Thread by @postquantum

Looks like the worst kept secret in quantum technology has finally leaked into the mainstream press ahead of publication. While this is a milestone, it is *very* far from being a quantum computer that can compute anything useful. ft.com/content/b9bb4e…

The Google device solves a "problem" called random sampling. Essentially, you run a very short and random quantum circuit and then measure the output. We believe (although we can't prove) that a classical computer would not be able to produce the same distribution of random 2/

numbers without running for thousands of years. The @FT article says the device is called "Sycamore" (I'd understood it to be Google's Bristlecone device in John Martinis group, which has 72 very noisy qubits). Needless to say, the scientific paper hasn't made it out, so we /3

should be cautious here. I'd heard back in June a draft was ready for submission, but I have no idea at what stage of review the paper currently is. Looks like someone at NASA jumped the gun and posted it on the website (they were using classical computers to solve the 4/

sampling problem for benchmarking). Someone posted a screen shot of the article on twitter (I now can't find it), and the FT got a draft. To learn more about quantum supremacy, check out @skdh's explainer & @KSHartnett's article in @QuantaMagazine. You /5

should also follow @BullshitQuantum to help cut through the hype! 6/

@BullshitQuantum Yikes, from the replies it looks like the toothpaste is out of the tube -- a cache of the draft is available. No doubt the paper is being held up by referee #3, who wants them to verify that their device is doing what they claim by running the circuits on a classical computer. 7/

Since the best known classical algorithm would take 10,000 years to verify this, we can expect the paper to be out in 12019. Peer review takes time! 8/

Nope, wrong guess -- I've been told the paper cleared peer review and is ready to go. Standards have sunk so low neither the referees nor the authors will spend the time required to verify the distribution classically! But the journal (which shall remain nameless) is... 9/

[wait for it]
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
waiting 10/

Okay now the American far-right / infowars crowd has gotten hold of this: infowars.com/googles-quantu…. At this point, I think it would be helpful if @nature (woops) published the paper or at the very least, let the authors comment to tamp down on this nonsense. 11/

There is a real vacuum of information right now and most of the commentary and media coverage is bonkers! 12/

There's been a debate in the quantum information community about the term "quantum supremacy" (because it sounds a bit like white supremacy). Well, now we know... White supremacists hate quantum supremacy! naturalnews.com/2019-09-21-goo… . I can't believe we didn't see that coming. 13/

Another drawback with the term, is it suggests the quantum computer is better than the classical computer, and that it's solving a problem. When really it's just making a lot of noise that's difficult to simulate. Something quantum supremacy & white supremacy have in common! 14/

Some commentary out there saying: lots of quantum systems are difficult to simulate so this is no big deal. But for random circuit sampling, one needs to create a huge family of states. If you can do it, you're able to explore a large Hilbert space, and that's impressive. 15/