This article was reviewed by a Caltech faculty member.
As part of Conversations on the Quantum World, a webinar series hosted by the Caltech Science Exchange, John Preskill, Richard P. Feynman Professor of Theoretical Physics and the Allen V. C. and Lenabelle Davis Leadership Chair for the Institute for Quantum Information and Matter, or IQIM talks about the promise and challenge of building quantum computers.
As a theoretical physicist, Preskill is interested in the potential of quantum computers to answer some of the most complex physics questions, such as those about the nature of space and time. In conversation with Caltech science writer Whitney Clavin, he discusses how quantum computers can potentially help crack hard problems in a variety of scientific fields, and how the late Richard Feynman had similar realizations when he first proposed quantum computers in the 1980s.
Highlights from the conversation are below, including Preskill's views on creativity in science.
The questions and answers below have been edited for clarity and length.
What is a quantum computer, and how does it differ from a normal computer?
I thought you'd start me off with an easy question! This is actually a pretty hard one to answer, but I'll try. I should begin by saying what a quantum computer is not: It is not just a better, more powerful, faster version of the conventional computers we use now. It processes information in a fundamentally different way using the principles of quantum physics.
We've known for a long time, nearly a century, that all matter is governed by quantum theory, and that's led to a lot of new technologies like lasers and magnetic resonance imaging (MRI), and to putting billions of transistors on a chip. But those technologies have only scratched the surface of how quantum theory has modified our view of what's possible in the universe.
In particular, they don't take into account that when we have many subatomic particles that are strongly interacting with one another, quantum mechanically, those particles speak a very extravagant language, which is very different from the language that we understand and that our computers understand. There's no way to succinctly translate that quantum language to ordinary bits that our computers know about. If you want to describe what a few hundred particles are doing using bits, you'd need more bits than the number of atoms in the visible universe. There's this extravagant complexity, and we want to take advantage of it for speeding up the solutions to some really hard computational problems.
There is a lot of hype around quantum computers. What are they really going to do?
The hype is natural in a way. Everybody understands that computation is important, that it affects our daily lives, that it has economic value. We've seen in recent years a sharp ramping up of interests in the tech industry and from investors in quantum computing. That's a good thing in some ways. It accelerates progress and provides opportunities for people to work in the field. But we should be realistic about the timescale for quantum computing having a big practical impact. And we should also appreciate that quantum computers probably won't be able to speed up everything we want to do with computers but will apply to a special class of problems—and we still have only a partial understanding of what those problems are. We'll understand it better when we have quantum computers and can experiment with them.
How long do we have to wait? One year? 10 years? 100 years?
Well, it depends on what you want. We're at a very early stage of the development of quantum computers, but even now, from a scientific perspective, the quantum computers we already have are empowering. They enable us to explore the behavior of complex quantum systems in ways that we've never been able to before, and that will fuel scientific discovery over the next five or 10 years. But for widespread practical impact, I think a reasonable estimate is decades, or more than 10 years.
What potential applications excite you the most?
Well, remember I'm a scientist, so I like to think of computers as tools for advancing science, for scientific discovery. We use computers to illuminate how nature works in a variety of ways. That applies to much of chemistry and the science of materials. We know the equations; they describe how electrons interact with one another electromagnetically and interact with atomic nuclei. But they're just too hard to solve for large molecules or complex materials. Quantum computers will be good for solving problems like that.
And that will also have practical impact eventually. Computational chemistry, for example, can facilitate the discovery of new pharmaceuticals and chemical catalysts and that can have a broad impact on humanity, but it may be a while before we see that impact.
Another way we use computers is to help us discover new laws of physics, and I think quantum computers will be useful for that too. One thing I've been interested in for a long time is, how should we think about the quantum mechanics of spacetime itself? That's important if I want to understand what happens when something falls into a black hole or what was happening very early in the history of the universe. Quantum computers will help us to get a better grasp of those things by allowing us to simulate quantum phenomena that would otherwise just be too hard to study.
Is this what Richard Feynman was thinking quantum computers would be used for when he first proposed them in the 1980s?
I knew Feynman; we overlapped on the Caltech faculty for about five years between the time I arrived and when he passed away, and we talked a fair amount about science. We didn't talk about quantum computing, but we did talk a lot about subnuclear particles and how they behave. That's a case where we think we know the equations. We have a theory called quantum chromodynamics that describes how protons and neutrons, and their constituents behave. But as with chemistry, the equations are just too hard to solve. Although we know the right equations, we can't compute how that process goes from first principles, and that's something a quantum computer would be able to do.
I think part of what aroused Feynman's interest in the possibility of quantum computing is that he realized that problems like that would be just too hard to solve. And of course, he also realized that if we could simulate how quantum systems behave, that would have other implications, including empowering us to do computations in chemistry and materials that would be out of reach otherwise.
Would you say there's a creative side to what you do?
Yeah, I hope so. I mean, scientists are problem solvers. Well, everybody is, right? I mean, it's part of the human condition as we all have problems, and we want to solve them. And so, you have to figure out how to solve problems. You also have to figure out what the problems are that you want to focus on, what questions to ask. And I think both of those things involve flashes of creativity and intuition. If you're working on a problem and it's worthwhile to know what the solution is, then if it were easy, somebody probably would have figured it out already. So sometimes you're stuck.
But then you see a connection between the problem you're currently thinking about and something you've thought about before. Those kinds of analogies actually are very helpful. And sometimes it just comes out of nowhere and sometimes you have to brute-force it. You realize you can solve a problem if you just work hard enough and the path is clear, but you have to work out all the details. It's those creative steps that are the most fun when you don't really see where the solution is going to come; if you look at a problem from a different angle, you realize the right way to think about it.
Does Caltech have a quantum computer?
Caltech has a partnership with Amazon Web Services, which has opened a center for quantum computing on our campus, and they're pursuing the approach based on very cold electrical circuits. That's a good partnership, I think, because both sides see the need to focus on the long-term problems. To me, it seems most likely that the applications for the next five to 10 years will be tools for scientific discovery rather than for solving problems that are of interest in business, for example. To get to the stage where we can have a big practical impact from quantum computing, we're going to have to have better and more qubits. We're going to have to solve some big system engineering problems.
Here are some of the other topics addressed in the video linked above:
- Surprises in quantum sciences that we've learned in the last 25 years
- The day-to-day life of a quantum physicist
- The phenomena of decoherence and entanglement and their role in quantum computing and quantum error-correction
- Amplifying microscopic quantum effects to the scale of computers
- How quantum computers could help us resolve open questions about black holes, the early universe, and quantum gravity
- How quantum computers are programmed and the connection between quantum computing and artificial intelligence, or AI
Learn more about the quantum world on the Caltech Science Exchange: