This month we need to deal with a big concept: quantum supremacy. Not the story of a subatomic particle trained as the CIA’s ultimate assassin but now suffering amnesia and uncovering a conspiracy. No, it’s the idea that a quantum computer might be able to solve some task that a classical computer cannot, or at least couldn’t do within the remaining lifespan of the universe.
Without this, you may ask, what’s the point of spending billions on developing quantum computers? And that’s a good point. The idea that quantum supremacy will one day happen, and that no calculation, no matter how complex, will be beyond us, is behind all the efforts to get one working. Quantum computing hinges on the idea that subatomic particles are weird, and that weirdness can be marshalled to do sums. Theoretically, a correctly configured quantum machine could do sums that ordinary computers would find impossible. Explanations for how it works vary from the idea that a calculation is being run in many parallel universes simultaneously, to the metaphor of ripples in a lake interfering with one another until only one possibility is left—which must be the answer. None of them are a satisfying explanation.
Unexpectedly, at the end of last year, a Google-backed team announced that it had achieved supremacy: carrying out a computational task (checking the outputs from a quantum random-number generator) in 200 seconds that a classical supercomputer would take 10,000 years to complete. Google used a 53-qubit processor named Sycamore, and broke the news in a paper published in Nature, one of the world’s top scientific journals and very much the PC Gamer of its field.
The paper’s 77 authors—including John M Martinis, Google’s chief scientist for quantum hardware and a professor of experimental physics at the University of California—claim in the paper’s abstract that they had achieved a “dramatic increase in speed compared to all known classical algorithms [and this] is an experimental realisation of quantum supremacy for this specific computational task, heralding a much-anticipated computing paradigm”.
Whenever somebody is celebrating, however, someone else always likes to come along and pour cold water on the festivities, and this time that turned out to be IBM, no stranger to the quantum computer or, indeed, classical supercomputing. And rather than a ding-dong between robotic spokes-AIs, the whole thing played out in the genteel language of the academic paper. Big Blue pointed to a different classical computing technique that could do the random-number-checking in two days “and with far greater fidelity. This is in fact a conservative, worst-case estimate, and we expect that with additional refinements, the classical cost of the simulation can be further reduced”.
Google’s 200 seconds claim is still faster, but it’s not the complete owning that was initially reported. More of a quantum advantage than full supremacy.
In a blog post credited to four writers including Jay Gambetta, the company’s vice president of quantum computing, IBM quoted John Preskill, professor of theoretical physics at Caltech who coined the term ‘quantum supremacy’ in 2012. “It won’t change anything overnight, but it is significant that quantum computers are now at the stage that at least in some arena, they can outperform the best computers on Earth.” IBM went on to praise Google’s achievement, saying, “Google’s experiment is an excellent demonstration of the progress in superconducting-based quantum computing, showing state-of-the-art gate fidelities on a 53-qubit device, but it should not be viewed as proof that quantum computers are ‘supreme’ over classical computers.”
Preskill himself then jumped into the debate with a column in Quanta magazine, a publication that aims to aid the public understanding of a very complex subject and therefore very much the PC Gamer of quantum physics. In it, he discussed problems with the term ‘quantum supremacy’ itself, before congratulating the Google team on its achievement and pointing out the main issue, “The catch, as the Google team acknowledges, is that the problem their machine solved with astounding speed was carefully chosen just for the purpose of demonstrating the quantum computer’s superiority.
“It is not otherwise a problem of much practical interest. In brief, the quantum computer executed a randomly chosen sequence of instructions, and then all the qubits were measured to produce an output bit string. This quantum computation has very little structure, which makes it harder for the classical computer to keep up, but also means that the answer is not very informative.”
So if quantum supremacy has been achieved, it’s only in a way that doesn’t benefit us much—although streams of random numbers do have applications in areas such as gambling, cryptography, and anywhere that unpredictable results are desired. The possibilities, however, are endless. Google could have used Sycamore to factor a couple of enormous prime numbers, which would have the same effect on the RSA encryption ecommerce depends upon in the same way that a tin opener does on a can of beans. It could have identified a room-temperature superconductor, paving the way for breakthroughs as quantum processors would no longer need to be cooled to a few degrees above absolute zero to work.
Quantum computers should also be good at solving optimisation problems, balancing cost and utility in areas such as telecoms infrastructure, road planning, and games of Spacechem. So while we shouldn’t be expecting quantum AI minds to start controlling our post-scarcity interplanetary society just yet, once quantum supremacy is fully here, and scientists start moving across to quantum computers to carry out tasks, things will start to get very interesting.