Twenty-five years ago, I started working on the problem of quantum computing, which is how atoms and molecules, photons, elementary particles, process information. At that point, there were only half a dozen people in the world looking at this problem, and now there are thousands. There goes the neighborhood. In any field that expands by so much so rapidly, there are now all kinds of branches on this tree. There are still branches of the fundamental questions of how we understand the world, in terms of how it processes information.
Right now, there's been a resurgence of interest in ideas of applying quantum mechanics and quantum information to ideas of quantum gravity, and what the fundamental theory of the universe actually is. It turns out that quantum information has a lot to offer people who are looking at problems like, for instance, what happens when you fall into a black hole? (By the way, my advice is don't do that if you can help it.) If you fall into a black hole, does any information about you ever escape from the black hole? These are questions that people like Stephen Hawking have been working on for decades. It turns out that quantum information has a lot to give to answer these questions.
Then there are a lot of practical questions about understanding what's going on in nature. For instance, it has become clear over the last decade that photosynthesis—where a particle of light comes in from the sun, is absorbed by a chlorophyll molecule, the energy rattles around inside a leaf and gets turned into more leaves—is operating in a very quantum mechanical fashion.
Exactly the same kinds of models that we use to look at quantum computation allow us to understand what's happening in photosynthesis. It turns out that photosynthetic plants, bacteria, and algae are extremely sophisticated in the way they use quantum mechanics. They use quantum coherence and funky effects like entanglement to get very high efficiency of energy transport.
My motto about this is that if a little quantum hanky-panky will allow you to reproduce a little faster, then by God, you're going to use quantum hanky-panky. It turns out that plants and bacteria and algae have been using quantum hanky-panky for a billion years to make their lives better.
Indeed, what's happening in quantum information and quantum computing is it's become more and more clear that quantum information is a universal language for how nature behaves. A few years ago, there was a centerfold in Physics Today, the journal for physicists for the American Physical Society. It wasn't a very sexy centerfold, but it had all the different parts of physics: high-energy physics, solid-state physics, string theory, physics of mechanics, nanophysics. And right in the middle, they put quantum information.
The reason was that this centerfold showed which parts of physics were talking with other parts of physics, who in this field was talking with this other field. They had to put quantum information right in the middle because everybody was talking with the people in quantum information. So, physical chemists—the people who study photosynthesis—all of a sudden were talking with people like me, and we were doing experiments on plants and bacteria.
Now we're doing artificial experiments, in our case, they are systems—man- and woman- and virus-made systems—that mimic the high efficiency of energy transport in photosynthesis. In fact, by using ideas from quantum information, we've constructed systems that are much better than even the most efficient, naturally occurring system.
At the same time as all these theoretical developments, there have been major advances in how we build devices that process information, in building quantum computers. There's a company, D-Wave, that builds these special-purpose quantum computers that are not general-purpose quantum computers that could break codes and strike fear into the heart of the NSA—assuming, of course, the NSA has a heart.
These systems are up and running, and people are buying them and trying to figure out if they can solve hard problems faster than classical computers. The jury is out on this question; we don't know if they can or not. At the same time, people who are building superconducting systems and systems made of atoms or ions and optical systems have gotten much better at building quantum computers, so expect in the next five to ten years to have quantum computers that are large enough to do things no classical computer could ever do. This is a very exciting time in quantum computation for coming up with stuff to do with quantum computers.
There is an old idea, originally due to Richard Feynman, that you could use quantum computers to simulate other quantum systems—a quantum analog computer. Twenty years ago, I wrote the first algorithms for how you could program the quantum computers we have now to explore how their quantum systems behave. In the next few years, we're going to have devices that will allow us to build quantum mechanical simulations of, say, what happens inside a black hole. We can look at what these models do.
A few years ago, some friends of mine and I used a small quantum computer to simulate what happens in the process of time travel, because the theory of time travel is intrinsically quantum mechanical. If you want to find out what happens when you send a photon a few billionths of second backwards in time and have a try to kill its former self, well, we have experiment that tests to see what happens when you do that.
It's lucky that there is no Society for Prevention of Cruelty to Photons, because a lot of photons died in that experiment. It turned out that one photon that tried to kill itself in the past always failed, because our quantum theory of time travel shows that you can't go back and do something self-contradictory in the past, like kill your former self.
One of the most fun applications of quantum computing right now mirrors what's happening in classical computing. Amongst the biggest advances in classical computing these days are programs for machine learning—to take computers and have them process humongous amounts of data and figure out what the patterns are. The NSA uses this to spy on us, Google uses it to spy on us, Amazon uses it to spy on us. Everybody uses this to spy on us.
In machine learning, it's no secret we live in an era of big data. Human beings are generating Avogadro's numbers worth of bits every day or so. Companies like Google and Amazon and Microsoft are processing this data to try to find out every aspect of our lives so they can sell us stuff. Computers are getting very good at processing data and finding patterns in it.
Quantum mechanical systems have this feature where they can generate patterns that are very hard for any classical system to generate. It also turns out that quantum computers can detect and identify patterns that are very hard for a classical computer to detect. For example, if you have a huge dataset like the tick-by-tick history of all the stocks in the Dow Jones over the last fifty years, it's a big dataset.
If you say, "I'd like to process this to find out what a good portfolio would be for me if I can tolerate a certain amount of risk and I want to have a certain amount of return." Well, with a pretty small quantum computer, the kind that we're going to have in the next five years or so, you could find the answer to that question much more accurately then you could do on a classical computer.
Quantum computers work by storing and processing information at the most microscopic level. For instance, if you have an electron, you could call an electron spinning out like this zero, you could call an electron spinning like that one, and you could call an electron spinning like that zero and one at the same time, which is the primary feature of quantum computation. A quantum bit—qubit—can be zero and one at the same time; this is where quantum computers gain their power over classical computers.
For the last twenty years or more, my colleagues and I have been working to build quantum computers using electrons, using particles of light. So, a photon with its electric field wiggling like that is zero, a photon with its electric field wiggling like that is one, and a photon with its electric field wiggling like that is zero and one at the same time. We've been building these quantum computers and quantum communication systems.
I, myself, am a theorist, so the experimentalists don't like me to use a screwdriver in their lab because I tend to break things, but I've been working closely with experimentalists for more than two decades now to build these devices. They started off small, with a couple of quantum bits, but it turns out that just having a handful of quantum bits is enough to do good demos of the ideas of quantum computation. The quantum computers are now getting a lot bigger. Now, we have tens of bits, soon we'll have 50 bits, then we'll have 500 bits, because now there's a clear path to how you build a larger scale quantum computer.
Classical computers famously obey Moore's law. It's not a law of nature; it's just an observation about technological progress, where the size of the components in the computer gets smaller by a factor of two every couple of years and the number of components doubles. Quantum computers haven't been obeying Moore's law. The reason is that to build quantum bits and put them together is a difficult process: You're operating at the most microscopic scale. It's tough to do. You have to control things very precisely.
There is a parallel Moore's law that goes along with the ordinary Moore's law. In fact, it's responsible for this. As time goes on, we're getting better and better at controlling things at very microscopic scales. This same ability to control things is allowing us to make more and more powerful quantum computers. Our quantum computers are still piddling compared with a classical computer. I remember I had a classical computer that had 16K of memory, and then a few years later, it was 64K of memory. Now, it's like 100 GB of memory, or a terabyte of memory.
Quantum computers are still at the stage where we have a small number of bits—10 bits that we can use, soon 50 bits, 100 quantum bits that we can use. Even though this is piddling by comparison with the classical computer, because quantum computers for specific problems are so much more powerful than classic computers, this means that over the next five to ten years, once we get up to something like a few hundred quantum bits—which is going to happen soon—we'll be able to solve problems you couldn't solve on a classical computer.
What problems are we going to solve? A quantum computer with, say, 500 quantum bits, the kind we're going to see soon, would not be able to factor large numbers, break codes, and strike fear into the heartless NSA, but it would be able to do some of these problems, like quantum machine learning—finding patterns in large amounts of data.
Last December I organized a quantum conference on quantum machine learning at NIPS, this gigantic machine learning conference in Montréal. We were expecting a few dozen people to show up, and there were 150 people so that you couldn't get into the room. People in classical machine learning are always on the lookout for new ways of doing things.
They were very surprised to find machine learning problems like looking at the topology of something, figuring out the number of holes in a piece of data. You know, topology studies whether things have holes or gaps or voids or connected components—these are features of the world that people who are analyzing data would like to find. But the classical algorithms that do this, while effective, are only effective on very small numbers of holes, for example, because they just can't process that data.
By contrast, with a small quantum computer, even one with a few hundred quantum bits, you'd be able to find complicated patterns and topological systems like holes and gaps and voids that you could never find classically. We've progressed to a new stage. The first twenty years of quantum computing were very interesting ideas from theory, establishing connections with other branches of physics, coming up with different algorithms that you would love to perform if you only had a quantum computer that was big enough to perform them.
We're on the brink of having quantum computers that are big enough to perform these kinds of analyses, do simulations of other quantum systems that no classical system could do, find patterns in data that no classical system could find. It's going to be a very exciting time for quantum computation coming up.
Who has quantum computers right now? Everybody's got quantum computers. Over at MIT there are five or six laboratories with quantum computers sitting in them, and people are trying to expand them and make them bigger. There are hundreds of groups around the world that are building quantum computers.
These quantum computers in these laboratories are pretty interesting; they look different, according to what you're using. A superconducting quantum computer whose quantum bits are supercurrent, going forever around this circuit in a clockwise fashion—that's a zero; a supercurrent going around the loop forever in a counterclockwise direction—that's a one; a supercurrent going around the loop in both directions at once, which is hard to imagine, but that's what happens—that's zero and one at the same time.
The actual device itself, in the guts of the quantum computer, is a chip. It's a chip that's etched with little superconducting circuits using relatively conventional technologies. Then, the chip itself is connected to wires that come in from the outside world, because superconductors have to sit inside a helium dilution refrigerator at 15 thousandths of a degree above absolute zero. You've got this big thing that sits there looking like a beer keg going clickety‑click-click as it cools down. Then you talk to it using an ordinary computer. You type on your keyboard, which sends signals into the chip, the chip then processes those signals, and it does its weird quantum mechanical thing to get the answers. These things are pretty large right now just because of the dilution refrigerator that contains them. You wouldn't want to put it on your lap because it would squish you. But they are sufficiently compact that you could just have them sitting in your office, if you like.
There is a special-purpose quantum computer, a quantum annealer, that's made by a D‑Wave. This is a commercial device, and now a number of people have bought them. Lockheed Martin has bought a D‑Wave computer, Google and NASA have bought them, the Army is buying some of them. They're buying them because they're very interesting devices. Nobody understands exactly what's going on inside them. They're doing things that are quite mysterious in their own quantum mechanical way, because being mysterious is a quantum mechanical thing. People are buying them to put them through their paces to see if maybe you could solve hard problems that you couldn't solve on a classical device.
This D-Wave device is based on a couple of papers that my graduate student, Bill Kaminsky, and I wrote in 2002. We said that you could take a superconducting chip, make a quantum annealer on it, a special-purpose quantum computer, and gave instructions on how to do it. We didn't patent it because we knew from a simple analysis of our theory that this device would not operate in the way that you wanted to operate, which is staying at its lowest energy state throughout the computation.
Lke fools, we didn't patent it. D-Wave went along and built it; they admit this freely, and why shouldn't they, there's no patent. When they built it, it was indeed true that it didn't behave in the way that you wanted it to behave—staying in the lowest energy state throughout the computation. It got excited to higher energy levels, but it still solves the hard problems. Why is this? Nobody knows. I've been working with the folks at D-Wave to try to figure out why they are successful when they shouldn't be. Ever since then, I patent everything by the way, even if I don't know whether it's going to work or not.
Another fun kind of quantum computer relies on quantum optics, on light. For many years, these devices were massive because they consisted of a bunch of big lasers sitting on an optical table, covered with a million mirrors, carefully aligned by graduate students so that all the beams of light were going in just the right fashion.
There's been an amazing development in this particular field, because, based on techniques from telecommunications, people can now put the whole thing on a chip. You take an optical table the size of a football field, miniaturize it, and pop the entire thing on a little chip this big. The chip is then etched with little lines of silicone, and the photons go zooming along these lines, bounce into each other, bounce all around, interact with each other, and then come out the other side.
These are great devices, and very fun to play with. One of the funky things about these devices is it's become clear, over the last five or six years, that even though in some sense these devices should be very simple—just light moving through a chip, photons bouncing off mirrors and interacting with each other—their behavior can be very mysterious. If you send 20 photons into these little ports going into the chip, and you ask what the probability is that they come out of these other 20 ports coming out of the chip, this turns out to be extremely hard to calculate classically.
Nobody knows how to do it. Yet, this chip can do it automatically. It can generate patterns that nobody knows how you could possibly generate them on a classical computer. They have some kind of weird quantum feature that we can't generate, even using the world's largest classical supercomputers.
One possibility for such devices is for learning. A common feature of machine learning is if you have a device that can generate a certain set of patterns, it can also recognize the same set of patterns. Right now, we're working on an experiment to try to see if we can make this happen. Can we have the patterns that are generated by one of these chips, and then train another chip to recognize a set of patterns? If we can do that, then we will have trained a quantum device to recognize patterns that couldn't possibly be generated or recognized by a classical computer.
These patterns are so weird. Because they can't be generated by any classical object, they're like nothing you've ever seen before, by definition. In addition to having these funky patterns that we have no idea what they would be like, how you generate them classically, quantum computers could do ordinary machine learning tasks, like recognizing large-scale patterns in data—the kind of everyday thing that we now use all the time, things like facial recognition, voice recognition, character recognition.
A very important question if you're investing in the stock market is, are there some hidden dynamics that's driving all the stocks together in some pattern? If you knew what that dynamic was, then you could make a lot of money. A quantum computer could find such patterns much more efficiently than a classical computer could. There are plenty of ordinary problems where a quantum computer, even a very small one with a few hundred quantum bits, could do things that a classical computer couldn't.
Then there are the crazier things like these patterns that are generated by these devices that are never generated classically. I don't know what these patterns are good for in terms of recognizing them, but they are very useful in terms of problems that involve cryptographic applications. For example, encoding information in ways that nobody can decode. If you take your information, put it together with one of these patterns that nobody can decode, then, by God, nobody can decrypt your information.
It's useful to compare the current state of quantum computation and what's been going on over the last twenty years with what happened with digital computers over the first twenty years. The idea of building a digital computer was proposed in the mid-1930s by Claude Shannon—it was part of his very influential master's thesis at Harvard—and by Konrad Zuse, in Germany.
The first devices started to be built then, effectively, in fact, during the Second World War. By the mid-1950s, people had these gigantic, extremely expensive devices. There were very few of them, and they cost a lot to build. It was a huge effort for a very small number of bits that would break down all the time.
The idea of building a quantum computer was something that I proposed in 1993. Quite soon after that, people started building simple quantum computers. It's been tough to do, just in the same way the first twenty years of building classical computers was tough. Now, we're at the stage where we have these quantum computers that fill rooms and have lab technicians in white coats tending to them. They're hard to operate, they break down, and they still only have a few tens of quantum bits; although, of course, we're making progress.
What's interesting about this is that, with computers there's this "If you build it, they will come." I have many senior colleagues at MIT who were participants in the early days of computation, people like Marvin Minsky, Bob Gallager. When they tell me about the good old days, as they're fond of doing, one thing that comes across was that the origins of computer science in the 1950s were tremendously exciting. There was a huge change as soon as they had a device that they could run their algorithms on, even a device that was huge, physically, but incredibly weak and puny by comparison with today's standards.
As soon as people had developed the first computers that they could run programs on, within a few years, these pioneer computer scientists—it wasn't even called computer science at that point—developed many of the most powerful methods that we know today, things like Monte Carlo and Simplex algorithm, all these algorithms that people now use for everything.
It was a tremendously exciting time, and it was exciting because all of a sudden these smart people who had been working in the realm of theory had a toy they could play with. Very fast, they came up with a huge number of fun games that they could play with this toy, this pretty expensive toy.
The field of quantum computing is in that stage right now. We have these toys, these complex, not so powerful, quantum computers, but we can play games on them. We can try out the things that people come up with, and we can see what happens. People are coming up with very fun games. As a result, the field of quantum computing is tremendously exciting for someone like me because it's full of young people with fantastic ideas.
Some of the most brilliant young scientists I know in the world have gravitated to this field, because it's fun: You can play awesome games; the questions are big; you can ask questions about the nature of the universe. You can ask questions like, can I recognize a scrawled five or seven? But then you can work with people and say, "Hey, I've got this idea, can we try it out?" You walk down the hallway at MIT, and somebody says, "Yes, we can try that out. Let's see what happens."
The field of quantum computing right now is a tremendously enjoyable place to be, just from the point of view of intellectual play. For me, it's great, because I get to work with these people who are heck of a lot smarter than I am, and that is a lot of fun too.
Thinking about the future of quantum computing, I have no idea if we're going to have a quantum computer in every smart phone, or if we're going to have quantum apps or quapps, that would allow us to communicate securely and find funky stuff using our quantum computers; that's a tall order. It's very likely that we're going to have quantum microprocessors in our computers and smart phones that are performing specific tasks.
This is simply for the reason that this is where the actual technology inside our devices is heading anyway. If there are advantages to be had from quantum mechanics, then we'll take advantage of them, just in the same way that energy is moving around in a quantum mechanical kind of way in photosynthesis. If there are advantages to be had from some quantum hanky-panky, then quantum hanky‑panky it is.
One should always ask with these technologies if it's really worth it. Is the Internet worth it? It probably is. On the other hand, there's a huge amount of stuff that's not worth it on the Internet. Is it a great thing to have a smart phone? Sure, on certain occasions it is. A lot of times, it's just a distraction, however, that prevents you from paying attention to what's going on around you, makes you run into lampposts while you're texting.
As a professor at MIT, and I'm a professor of quantum mechanical engineering, I'm exposed to novel technologies all the time, most of them are super cool technologies that will never be turned into something that people will use, but they're super cool nonetheless. The strategy I've learned is that there're a huge number of technologies out there, and we don't have to adopt them. You don't have to adopt these technologies.
You can use the ones that you like. You can not use the ones that you don't like. I don't use Facebook or Twitter or other social media, because I feel that there's presence and there's absence, and then there's cyberpresence, and cyberpresence is a heck of a lot closer to absence than it is to actual presence.
Similarly, when I look at my parents and their friends, they have friends whom they go to visit, they stay overnight, they have dinner together, they talk with them; that's what a friend is. Maybe some of your friends on Facebook are friends like this, but probably not most. Most of the things that make human life rich and rewarding are just about being human and not about technologies.
If we're lucky, we can find a technology that helps us understand what's going on better, or is useful for certain purposes, and that's great. We should use those. Of course, a lot of these technologies get used for bad, sloppy, annoying purposes as well. In general, I would say that if we're lucky with technology, where you have the good things and bad things, and if the average is a little better than zero—that's good.
Technology is not neutral. In fact, one of the primary uses of technology, just very naturally in the way that the economy works, is that wealthy and powerful corporations use technology to exploit ordinary people.
I've experienced this a lot just in my own job. When I started off as a scientist, I would spend my time calculating by hand and writing formulas on blackboards. When I wanted to write a paper, I would maybe type it up or write it out by hand and give it to someone who typed it. Then it would go to a journal and there would be a typesetter at the journal. It was a very skilled job to typeset scientific equations.
The typesetter typed up these scientific equations. Now what happens when I submit a paper to a journal is I have to typeset it myself. There is a very nice program called LaTeX, that allows you to typeset scientific equations, but it takes a lot of work. To write a paper, I spend a lot more time than I did in the past.
A bunch of other people lost their jobs. One of the very natural things about technology is that, at the same time it makes some jobs more efficient and easier to do, it means that the people who are doing them end up having more work because your employer is making you do more stuff. Then there are a bunch of other people out there who are out of a job. Technology is not neutral. It can make things more efficient, but it doesn't necessarily make our lives easier or better. In fact, it often makes a lot of us work a lot harder, which I object to.
I find it's easiest to tell people the truth about what's going on, and that's good for them too. There are a number of Fortune 500 companies that are investing heavily in quantum computing. IBM has always been heavily invested in quantum computing. Microsoft, Google, now Intel is making a very significant investment in quantum computing; NEC, in Japan, has a big investment in quantum computing. There are quite a few companies that have decided they want to invest in this field.
When they ask me, "Are we going to have a quantum computer that we can build and sell to people soon?" I say, "Well, maybe not." Though we're much closer now. In fact, it's quite likely with these new advances in the technologies, particularly with things like superconducting quantum computing and optical quantum computing, that we will have quantum computers that people can build and they might be able to sell.
There's a very good reason for a company like Google or IBM or Microsoft or Intel to invest in quantum computing. This is a technology that has tremendous promise, even though at the moment, it's not something that is incorporated in the everyday smart phone. The reason has to do with the nature of computing in general.
When people first built these humongous computers the size of a gymnasium and put them in gymnasiums, they didn't have the slightest clue about what computers would be used for. They were thinking, "Oh, we'll use it to analyze things like shell trajectories, material properties, things like this." But one thing that we've all experienced over the last few decades is that computers can do things that you would never have thought they'd have been able to do. Moreover, information processing technologies have exploded in a way that nobody would have ever expected.
Now it doesn't make sense to just talk about computers, because everything is computing. Your smart phone is a tremendously powerful computer. Your car engine contains twenty to fifty microprocessors that are computing away all the time, and this is the secret for getting much better fuel efficiency with pollution controls, etc. It also turns out to be the secret to cheating pollution controls.
Computation is present and information processing is present in a huge number of devices. This notion that almost anything you touch is capable of processing information in a sophisticated fashion is now commonplace. If you're a company that deals with questions of information processing, it's important for you to know what's going on.
IBM is very invested long-term in quantum computing. They've been strongly invested in quantum computing from the beginning, for more than two decades. That's because they had very good people who were amongst the founders of the field who were developing it—Rolf Landauer and Charlie Bennett. It was clear that amazing things would come out of this. It's not like they're investing $1 billion in quantum computing, but they are investing tens of millions of dollars a year. I don't know what their actual investment is. As a result, they have some of the best and brightest people in the world working with them, coming up with the ideas, who know what's going on and are playing with their own quantum computers that they're building.
The same is true of all these other companies. They're fantastic places for young people to work, and they're some of the main places where new ideas are being developed.
DARPA has always had a close relationship with quantum information processing. By its very nature, program managers at DARPA are always wanting to hit it out of the park, and it's been clear from the beginning that quantum computing is potentially a technology where you can hit it out of the park.
I was a co-principal investigator on the first quantum computing grant from the government, which came from DARPA back in 1994. Jeff Kimble was the leader of this group. DARPA realized right away that this was something they could look at. In fact, over the last twenty years, there have been a wide variety of programs of DARPA investigating different aspects of quantum computation, many of them very successful.
A lot of the fundamental advances in quantum computing have ended up being funded by DARPA in one form or another. I regard this as a good thing, although I'm not sure if the head of DARPA regards this is a good thing. Frequently, what ends up being developed by a particular program is not what they set out to do in the beginning. But it turns out, there's some spinoff that comes out of this program that is tremendously powerful.
DARPA was the first funding agency to recognize that this role of quantum mechanics in photosynthesis was a very important thing. They created the first program to fund looking at funky effects like quantum coherence and entanglement in photosynthesis and in energy transport. It was a very successful program that had wonderful results. Some of the spinoffs that came out that I'm working on right now are these man- and woman- and virus-made systems that are much more efficient in their energy transport than anything that's found in nature.
DARPA has fingers in many pies, and it has got fingers in many quantum pies. It has been part of developing many of the fundamental ideas in quantum computing. IARPA spun off of DARPA. IARPA has also taken a major role in investing in the forefront of quantum information processing.
Because it's grown so rapidly, because it has impacted on so many other fields, there are now a lot of subfields in quantum computation and quantum information processing. There are the techie guys, who are building quantum computers. There are some remarkable people out there. In superconducting quantum computers: John Martinis, who's just been hired by Google; my colleague, Will Oliver, at MIT; the group at Delft. Then there are people who are looking at wide-eyed and crazy ideas about what kinds of new quantum algorithms you can come up with.
Scott Aaronson, my colleague at MIT and the famous blog poster on Shtetl-Optimized, has a remarkable set of ideas. He and his colleagues are mapping out the set of questions that you might be able to solve on quantum computers, and they're doing a wonderful job. One of the most successful, powerful types of devices you can make to build a quantum computer out of is ion traps, where you take a bunch of ions, atoms, you strip electrons off, you trap them a little trap, and zap them with lasers. My colleague, Chris Monroe at the University of Maryland, is a pioneer in this field. Rainer Blatt, at the University of Innsbruck, has done amazing things on this. My colleague, Ike Chuang at MIT, has made lots of progress on building these devices.
My favorite part about quantum information is the wild and crazy stuff, where we say, "Hey, let's understand how the universe is made and how it's put together from thinking about it in terms of quantum information." Thinking about quantum gravity, which is something nobody understands in terms of quantum information, is something I've been doing now for fifteen or twenty years, and now there are quite a few people working on this. It's quite fun.
Senior people include John Preskill at Caltech, and Alexei Kitaev, who is a MacArthur Prize winner. They're both brilliant people, making great strides in this regard. As I said, one of the things that characterizes the field of quantum information is the tremendous quality amongst the younger researchers in the field. Patrick Hayden, who has just been hired at Stanford to look at questions of quantum mechanics and quantum gravity. Brian Swingle, who came up with one of the main connections between quantum gravity and quantum information when he was a graduate student at MIT, all on his lonesome. Let's face it—the guy is brilliant. What can I say?
China is a bit of a latecomer to the quantum information game, although about four years ago, they established an institute for looking at quantum computing at Tsinghua. This is an excellent institute. They're doing great things. There have always been some wonderful individual experimentalists in China doing quantum information, like Pan, for example. Singapore has an amazing program in quantum information processing at the National University of Singapore. It's one of the leaders in the field.
There are many fantastic researchers in Japan on quantum information. My colleague, Yasunobu Nakamura, and groups at the Tokyo Institute of Technology. NEC has always had a great group there. There are many great people there. The largest group or concentration of people working on quantum computation are in Canada at the Institute for Quantum Computing, in Waterloo, where Mike Lazaridis, the founder of Blackberry, has donated hundreds of millions of dollars as seed funds to create a remarkable group of researchers. It's headed by Raymond Laflamme. This is the biggest concentration in the world these days.
There are many fantastic quantum groups in Europe. The University of Vienna has amazing people. Anton Zeilinger has been there for a long time, and Philip Walther. Oxford and Cambridge have great programs, with Vlatko Vedral at Oxford and Richard Jozsa at Cambridge. And of course, David Deutsch at Oxford, who is a founder of the field. It's difficult to see him because he only goes out at night.
I've had many conversations with David Deutsch over the years. One of the most fun conversations was when I was running a session of a conference at MIT, and David was appearing via video link. I was in this gigantic auditorium with a 40-foot tall screen in front of me. I was seated in the front row of the auditorium, and David's 40-foot tall head was talking to me from the screen. There was nobody else in the room, and we were just talking about physics. It was like talking with the Wizard of Oz.
David is a brilliant person, and a deep thinker. I don't need to say it, but it is true. He was the first person to realize that quantum computers could do something fundamental that classical computers couldn't. He realized this in the mid-1980s. It took him a long time to come up with an example of something where a quantum computer could do better. He had this intuition, and he came up with a formal notion of a quantum computer. But for more than five years or so, he couldn't come up with something where it could do better.
Then when he finally came up with something, he showed where a classical computer takes two or three steps on average to this problem, a quantum computer can do it in one. Even that was an advance, but it wasn't a problem that anybody would like to solve. But he was so fixed on the idea of coming up with the idea and figuring out what you could do, that in the end he, by sheer force of intellect and willpower, brought the community around to realize this was important.
Then other people started working on this and coming up with ideas that could be more useful. Quantum theory transformed the problem of finding periods of functions, and then Peter Shor used this to come up with his famous algorithm for factoring numbers and breaking codes. And then it was off to the races. During the last two decades, since the Renaissance of quantum computing, since Shor's algorithm in 1994, it's gone from a half dozen people in the field to thousands of people in the field working on a vast variety of things.
During this entire time, David has stuck by his own lights, and he's continued to work on what he regards as the most important things. For the last ten years, he's been working on what he calls a quantum constructor theory, where so far as I can tell—and I can't say that I understand it very well—he's trying to derive the very nature of reality from ideas based on quantum computing. I wish him good luck in doing so.