[LINK] Google's Quantum Leap

[Stephen Loosley]

Bits: Google's Quantum Leap

The New York Times 26/10/2019


In a paper published in the journal Nature, Google researchers described how they had used a quantum computer to perform, in 200 seconds, a series of calculations that, they claimed, would take the world’s most powerful supercomputer at least 10,000 years to closely replicate.

https://www.nature.com/articles/s41586-019-1666-5

As our colleague Cade Metz wrote, this is the first example of what researchers have called “quantum supremacy” — the point where a quantum computer can perform a task that would take traditional computers a very, very long time.

This achievement won’t turn computing on its head overnight.

First, consider the device itself: A behemoth of exotic lab equipment called Sycamore, it is probably rivaled by only one or two similar devices in the world. This is nascent technology of the bleeding edge. Second, the calculations were esoteric and of little practical application; we are still a long way from such devices doing something as practical as breaking encryption. And third, it is by no means the case that quantum computers are set to totally replace traditional ones, as they could be good at only specific kinds of tasks.

So when Google’s chief executive likened his company’s achievement to the Wright brothers’ first flight in an interview with M.I.T. Technology Review, was that reasonable?

Well, probably. “It’s not particularly useful in itself,” Andreas Wallraff, a quantum physicist from ETH Zurich, said. “But is shows that it can be done. I think at this stage we can be sure that quantum devices will become ever more challenging for conventional computers.”

(You may have read, by the way, that researchers at IBM, who are themselves racing to build working quantum computers, disputed Google’s claim, arguing that supercomputers could perform the task in a couple of days rather than 10,000 years. But arguing that point seems to underscore how big a deal the achievement is.)

The big unanswered question now is how long it will take for these kinds of devices to run genuinely useful algorithms, which many researchers think might be able to simulate problems that are quintessentially quantum themselves, like interactions between molecules.

Google’s research paper gives a suitably noncommittal nod in that direction: “We are only one creative algorithm away from valuable near-term applications,” it concludes.