Top banner


An interview with Seth Lloyd

Greg Ross

Computers get faster every year, but Seth Lloyd thinks he has the limit finally in sight. The MIT mechanical engineer envisions an "ultimate laptop" that thinks with every elementary particle in its 1-kilogram structure. The good news: It will perform 10 million billion billion billion billion billion logical operations per second. The bad news: It will operate at a billion degrees Kelvin.Seth LloydClick to Enlarge Image

Such are the dreams of quantum computing, an emerging science that enlists elementary particles to process information. To a quantum engineer, everything in the universe is a store of information, because of the inherently "digital" nature of subatomic states. Every time two particles interact, their "bits" are transformed; thus, Lloyd says, the universe itself is a giant computer. What does it compute? "Its own dynamical evolution," he says. "As the computation proceeds, reality unfolds."

Lloyd has become a de facto ambassador for this new discipline, turning up variously (as author and subject) in The New York Times, Nature, The Economist and Wired. In his latest book, Programming the Universe (Knopf, 2006), he expounds on the quantum effects that permit huge computing efficiencies at tiny scales. American Scientist Online managing editor Greg Ross interviewed him by e-mail in May 2006.

When people ask what you do for a living, what do you tell them?

I am a quantum-mechanical engineer: I engineer atoms. I build quantum computers, devices that store and process information at the level of individual atoms. One atom, one bit. In fact, you might even call me a quantum mechanic. If your quanta are broke, we fix 'em.

Do you see quantum computing mostly as a metaphor, a way to think about nature? Or are you focused on actual applications?

Quantum computing is a technological reality. My colleagues and I build quantum computers. If you have a device that works, that is hardly a metaphor. Or, if quantum computers are metaphors, they are so in the same way that a car is a metaphor.

Quantum computers do provide us with new ways to think about nature. That application, to me, is their most important. Indeed, we can build special-purpose quantum computers with a billion billion bits that allow us to simulate and reproduce aspects of nature that could never be simulated on any classical computer, even one the size of the universe itself. These quantum analogue computers, which allow us to construct quantum analogues of other quantum systems, are themselves devices, not metaphors. But boy oh boy, are they helpful in making sense out of nature. I'd say that it is a practical application of quantum computers.

What's the largest working quantum computer today?

For general-purpose quantum computation, only a dozen bits or so. Tiny steps for tiny bits. But for special-purpose analogue quantum computing, as described in the previous paragraph, we can construct quantum simulators with a billion billion bits, larger than any classical computer, and capable of performing computations that no classical computer could ever perform, even if it enlisted every elementary particle in the universe in that computation. Every few years, quantum computers get bigger and bigger.

What are the advantages of working at such a small scale?

First of all, bits are packed more densely together, with only one ten-billionth of a meter—one angstrom—between bits. But the best part of working at such a small scale is that atoms and bits behave in an intrinsically quantum-mechanical way at that scale. Quantum mechanics is weird and counterintuitive. By working at a small scale, we can enlist that quantum weirdness to do computations that no classical computer could perform.

You've said that, seen in this way, the universe itself is a vast computer. If that's so, what happens to free will?

Free will is safe. Even if the universe is completely deterministic, then we (and computers, and God knows who else) possess free will. At first, the deterministic nature of the laws of physics would seem to forbid free will: No choice is available. In fact, however, the computational nature of the universe actually guarantees free will.

Let me explain. Free will arises when we make decisions—decisions that we and we alone are responsible for. For example, every morning I decide whether to have coffee or tea. The decision is mine, and mine alone. Until I make it, I have no idea whether I will have coffee or tea. My decision process is a kind of computation: I weigh the relative merits of coffee or tea, thinking about my day ahead, and then make a decision.

But exactly because the decision process is a kind of computation, the outcome of this process is intrinsically unpredictable. Why? Because any process that involves logical reasoning is intrinsically unpredictable: The result of such a process—and my eventual decision for tea or coffee—can only be determined by going through the same reasoning process oneself. Until one has actually gone through the reasoning process of making the decision, the actual decision will be unpredictable. This verbal argument can easily be made mathematically precise by restating it in terms of mathematical logic, of the sort that computers practice.

One of the most famous results of computer science is the so-called "halting problem," which states that the result of any computation is itself impossible to compute without going through the same sequence of logical steps that the computer programmed to perform the computation undergoes. Ironically, it is exactly when we are most rational and deterministic that free will shows up.

Do human societies process information?

They certainly do. It is exactly because human societies exchange, mull over and process information—both verbal and mathematical—that human beings are such a blight on the world (at least according to other large mammals, such as elephants or whales). Human beings process information in a special way, using universal language and grammar.

This universal mode of communication allows human beings to form societies, organize those societies in complex fashions, construct elaborate social rituals, etc. Exactly because of their ability to process information in this special way, human beings are special.

Has this work influenced your own vision of the world?

My vision of the world as processing information arose out of my day-to-day work building quantum computers. Since I made this realization, my own vision of the world has changed and evolved. The more information I process in my own thoughts, the more convinced I am that the theory of the computational universe is the right theory.

What are the next hurdles to be crossed?

It's hard to build large-scale, general-purpose quantum computers. I hope that we can cross that hurdle soon and build general-purpose quantum computers with hundreds or thousands of quantum bits. But the most important hurdle is in our own understanding. Unless we can understand how the world processes information at a quantum level, we will remain in the dark.

Programming the Universe is reviewed in the July-August 2006 issue of American Scientist.

comments powered by Disqus


Bottom Banner