going to take one more, and the chair is going to save the last question for the chair.
This gentleman here in the fourth or fifth row.
QUESTIONER: Pete Mansoor, Council on Foreign Relations.
Congress has mandated that by 2015, all — one-third of the aircraft that the Air Force flies and one-third of the vehicles that the Army drives must be unmanned.
Given what you’ve said tonight, is that too conservative a goal? And I’m just wondering if you could comment a little bit on what you see as the future of weapons, given what you’ve talked about?
KURZWEIL: Well, I’ve been pushing the Army in this direction. I think weapons will get smaller, more autonomous and will be unmanned. And I think it will move faster than the official plants.
The armed Predator was actually not a plan. It just kind of happened, and it worked so well, they’ve started just at an ad hoc basis using it quite extensively.
And we’re spending now I think $100 billion on the Joint Strike Fighter that’s to come online in 2025. I think that’s going to be the last major project of a very expensive manned fighter aircraft. That is where military technology is moving, which is the same direction that a lot of other technology is moving, toward more autonomous distributed systems.
The — in fact, DARPA created this worldwide mesh concept where — right now, your cellphone and your laptop are not part of the network. They’re spokes into a network, and then there’s this network out there that organizes the information. But it actually would make more sense to have every device be a node in the network that not only send and receive your own messages, but to passed on other messages and you’d have this mesh of self-organizing devices.
So the Army actually created this system so they could drop a battalion in place and the communication would self organize, and the pieces that went down, it would still be very stable. There would be no central hub to the information, and Intel and Microsoft have actually adopted now this standard of the worldwide mesh, and that’s going to be — in five or six years from now you will see that all your devices will actually be parts of the network, and the network’s going to be constantly self organizing.
So it’s all part of this trend towards these very centralized systems we have now towards highly decentralized self-organizing systems. And for, you know, better or worse, for promise or peril, that is where military technology is moving, as well.
FOSTER: I’m going to take the opportunity for the last question here, Ray.
You’ve — you talked about these technologies that transcend national boundaries. On the other hand, the nation is the bedrock of foreign affairs and the organization of the world as we know it today. It’s quite possible to imagine the international forces that you have described in breaking down national boundaries. But it’s also quite possible to imagine these technologies seen as national systems to defend and protect the national interest and actually build futher walls among nations.
How do you think it’s going to go and why? If we are going to go this international world that you believe is a requirement for managing this technology, how are we going to actually get there from where we are today?
KURZWEIL: Well, it’s here. I mean the Internet exists, and we have probably a billion users. And it’s growing. And it has deep roots, and no one government controls it; no one organization controls it. And more and more of our lives are going to be on the Internet. I mean ultimately the Internet will be a whole virtual reality environment where we can share information and — I just actually had the pleasure of my first board meeting at MIT, and we’re giving away 60 percent of our courses. It will be 100 percent by 2007. And there are thousands of schools around the world — for example, a school in Pakistan, they’ll just a computer, put it on the Internet, and the teacher will, you know, have the students huddle around this one computer, and they will take an MIT course.
And they have access to all the course material, and ultimately this open courseware system will include attending the lectures. And when we have full-immersion virtual reality in — early in the next decade, it really will be no different than attending class. And the students officially enrolled will also just be doing it virtually. And it’s going to be this one worldwide activity where educational, commerce is going to be this one integrated world economy based on this international network of communications.
And the old paradigms don’t slip away instantly. We still have nations. We still have boundaries. We still have customs officials. But a lot of commerce just ignores that. It just — you know, anything that has to do with information just is done on the Internet; it’s a virtual space. There is no conflict of nations or of foreign policy in that world, and that world is growing in power.
Now, the old paradigms don’t disappear instantly. I mean, horse and buggies didn’t go away when the car first emerged. Generally, these technologies when they first emerge are crude and don’t supplant the older technologies, but gradually older concepts do reach antiquity.
And it affects not just products and technologies but social and cultural institutions. We’re going to have to rethink the nature of education, of governments, of social institutions, the concept of religion. I mean, all these things are going to have to be rethought. We’re certainly already overturning many of the business models. You know, every business you talk to is really being transformed already through the advent of these kind of international markets. And nobody can really rest easy, because we’re constantly — you know, any one business model — I mean Google looks very strong now with its business model, but that’s not going to last forever either.
So the concept of the nation, I think is — ultimately is going to become less important, and the concept of this one integrated world — economy integrated world communication system, integrated social and cultural environment — is going to gain in strength and ultimately predominate.
FOSTER: I’d like to thank you very much on behalf of the Council on Foreign Relations for a very stimulating evening.
(C) COPYRIGHT 2005, FEDERAL NEWS SERVICE, INC., 1000 VERMONT AVE.
NW; 5TH FLOOR; WASHINGTON, DC - 20005, USA. ALL RIGHTS RESERVED. ANY REPRODUCTION, REDISTRIBUTION OR RETRANSMISSION IS EXPRESSLY PROHIBITED.
UNAUTHORIZED REPRODUCTION, REDISTRIBUTION OR RETRANSMISSION CONSTITUTES A MISAPPROPRIATION UNDER APPLICABLE UNFAIR COMPETITION LAW, AND FEDERAL NEWS SERVICE, INC. RESERVES THE RIGHT TO PURSUE ALL REMEDIES AVAILABLE TO IT IN RESPECT TO SUCH MISAPPROPRIATION.
FEDERAL NEWS SERVICE, INC. IS A PRIVATE FIRM AND IS NOT AFFILIATED WITH THE FEDERAL GOVERNMENT. NO COPYRIGHT IS CLAIMED AS TO ANY PART OF THE ORIGINAL WORK PREPARED BY A UNITED STATES GOVERNMENT OFFICER OR EMPLOYEE AS PART OF THAT PERSON’S OFFICIAL DUTIES.
FOR INFORMATION ON SUBSCRIBING TO FNS, PLEASE CALL JACK GRAEME AT 202-347-1400.
THIS IS A RUSH TRANSCRIPT.
George Washington University's Peter Bock, the Defense Advanced Research Projects Agency's Paul Cohen, and MIT's Andrew McAfee join Amy Alving, former chief technology officer of Science Applications International, to discuss recent innovations in artificial intelligence as well as the economic and security implications of these technological advances.
George Washington University's Peter Bock, the Defense Advanced Research Projects Agency's Paul Cohen, and MIT's Andrew McAfee join Amy Alving, former chief technology officer of Science Applications International, to discuss recent innovations in artificial intelligence as well as the economic and security implications of these technological advances.
George Washington University's Peter Bock, the Defense Advanced Research Projects Agency's Paul Cohen, and MIT's Andrew McAfee join Amy Alving, former chief technology officer of Science Applications International, to discuss recent innovations in artificial intelligence as well as the economic and security implications of these technological advances.