Does quantum computing have a future?
On Tuesday, Canadian company D-Wave Systems demonstrated a 16-qubit, specific-purpose quantum computer to a room packed with observers and thick with doubt and awe. Reporters watched as the machine solved a Sudoku puzzle and a seating arrangements problem, and, most impressively, searched for molecules similar to the drug Prilosec from a database of molecules.
But the final significance of D-Wave's demo is as uncertain as the fate of Schrödinger's cat -- opinions are all over the place, within the scientific community and without. To cut through the fog, Wired News sought out the father of quantum computing, Oxford University theoretical physicist David Deutsch.
Spooky Computing Quantum computers hold the possibility of solving what computer science calls "NP-complete" problems, the problems that are impossible or nearly impossible to calculate on a classical computer. Picking out a single pattern from a collection of patterns, such as your mother from a photo of people, is easy for you, but beyond the reach of your PC.
The D-Wave machine, called Orion, can't do that yet, but it's a big step in that direction.
The trick in quantum computing is to harness the entanglement of different particles -- what Einstein called the "spooky action at a distance" -- that allows one particle to affect another somewhere else. Orion does this by using rings of current flowing through superconductors. The current can flow clockwise, counterclockwise or, significantly, both directions at once, allowing it to hold two values simultaneously due to quantum mechanical strangeness.
The machine is programmed by changing the magnetic conditions around quantum bits, or "qubits," creating relationships between them that model the physical embodiment of the equation the programmer is trying to solve. The results are read by detecting the direction of the current within the qubit when the calculations are complete.
But significant challenges confront D-Wave in building a useful quantum computer. A key part of making a practical machine will be error correction -- something Orion doesn't do yet, and which requires many more qubits than are currently feasible. Right now, Orion runs its calculations multiple times and determines which answer has the highest probability of being right.
Moreover, scaling up a quantum computer might cause it to lose "coherence," i.e. the entanglement of a distant particle might fail when you introduce too many qubits. Nobody's certain.
Finally, engineering the whole system to be fast enough for practical use and modular enough to deploy at a customer's site remain daunting problems, even if the laws of physics decide to play along.
Deutsch invented the idea of the quantum computer in the 1970s as a way to experimentally test the "Many Universes Theory" of quantum physics -- the idea that when a particle changes, it changes into all possible forms, across multiple universes.
Deutsch is a leading proponent of the theory, so, while he wasn't in attendance at the D-Wave announcement, perhaps it's safe to say as well that he was. Wired News pulled him away from dinner to talk about what a quantum computer really is, what it's good for and what D-Wave's announcement might mean for the future.
Wired News: D-Wave announced 16 qubits, and they want people to play with them, so they're talking about having a web API where people can try to port their own applications and see how it works. Do you think that's a good approach to gaining some acceptability and mind share for the idea of quantum computing?
David Deutsch: I think the field doesn't need acceptability. The idea will either be valid, or not. The claim will either be true, or not. I think that the normal processes of scientific criticism, peer review and just general discussion in the scientific community is going to test this idea -- provided enough information is given of what this idea is. That will be quite independent of what kind of access they provide to the public.
However, I think the idea of providing an interface such as you describe is a very good one. I think it's a wonderful idea....
WN: Can you give a couple of examples of what kind of things can be done with quantum computing that either can't be done, or can't be done practically, with classical computing?
Deutsch: The most important application of quantum computing in the future is likely to be a computer simulation of quantum systems, because that's an application where we know for sure that quantum systems in general cannot be efficiently simulated on a classical computer. This is an application where the quantum computer is ideally suited.
Perhaps in the long run, as nanotechnology becomes quantum technology, that will be a very important generic application.
Another thing I should say is, that application is the only one of the major applications -- apart from quantum cryptography, by the way, which is already implemented and is really in a different category -- that might be amenable to a non-general purpose quantum computer. That is to say, a special-purpose quantum computer.
WN: Can you talk a little about the importance of simulating quantum systems, and give an example?
Deutsch: Yes. Whenever we design a complex piece of technology we need to simulate it, either in theory by working out the equations that govern it, or as a computer simulation, by running a program on the computer whose motion mimics that of the real system.
But when we come to designing quantum systems, we're going to have to simulate the behavior of quantum super positions, which is, in Many Universes terms, when an object is doing different things in different universes. On a classical computer you'd have to work out what every single one of those was, and then combine them in the end with the equations governing quantum interference.
WN: And that becomes computationally impossible?
Deutsch: That becomes infeasible very, very quickly, once you've got more than three, four, five particles involved, whereas a quantum computer could mimic such a process directly by itself doing that number of computations simultaneously in different universes. So it is naturally adapted to that kind of simulation, if we wanted to work out, let's say, the exactly properties of a given molecule.
Some people have suggested this might be useful for designing new drugs, but we don't know if that's the case or not. Although quantum processes are needed in general for atomic and molecular scale properties, not all of them (need quantum processes). An example of that is we've been able to do a lot of biotechnology without having any quantum simulators.
WN: Do you think a quantum computer could eventually build a slightly more macro simulation, something like an immune system, in order to see how it interacts with a drug?
Deutsch: No, that's not what it would be used for. It would be used for smaller things, not things on a larger scale than a molecule, but things on a smaller scale. Small molecules and interactions within an atom, subtle differences between different isotopes, that sort of thing. And of course things on an even smaller scale than that. Nuclear physics, and also artificial, atomic-sized things which will be used in nanotechnology.
Of which at the moment the only ones planned are quantum computers. Of course quantum computer designing other quantum computers is undoubtedly going to be one of the applications.
WN: The other field I can see ... this revolutionizing is materials science.
Deutsch: Yes, yes. Again we don't know how revolutionary it will be, but certainly on the small scale, it will be indispensable.
WN: What would you like to see the field trying to do?
Deutsch: I'm probably the wrong person to ask that because my own interest in this field is not really technological. To me quantum computation is a new and deeper and better way to understand the laws of physics, and hence understanding physical reality as a whole. We are really only scratching the surface of what it is telling us about the nature of the laws of physics. That's the kind of direction that I'm pursuing.
The pleasant thing about that is that can be done before one even makes a quantum computer. The theoretical conclusions are already there, and we can work on them already. It's not that I don't think technological applications are important, but I watch them as an eager spectator rather than participant.
WN: For your purposes, the importance of quantum computing is in the general case more than in the specific-use case.
Deutsch: Yes. The fact that the laws of physics permit themselves to be simulated by a quantum computer is a deep fact about the nature of the universe that we will have to understand more deeply in the future.
WN: How do you think using quantum computers will change how people think about computing, and consequently the universe and nature?
Deutsch: "How they will think about it" is the relevant phrase here. This is a philosophical and psychological question you're asking. You're not asking a question about the physics or the logic of the situation.
I think that when universal quantum computers are finally achieved technologically, and when they are routinely performing computations where there is simply more going on there than a classical computer or even the whole universe acting as a computer could possibly achieve, then people will get very impatient and bored, I think, with attempts to say that those computations don't really happen, and that the equations of quantum mechanics are merely ways of expressing what the answer would be but not how it was obtained.
The programmers will know perfectly well how it was obtained, and they will have programmed the steps that will have obtained it. The fact that answers are obtained from a quantum computer that couldn't be obtained any other way will make people take seriously that the process that obtained them was objectively real.
Nothing more than that is needed to lead to the conclusion that there are parallel universes, because that is specifically how quantum computers work.
WN: So what prompted you to start thinking about quantum computing?
Deutsch: This goes back a long way before I even thought of general purpose quantum computing. I was thinking about the relationship between computing and physics.... This was back in the 1970s....
It had been said, ever since the parallel universes theory had been invented by Everett in the 1950s, that there's no experimental difference between it and the various (theories), like the Copenhagen interpretation, that try to deny that all but one of the universes exist.
Although it had been taken for granted that there was no experimental difference, in fact, there is -- provided the observer can be analyzed as part of the quantum system. But you can only do that if the observer is implemented on quantum hardware, so I postulated this quantum hardware that was running an artificial intelligence program, and as a result was able to concoct an experiment which would give one output from an observer's point of view if the parallel universes theory was true, and a different outcome if only a single universe existed.
This device that I postulated is what we would now call a quantum computer, but because I wasn't particularly thinking about computers, I didn't call it that, and I didn't really start thinking about quantum computation as a process until several years later. That lead to my suggesting the universal quantum computer and proving its properties in the mid-'80s.
WN: How many qubits (does it take) to make the general-purpose quantum computer useful?
Deutsch: I think the watershed moment with quantum computer technology will be when a quantum computer -- a universal quantum computer -- exceeds about 100 to 200 qubits.
Now when I say qubits, I have to stress that the term qubit hasn't got a very precise definition at the moment, and I've been arguing for a long time that the physics community ought to get together and decide on some criteria for different senses for the word qubit. What I mean here is a qubit which is capable of being in any quantum state, and is capable of undergoing any kind of entanglement with another qubit of the same technology, and all those conditions are actually necessary to make a fully fledged quantum computer.
If you relax any one of the those conditions it's much easier to implement in physics. For instance, if you call something a qubit but it can only be entangled with qubits of a different technology, then it's much easier to build. But of course a thing like that can't be made part of a computer memory. (With) computer memory you need lots of identical ones.
There's also the question of error correction. The one physical qubit is probably not enough to act as a qubit in genuine quantum computation, because of the problem of errors and decoherence. So you need to implement quantum error correction, and quantum error correction is going to require several physical qubits for every logical qubit of the computer. When I said you need 100 to 200, that probably means several hundred, or perhaps 1,000 or more, physical qubits.
WN: To get an effective 100 or 200 qubits.
Deutsch: Yes, and that is what would have to count as the watershed for quantum computation, for being a distinctive new technology with its own genuine uses.
WN: That's actually D-Wave's stated goal as well: essentially 1,000 qubits in two years. Do you think engineering-wise, and this is not completely within your realm, they will be able to maintain enough coherence at that level to create a practical computer.
Deutsch: As you said that really isn't my field. Maintaining coherence itself isn't quite enough. They've got to maintain coherence in the operation that I spoke of; that is, the arbitrary superposition, the arbitrary entanglement, and so on....
I don't know. The technologies I've seen so far have got way fewer than 1,000. They've got way fewer than 16. I always have to ask whether the claimed number of qubits are qubits that I would count as qubits by these stringent criteria, or whether it's merely two-state systems that can in some sense act in a quantum way. Because that's a much more lenient criterion.
WN: I don't have the sophistication to answer that, for D-Wave at least. If I were to ask you to cast your mind forward, saying everything goes well, what does a world that combines ubiquitous quantum computing and classical computing look like? And you've said that quantum computing would never replace classical computing.
Deutsch: It's not anywhere near as big a revolution as, say, the internet, or the introduction of computers in the first place. The practical application, from a ordinary consumer's point of view, are just quantitative.
One field that will be revolutionized is cryptography. All, or nearly all, existing cryptographic systems will be rendered insecure, and even retrospectively insecure, in that messages sent today, if somebody keeps them, will be possible to decipher ... with a quantum computer as soon as one is built.
Most fields won't be revolutionized in that way.
Fortunately, the already existing technology of quantum cryptography is not only more secure than any existing classical system, but it's invulnerable to attack by a quantum computer. Anyone who cares sufficiently much about security ought to be instituting quantum cryptography wherever it's technically feasible.
Apart from that, as I said, mathematical operations will become easier. Algorithmic search is the most important one, I think. Computers will become a little bit faster, especially in certain applications. Simulating quantum systems will become important because quantum technology will become important generally, in the form of nanotechnology.
WN: If we have practical nanotechnology, I imagine that's a huge change.
Deutsch: Nanotechnology has the potential of making a huge change. But the only involvement of quantum computers is that it will make it easier to design nanotechnological devices. Apart from that I don't think it's a big technological revolution.
What it is though, philosophically, is taking a quantum world view. That is rather a revolution, but that could happen today and the only reason it has been sluggish in happening is psychological, and maybe quantum computers will help with this psychological process. That's a very indirect phenomenon.
WN: It does allow people to play with it, and they often get things better when they play with them.
Deutsch: That's true.
WN: I wanted to ask you to describe your book a bit.
Deutsch: You'll remember I said for me the most important thing about quantum computation is the way it shows us the deep connections between physics on the one hand and computation on the other, which were previously suspected by only a few pioneers like Rolf Landauer of IBM.
My book (The Fabric of Reality) is about this connection between computation and fundamental physics, between those two apparently unconnected fields.... To me, (that connection is) part of a wider thing, where there are also two other strands, the theory of knowledge and the theory of evolution.
The Fabric of Reality is my attempt to say that a world view formed out of those four strands is the deepest knowledge that we currently have about the world.
