Quantum computing is the heir apparent of traditional semiconductor-based computing, with many scientists and engineers pooh-poohing skeptics who claim that a real general-purpose quantum computer is impossible. Recently, researchers tested a supposed quantum-computing system from D-Wave; the results indicated that this system outperforms a “classical” computer on certain difficult optimization problems. (An archetypal example of an optimization problem is the traveling salesman, which involves a set of cities and distances between them; the goal is to find the shortest loop that visits all cities exactly once and returns to the home city.) Additional research may lend further credence to D-Wave’s assertions, but questions remain.
Setting aside the matter of whether quantum computing is possible—a matter whose technical details go far beyond the scope of expertise and even interest of many IT professionals—perhaps the more relevant question is whether it is both practical and even desirable. Practicality is usually ignored when research yields a novel technology or concept, but it quickly imposes its tyrannical grip when matters of money and marketing enter the picture. Desirability, however, is seldom considered: a powerful technology is always considered desirable—at least until it proves to have inhuman ramifications.
Recently, The Data Center Journal considered whether a break from Moore’s Law would be beneficial—both for the technology industry and for society at large. Quantum computing, if successfully implemented as its most visionary proponents expect (scalable general-purpose computers that far exceed even the loosest theoretical limits of classical computers), could continue or perhaps far outpace Moore’s Law. Beyond even the matter of practicality, a relevant question in light of the recent NSA spying scandal and ongoing concerns over the role of major corporations (e.g., Google) in our lives is whether society (read: humanity) can handle the theoretical powers enabled by quantum computing.
A Different Evolution
Classical computing based on semiconductor technology started with a simple transistor model built using macroscopic, readily available materials under normal conditions of temperature and so on. The history of classical computing since then has been to miniaturize, enhance and combine many of these transistors into increasingly complex integrated circuits. Quantum computing is following a much different path: instead of starting with a simple macroscopic model under normal conditions, it starts with a microscopic (meaning invisible to the naked eye) model—the quantum bit, or qubit—under other-worldly conditions of extremely low temperature. The D-Wave computer, for instance, operates at 20 millikelvins above absolute zero; cryogenics accounts for most of the machine’s bulk (it’s “about the size of a sauna,” according to Nature). By comparison, liquid nitrogen has a maximum temperature of 77 kelvins, and liquid helium has a maximum temperature of 4 kelvins.
Thus, unlike classical computing, quantum computing aims not so much to make a practical prototype scalable and marketable, but to make an arguably impractical prototype scalable and marketable. (Qubits also suffer from other problems when used in “large” numbers—so-called quantum decoherence, for instance.) Naturally, given the challenges that other researchers have faced, the D-Wave device (the only product presently claiming to be a quantum computer) is the target of much skepticism, although recent research may allay some of those questions. This particular approach carries some limitations, however, in that it is not designed as a general-purpose quantum computer, but a machine designed for specific tasks. Comparisons with general-purpose classical computers are therefore at least somewhat biased, although this bias may not warrant entirely discarding the machine’s advantages.
It’s Not All About Compute Power
Say a company created a quantum computer that was 10,000 times faster than the fastest semiconductor-based processor at every conceivable algorithm. That product would take over the data center and the desk of the common individual everywhere, right? Well, not so fast. First, what if that computer costs $1 million, compared with, say, a $100 chip from Intel or a similar chip manufacturer? Already, the quantum computer has lost its edge in compute power per dollar. Add to this the fact that it probably involves some form of supporting cryogenic system—consuming lots of space and, likely, power—and the combination of operating and capital expenses may just make the traditional computing route seem a bargain.
Of course, it’s easy enough to say the costs of such a quantum computer will fall over time, and advances in technology will enable smaller and more power-conscious models. But it’s just as easy to claim that the need for cryogenics and problems with decoherence may never be overcome in an economically feasible manner. Money is a greater driver of technological development than bare physical possibility. A flying car is conceivable (and likely possible, technologically), but the economics of development and marketing make it a practical impossibility. Researchers who wish to make the case that quantum computing is the future of data processing must not only show that a quantum computer is physically possible—they must, in addition to their roles as scientists, take on the role of economists and marketers. Of course, some will simply follow the possibilities where they lead without concern for whether the results will yield any value for the rest of humanity. Depending on the economics of the situation, quantum computing—even if possible—may simply become a pie-in-the-sky idea, like the iconic flying car.
But Should We Pursue Quantum Computing?
Our taste for the new and flashy makes the idea of retrofitting data centers with quantum computers seem quite palatable. Imagine a powerful network spitting out cures for debilitating diseases, checking software (before its commercial release) for bugs and on and on. Alternatively, imagine it spying on your every move, every moment, and feeding that information to government agents who will predict whether you’ll commit a crime (and jail you accordingly) and to companies like Google who will flood you with ads they know you’ll find hard to resist.
These latter considerations may seem like the stuff of whacky conspiracy theorists, but the recent revelations of broad and invasive government spying on anyone and everyone by way of the NSA makes such scenarios far more believable. Imagine the NSA or some other alphabet-soup agency outfitting itself (at your expense) with a few quantum computers to process data about you that much faster. Maybe we won’t end up with Skynet trying to wipe out all of humanity, but given too much power, people are plenty able to accomplish that feat themselves. The chilling words of J. Robert Oppenheimer regarding the development of the nuclear bomb should always call us to care in the development of technology. Maybe fighting against the creation of some new device is a losing cause, but we can work in other areas to ensure that society is ready to handle it. Presently, as the NSA’s actions show, society (or, at least, the U.S. federal government) is incapable of responsibly holding the kind of power that classical computers offer. Adding new, more powerful toys to an already frightening arsenal is unconscionable.
Researchers will continue to debate the possibility of scalable quantum computing, and even whether D-Wave’s products truly constitute quantum computers. But the ultimate test is not so much possibility as practicality: the development path of quantum computers is very different from that of classical computers, so attempts to simply extend Moore’s Law and treat quantum computing as a foregone conclusion are suspect. Even if products emerge in forms that are affordable to, say, major corporations and governments, they may still lack the ability to compete with classical computers when cost, space and energy requirements are considered. Given the increasingly urgent push to find lower-power solutions in data centers, quantum computing must do more than just pack a large compute punch. Only time will tell if quantum computers—the heirs apparent of classical computers—can achieve the same success as their older brethren.
But possibility—and even practicality—should not be the only consideration. In the near future, assuming quantum computers are implemented (or have been implemented, in the case of D-Wave), only governments and major corporations will have access to these systems. Given the collusion of technology giants like Facebook and Google with NSA spying (even if they did put up a mild fight), a huge infusion of compute capabilities will certainly be used to further any number of nefarious goals. The public response to the arrival of Google Glass—although this product is not nearly as exotic as a quantum computer—shows that people may be starting to consider whether new technologies should be embraced wantonly, or if they should be examined and, if not rejected, given only a limited role in society. Such critical thinking may end up being extremely important to maintaining social sanity, should quantum computers be implemented—lest the zealous scientists and engineers pursuing them one day repeat the words of Oppenheimer: “Now I am become death, the destroyer of worlds.”
Image courtesy of jurvetson