## Martinis, The Plot Against America, Kill Chain

Update (May 1): Check out this Forbes interview, where Martinis explains his reasons for leaving Google in much more detail.

I was never big on HBO (e.g., I still haven’t seen a single minute of Game of Thrones), but in the last couple of weeks, Dana and I found ourselves watching two absolutely compelling HBO shows—one a fictional miniseries and the other a documentary, but both on the theme of the fragility of American democracy.

The Plot Against America, based on the 2004 Philip Roth novel of the same name (which Dana read and which I now plan to read), is about an alternate history where the aviator Charles Lindbergh defeats FDR in the 1940 presidential election, on a fascist and isolationist platform, in events that—as countless people have pointed out—are eerily, terrifyingly prescient of what would actualy befall the US in 2016. The series follows a Jewish insurance salesman and his family in Newark, NJ—isn’t that what it always is with Philip Roth?—as they try to cope with the country’s gradual, all-too-plausible slide downward, from the genteel antisemitism that already existed in our timeline’s 1940 all the way to riots, assassinations, and pogroms (although never to an American Holocaust). One of the series’ final images is of paper ballots, in a rematch presidential election, being carted away and burned, underscoring just how much depends here on the mundane machinery of democracy.

Which brings me to Kill Chain: The Cyber War on America’s Elections, a documentary about the jaw-droppingly hackable electronic voting machines used in US elections and the fight to do something about them. The show mostly follows the journey of Harri Hursti, a Finnish-born programmer who’s made this issue his life’s work, but it also extensively features my childhood best friend Alex Halderman. OK, but isn’t this a theoretical issue, one that (perhaps rightly) exercises security nerds like Alex, but surely hasn’t changed the outcomes of actual elections?

Yeah, so about that. You know Brian Kemp, the doofus governor of Georgia, who’s infamously announced plans to reopen the state right away, ignoring the pleading of public health experts—a act that will fill Georgia’s ICUs and morgues as surely as night follows day? And you know how Kemp defeated the Democrat, Stacey Abrams, by a razor-thin margin, in a 2018 election of which Kemp himself was the overseer? It turns out that Kemp’s office distributed defective memory cards to African-American and Democratic precincts, though not to white and Republican ones. There’s also striking statistical evidence that at least some voting machines were hacked, although because there was no paper trail it can never be proved.

In short, what The Plot Against America and Kill Chain have in common is that they would be desperately needed warnings about the ease with which democracy could collapse in the US, except for the detail that much of what they warn about has already happened, and now it’s not clear how we get back.

### 106 Responses to “Martinis, The Plot Against America, Kill Chain”

1. James Babcock Says:

During the 2000 Presidential election, Volusia County, Florida reported a negative number of votes. Upon reinspection, the negative vote total disappeared, and the incident was mostly ignored against the backdrop of “hanging chads”, a recount of the subset of ballots which had a paper trail, and a 5-4 party-lines Supreme Court decision to halt that recount.

Several years later, Harri Hursti demonstrated a voting-machine security vulnerability which involves inserting a memory card into the process which has positive votes for one candidate, and negative votes for another candidate, in order to steal votes without messing up the turnout numbers. If done properly, this would be completely undetectable. But if the hack were performed improperly, and the negative-vote-count memory card were inserted into the process *in the wrong place*, then it would produce exactly what was seen in Volusia.

2. Avi Says:

I’ve been watching Scandal recently, in which a Republican gets elected president through election fraud using corrupted memory cards.

3. Doug S. Says:

Source for the Georgia election problem?

4. Will Says:

I’m really curious if you know what your election security friends think of voting by mail.

I’m not an expert, but I’ve always been skeptical of vote by mail. It seems to me like in-person, paper-ballot voting is able to be secured by the tight constraints of the event: basically every step at which fraud could happen is closely monitored by partisans on both sides, as well as supposedly non-partisan officials. Once you introduce not only the entire Postal Service into the equation, but also everyone’s mailbox on everyone’s front porch, and in some states intermediaries who are allowed to deliver your ballot for you–and also extend the time frame over which the election takes place–it seems like the election becomes much harder to protect from fraud.

To me, the system where people can only vote by absentee ballot if they really need to gives me more confidence. But I would be happy if someone could convince me I am wrong.

But of course, this is thinking from the pre-pandemic era. The costs and benefits of mail-in ballots are different now.

It seems very unfortunate that now advocacy against vote by mail seems to be partly motivated by attempts at voter suppression. Surely we can make elections both accessible and secure against fraud.

5. John Figueroa Says:

Oh wow. I was heartbroken when Abrams lost. But when I read that Abrams had refused to officially concede the election, I had sort of just assumed she was being something of a sore loser, and that it would’ve been better—you know, for the sake of democracy—for her to just concede. And that to whatever extent there was voter suppression, it was just the statistical institutional stuff like fewer voting machines per person in Democratic areas. Ya’ know, stuff that, even if there are specific people/groups to blame, it’s difficult to draw an unambiguous line of causality, and unclear exactly where the line was crossed.

But it looks like I was wrong to be so cavalier, and I have some research to do (and now that I think about it, review my distinction of where “refuse to concede” does more harm than good and maybe alter it a bit).

_________________

A while ago, I read an interview you gave, Scott, where you said something really powerful that stuck with me:

Which brings me to another aspect of “shtetl-optimized”: the grief over what was destroyed, and over the world’s indifference while it was happening. The Holocaust has been the central event in my mental life since I was probably seven. Like, if I see a digital clock that reads “9:43,” the first thought to cross my mind will be: “1943—over two million Jewish men, women, and children already lying dead in pits, but the Allies still could’ve bombed the train tracks to Auschwitz and Sobibor, had they wanted to…” In discussions about nuclear proliferation, global warming, and so forth, I never make the slightest apologies for being paranoid about humanity’s future, because those members of my extended family who weren’t sufficiently paranoid to flee to the United States were, as far as is known, all murdered. I never accept anyone’s assurances that “everything will probably work out for the best.” The question, for me, is never whether to be paranoid, but only which things to be paranoid about!

As a non-Jewish person, that was something I never really thought of or appreciated until I read that.

And one day it occurred to me that for many black Americans, it probably isn’t so different, complete with the fact that the overriding priority of the American president fighting the oppressor was not to end the nightmare. Except that their WWII was only partially successful, that their FDR/Churchill was murdered right before the war ended, that their first post-Reich president was a former concentration camp guard, that their government stayed mostly the same except with a few paragraphs amended, that included in those paragraphs was an explicit exception to the “no more Holocausts” rule , that they have no Israel, etc. (And I suppose that for many Native Americans it’s also similar, except that in their version they lost their WWII, and now the ghettos are optional and more of a cross between ghettos and enclaved Israels.)

6. Scott Says:

James Babcock #1: Yes, I’d heard about that, and didn’t pay nearly as much attention as I should have.

7. John Figueroa Says:

Will #4: I think one important question—or even the question—is how scalable a fraud attempt is. A remotely accessible vulnerability in a voting machine is very easily scalable—once you’ve found the exploit, you probably don’t need 1000 times as many people or 1000 times as much time to change 1000 times more votes. An attack based on bribing local officials or stealing things out of people’s mailboxes isn’t as scalable, I don’t think.

I don’t know enough about how mail voting works to know if there’s a scalable exploit based on bribing Postal Service workers or something.

Of course, one of the big issues here is that the winner-take-all nature of the Electoral College makes scalability matter less for the Presidency. Or for that matter (and more relevant to the Georgia case), the same goes for the winner-take-all nature of executive-branch races. I guess that’s another argument for making the legislature more powerful relative to the executive (compromising one House or state parliament race generally isn’t enough to compromise the House or state parliament).

8. Hash Says:

Voting, observe and counting process here:
Over 125,000 people trained for 48.000.000 voters and 195.000 voting box. Volunteer trained party members were reportedly deployed by political parties for each voting box (sometimes more than 10 different party member stay for observe and report results to their parties). They check for seals and count paper votes one by one. Eliminate invalid votes unanimously or party members can object. When counting over, report paper sign by clerk and party members. Party members take photos of voting count list, send it to their headquarters, so their system check if government numbers match.
They don’t leave votes. Clerk put counted votes in a bag and seal it. Party members follow bags to City Court (mostly soldiers or police protect votes). After election table accept the results and declare it “official”, volunteer members return their homes.
EVEN IN THIS SYSTEM GOVERNMENT PLAYING WITH NUMBERS AND STEALING THE ELECTION.. but at least we know this is our fault (we didn’t organize enough volunteer, or left boxes unattended or left votes early. Now system want adopt fully digital methods like in USA.

9. Jay L Gischer Says:

I’ve been using vote by mail in CA for several elections now. I think it’s solid. It isn’t perfect, but it is far, far better than the Diebold machines used in GA.

As we see from the description of the “negative votes” above, it’s better to not tweak the actual number of people voting. But the primary method of cheating in vote-by-mail does exactly that. Doing this in any kind of numbers is going to produce anomalies that are easy to spot. There’s a risk that your mail-in ballot using some registered voters name hits a conflict when that registered voter actually decides to vote. They check this, you know, both onsite and in mailin.

The biggest abuse of mail-in is when certain people, usually with multiple residences, vote in more than one location. This is a problem, and it’s hard to spot. But it doesn’t break in favor of Democrats, that’s for sure. But the numbers of this, as best as anyone can tell, are quite small.

And, you know, with sufficient infrastructure, you could do this kind of thing in person, too. Even with an ID. Just go to one place in the morning, hop on a plane, and vote at the other in the evening.

10. Scott Says:

Will #4: Briefly, my friends think that voting by mail, while not ideal security-wise (e.g., it has no way to prevent someone from coercing their spouse’s vote), is by far the best solution in the midst of a pandemic. And it can be made even better with receipts to verify that your vote was counted and other security features—indeed, Alex tells me that that’s a lot of what he’s researching right now.

11. Scott Says:

Doug S. #3:

Source for the Georgia election problem?

For starters, Wikipedia. Or Google, or the HBO documentary that I wrote about!

12. Oleg S. Says:

On a related note – Scott, what do you think is the best way to convince Bill Gates to run for president in 2020? Or is it too late?

13. Shmi Says:

You know, in some countries people actually want to have fair and free elections, even if those they vote for do not necessarily win. I have not heard of voting suppression and fraud being much of an issue in Canada, the UK and most of Europe.

What is it in the American psyche that makes the politicians to want to lie and cheat on such a grand scale? Gerrymandering, voting machine fraud, disenfranchising various parts of the population, accusing the political opponents (and foreign powers) of interfering in the elections… And it is not a new phenomenon, either, it dates back over 200 years. Why does America use the tactics more befitting dictatorships?

I think you should try watching Game Of Thrones or some other escapist entertainment once in awhile… real world too depressing. You’re going to give yourself a nervous breakdown…

15. Scott Says:

Adam Treat #14: Didn’t I read somewhere that Game of Thrones was a giant allegory for climate change? It sounded like it might’ve depressed me as well…

16. Scott Says:

Oleg S. #12: Unfortunately, I don’t think Bill Gates wants to run for president, and I think it would be extremely inadvisable if he did, since the sane vote would then be split between him and Biden.

17. Sniffnoy Says:

Gotta recommend against Game of Thrones. The book series, A Song of Ice and Fire, is still great, but the TV show… oy. The first four seasons, where they followed the books pretty closely, are pretty good, but after that, when they ran out of material to adapt… yeah, Benioff and Weiss don’t know how to write like GRRM. The result is not good.

And no, A Song of Ice and Fire (or Game of Thrones) is not an allegory for climate change. (I’m sure one can make the analogy, but “one can make an analogy” is very different from “it is an allegory”, and is true for many things.) Anyway the books are great, I definitely recommend them.

They’re adapting “The Plot Against America” into a TV series, huh? The book was distinctly disappointing. It seems like the plot has barely gotten started when suddenly it ends. Looks like they’re going for just a 6-episode miniseries here so I’m worried this will go the same way…

18. Jeffo Says:

If alternative history that explores racial conflict interests you, you should definitely make time for “Watchmen.”

19. Scott Says:

Sniffnoy #17:

They’re adapting “The Plot Against America” into a TV series, huh? The book was distinctly disappointing. It seems like the plot has barely gotten started when suddenly it ends

Not are, did! You can watch it. I agree with the criticism that it ends too early—they could have followed the trajectory of this world for 12 or 20 episodes, rather than wrapping everything up with an unconvincing deus ex machina right after the serious mayhem starts in episode 6. Even so, the creation of this world at all—and of the Levin family and its dynamics—seems to me like a significant literary achievement. And it was poignant to watch, two years after Philip Roth died, and as the United States was collapsing into a quasi-fascist state where millions rightly feared for their physical survival, not exactly in the way Roth predicted but close enough.

20. Koray Says:

I too recommend against starting watching GoT. “It is known” that the execution in later seasons is a total injustice to the early material from the books as well as superb acting by many actors and superb directing by various directors.

I do recommend Chernobyl from HBO, though.

21. Deepa Says:

I recommend the brilliant TV show from Denmark, called Borgen. Very well made. I barely noticed there were subtitles I was reading in English. Amazing plots, direction, and acting. About a female prime minister.

And what I’m currently watching: Swami and friends, based on a book by the same name, by R.K. Narayan. About the life of a 9 yo boy in India, pre-independence. Amazon prime has it with English subtitles.

22. asdf Says:

Kemp’s order was explained by someone on reddit:

hate to keep posting this but in case you missed it.

“heres the deal. Kemp mandates restaurants reopen, whether I reopen dining rooms or not. I file for business interruption insurance, it does not go through since I am ‘allowed’ to operate at full capacity. My landlord can demand all their money, since I am allowed to fully operate. Furloughed staff that is collecting unemployment insurance have to come back to work or I have to let them go. Their unemployment insurance then goes on my tab. If things blow up again they are still on my tab not on the states since they are no longer employed. Guys this is about screwing the working class and small business, not about helping us. Thank you so much to everyone that voted this malignant tumor into office.” John Gianoulidis.

23. asdf Says:

Meanwhile (should have included in above), Missouri is suing China for messing up coronavirus response. If that works, will the other US states be able to sue Georgia on similar grounds?

24. Job Says:

I can add: I’d often wondered how John had time to travel the world, giving talks about quantum supremacy, while also managing the lab’s decisions on a day-to-day basis.

Isn’t that more or less what you do? I thought you were also setting up a lab at Austin. You just delegate? 🙂

You there, bring up a quantum lab.

Potential analogies flood the mind: is this like a rock band that breaks up right after its breakout hit? Is it like Steve Jobs leaving Apple?

I do have an analogy for the current era in Quantum Computing, but it’s not a new one.

It’s like the moon landing. Or the decades that followed. You’ve landed on a QC, but now you have to establish a permanent presence there.

And each further iteration is even less tangible than the previous one, but somehow sees lowering public attention and you’re wondering how you can sustain the momentum without short-term commercial applications for the technology.

Are we going to live on a QC or not? I had already started packing, but now NASA’s eng are dispersing.

25. Scott Says:

Job #24: I lead a quantum computing theory group, not an experimental lab, much less one on the scale of Google’s. It’s extremely different.

26. Will Says:

Scott #10:

And it can be made even better with receipts to verify that your vote was counted and other security features—indeed, Alex tells me that that’s a lot of what he’s researching right now.

This is good to hear. I will try to read more.

Unrelated to security/fraud, another issue with mail-in ballots is that a lot of them don’t get counted. The entire mail-in voting process is a bit complicated, and there are a lot of opportunities for the voter to the mess it up:
* Signature doesn’t match the one on file closely enough? Vote doesn’t get counted
* Didn’t put the ballot in the secrecy envelope properly? Vote doesn’t get counted
* Forgot to put a stamp on the envelope? Ballot may or may not arrive
* Ballot sent by election day, but doesn’t arrive until after? May or may not be counted, depending on state.
Some of these can (and should) be fixed by state legislatures. But even with those fixes, a lot things can go wrong. I don’t have a source for this, but I’ve heard it is the case that all other things being equal (e.g. well-designed paper ballots), your vote is more likely to be counted voting in person than via mail (possibly significantly more likely, depending on state). Of course, other considerations apply during a pandemic.

Don’t get me wrong, I’m not against vote-by-mail. I live in Allegheny County, PA. For the nine years I have voted here, we’ve used these absolutely awful digital voting machines with no paper trail (fortunately, I think we have new machines starting this year). I always walk away from the voting booth with far too little confidence that my voted has actually been counted. But one year I voted absentee by mail, and I also did not feel a ton of confidence in the process. (One year I voted absentee in-person at the county election office. No particular complaints about that.)

But I just think, that outside of pandemic conditions, we should have a little bit of skepticism toward vote-by-mail.

27. X Says:

I read an intimation that the anti-Martinis faction is also the pro-DWave faction. Do you have any insight as to how true that might be? If Google pivots away from high-quality qubit work back to DWave-style whatever-it-is-they’re-doing, that seems like a serious blow to the future of practical QC.

28. Scott Says:

X #27: Does Google still even have a “pro-D-Wave faction”? I don’t know, and the issue was mentioned by neither Martinis nor Neven when I spoke to them.

29. Paolo G. Giarrusso Says:

Probably stupid question Scott: any chance that NP \subset BQP can be shown in 5 pages, in particular by https://arxiv.org/abs/2004.11347 ?
Found it from https://twitter.com/lorensipro/status/1254317055093989376. I don’t have a clue and my prior is “no”, but the article seems short (and clear(?)) enough to be judged on its merits.

My silly plausibility meter is very inconclusive:
– At least they’re not cranks — at least one author is an actual cited quantum computing academic: https://scholar.google.nl/citations?hl=en&user=GO7V6ZYAAAAJ&view_op=list_works&sortby=pubdate
– IIUC (from https://blog.computationalcomplexity.org/2007/02/on-np-in-bqp.html), this result is “_almost_ impossible” due to a relativizing lower bound (Bennett, Bernstein, Brassard, Vazirani 1997, https://arxiv.org/abs/quant-ph/9701001), but they they do cite that paper correctly.

30. Scott Says:

Paolo G. Giarrusso #29: Good that they cite BBBV, I guess, but I glanced at the paper and it still looks like it’s clearly, obviously ruled out by BBBV, and like it joins 25+ years of similar naïve attempts to put NP in BQP. BBBV just completely slaughters all this stuff. For it even to be worth my time to look further at this, someone would need to explain to me how the authors propose to get around BBBV—specifically, what structure do they exploit in NP-complete problems?—rather than merely asserting that BBBV is not a barrier when everything here screams to me that it is.

31. Mean Southern Sergeant Says:

I see in the thread the argument for bombing the rail lines to the death camps. The big death camps were in Poland and had except for Auschwitz shut down by late 43
8th Air Force had a tough time getting even to Berlin and back without fighter support, which was not available until early 1944.
17 Aug 1943 was the dual raid on schweinfurt and regensburg, unescorted by fighters,it cost us 60 B17s, each with a crew of 10 highly trained airmen.
I think those directing strategy knew of the suffering in Europe and made the decision that the best way to help was to put a few million men into France and start killing Germans, and directed their efforts to that end

32. Michael Says:

Scott, Tom Cotton talked about how Chinese students shouldn’t come to America to learn quantum computing:
https://www.forbes.com/sites/alexandrasternlicht/2020/04/26/senator-tom-cotton-ramps-up-anti-china-rhetoric-says-chinese-students-should-be-banned-from-us/#5f4c89d299a2

33. Anonymous Says:

Mean Southern Sergeant, 31

Please, the British were bombing Berlin as early as 1940, with poor objective results and The head of the Air Staff of the RAF, Sir Charles Portal, justified these raids by saying that to “get four million people out of bed and into the shelters” was worth the losses involved.

Really if the locations of the camps in Poland contained strategic military industry they would have been reached one way or another.

34. Gerard Says:

@Scott #30

“what structure do they exploit in NP-complete problems?”

It certainly doesn’t appear that they do that in an way.

The first sentence of the abstract says

“In this work, we present a multi-layer quantum search method that generates an exponential
speedup of the standard Grover’s algorithm.”

They then imply that their algorithm only works for NP problems because only in those cases can the multi-layer Grover oracle they use be constructed efficiently.

However the actual application of the algorithm to an NP complete problem, specifically SAT, occupies a single paragraph in which the particular SAT instance is only represented by the variable E. Moreover the equations in this paragraph look identical to a previous paragraph where they describe the use of the algorithm for a function f (that they call an oracle). The only difference is that in the second paragraph E has been substituted for f. Hence though I don’t know enough about how the Grover algorithm works to understand the rest of the paper, I’m fairly confident that they do not meet the criterion you specified.

35. Bill Kaminsky Says:

Here’s what I’m 99% sure is the explicit fatal flaw in the NP \in BQP paper [1] mentioned by Paolo G. Giarrusso (Comment #29) and replied to by Scott (#30) and Gerard (#34).

**Full Disclosure:** I’ve done the exact same idea with the exact same fatal flaw. Heck, I even pitched it to Seth Lloyd, who I notice the authors of [1] thank in their acknowledgements. I can’t speak for Seth, but I can only assume they didn’t make it clear to Seth what exactly they were doing, because Lord oh Lord, Seth and (especially) I have spent time making sure this fatal flaw is truly as fatal as it first looks. (Spoiler alert: It is! Less of a Spoiler Alert: Seth was sure at the outset I’d eventually convince myself that it was!)

**Perhaps-Needed Context:** The authors in [1] speak of “multi-layer quantum search”.
What’s that in comparison to good ol’ Grover search? Well, their idea essentially is that if you have a constraint satisfaction problem with $C$ clauses, then instead of running your favorite rendition of Grover’s algorithm [e.g., 2, 3, or 4] with an oracle doing a conditional phase rotation on computational basis states encoding satisfying assignments of *all* $C$ clauses, run it first with an oracle just involving 1 clause. And then run it again with an oracle involving 2 clauses. And so on and so on, until you’ve done all $C$ clauses. Voila! NP \in BQP!!

**The Fatal Flaw:** Implementing oracles that recognize only a subset of the clauses is definitely doable. HOWEVER, it’s NOT the operator involving those oracles in Grover-style algorithms — namely the conditional phase flips of the computational basis states that represent satisfying assignments of some desired set of clauses — that’s your problem in seeking to do “multi-layer” Grover in such one-clause-at-a-time way.

INSTEAD, it’s the other operation in Grover-style search that’s the insuperable obstacle, namely: the “inversion about the mean” operator in traditional Grover [2] or, more generally [3, 4], the conditional phase flip on *the projection onto a uniform superposition over computational basis states representing the satisfying assignments of the first $c < C$ clauses.* You can't approximate that operation with sufficient accuracy to not have errors — which will blow up *exponentially* as you go from "layer" to "layer" in the multilayer search! — screw up your final result.

Think about it: given 1 copy of an *unknown* quantum state $\ket{\Psi}$, how do you implement the $\ket{\Psi}\bra{\Psi}$??.

I mean sure, if you have $N$ copies, you can do a weak SWAP-gate-based measurement to *approximate* the projector, but that'll still leave a $O(1/N)$ error. And near and dear to Scott's heart, given a reasonable number of copies, you could do "shadow tomography" [5] to approximately ascertain a polynomial number of local expectation values in an efficient manner. And you may then think for a second, "Great! THAT'S THE STRUCTURE I'M GOING TO EXPLOIT TO AVOID THE BBBV [6] argument. k-SAT has clauses, and thus by learning those clause-related expecation values I'm NOT doing unstructured search." But, not so fast! Shadow tomography — or whatever way you use to get those expectation values — still has errors. These errors will lead to errors in the needed projectors onto the uniform superposition of computational basis states that are satisfying assignments of the first $c < C$ clauses. And, once again, *you're going to exponentially blow up those errors* by doing *multi-layer* search with the approximate projectors you form from the knowledge of the expectation values you've generated by repeatedly running the algorithm up to some $c < C$ stages.

References:

[1] Shan Jin, Xiaoting Wang, Bo Li, "Multi-layer quantum search and inclusion of NP into BQP", https://arxiv.org/abs/2004.11347

[2] For "traditional" Grover's algorithm, see for example, Chapter 6 of Nielsen and Chuang's textbook.

[3] For the original version of a "fixed point" Grover's algorithm, see for example, L.K. Grover, "A Different Kind of Quantum Search", https://arxiv.org/abs/quant-ph/0503205

[4] For the 2014 optimized version of [3], T.J. Yoder, G.H. Low, and I.L. Chuang, "Fixed-point quantum search with an optimal number of queries", https://arxiv.org/abs/1409.3305

[5] For Scott's "shadow tomography", well… what reference do you prefer, Scott? I assume S.J. Aaronson, "Shadow Tomography of Quantum States", https://arxiv.org/abs/1711.01053, yes?

[6] Last but not least, this "BBBV" no-go theorem that's always worth remembering! Well, Chapter 6 in [2] does have you covered, but you can go to the source: C.H. Bennett, E. Bernstein, G. Brassard, U. Vazirani, "Strengths and Weaknesses of Quantum Computing", https://arxiv.org/abs/quant-ph/9701001

36. Hendrik Says:

I think the problem with the BQP-in-NP paper is that after the first layer, the actual solution t is disconnected from the rest of the wave function, whenever the first bit of t is zero (all other nonzero terms have their first bit set to 1). Which means, you don’t have any way to provide the additional constructive interference that is needed to boost the probability to find t. From that point on, the rest of the algorithm is just LOCC in disguise.

37. Craig Gidney Says:

Paolo G. Giarrusso (#29): The red flag in this paper is that this algorithm is not complicated, and its claimed efficiency is high, but the authors didn’t implement it and simulate it to verify that it works. The authors are not sufficiently paranoid about finding a trivial solution to a famous problem.

Based on my quick skim, I have two candidates for what the problem might be. First, I think they are assuming that adding one clause to a 3SAT problem will only reduce the number of solutions by some reasonable proportion, so that they can gradually refine the search. But there are situations where adding one clause can finish off a loophole and make the number of remaining solutions drop by arbitrarily large amounts. Second, I think they are using an operation to reflect about the superposition state of all partial solutions but not properly accounting for the cost of that operation (e.g. each time it is used it requires a recursive call, and the resulting recurrence relation for the total cost has an exponential blowup in it).

38. luca turin Says:

When I last visited Google (LA) to give a talk after a few years’ absence, I felt that the actors were just as smart as ever but the vibe was different, weary and bureaucratic. Came away feeling the party was over. Shame: they almost were this century’s Bell Labs.

39. gentzen Says:

Looks like the reassignment to an advisory role for Martinis happened long after the proclamation of Neven’s law. So it should have nothing to do with it. And I guess it is also totally unrelated to the “surprise” I expressed in a comment after watching a panel/debate on quantum supremacy moderated by Sandy Irani and featuring Scott Aaronson, Dorit Aharonov, Boaz Barak, Sergio Boixo, Adam Bouland, Umesh Vazirani, and Gill Kalai:

Watched the panel/discussion and was surprised that google does not plan to demonstrate error detection first, before shooting for error correction. Demonstrating error detection, and repeating a given computation until no error gets detected (with as many qubits as the hardware allows) seems like an important and interesting experiment to me. Especially the remaining error distribution for the runs where no error was detected will be very interesting.

I do see that google’s preprint Hartree-Fock on a superconducting qubit quantum computer mentions “We also demonstrate error-mitigation strategies based on N-representability which dramatically improve the effective fidelity of our experiments.” So maybe they did something even better, and I would have to read that preprint to understand details.

40. Paolo G. Giarrusso Says:

On NP ∈ BQP, I think I found a gap, but it’s different from what most comments suggest (except maybe Hendrik #36?); something interesting but not violating BBBV might be left.

Answering to #34, #35 and #37. Bill Kaminsky #35: they’re not using the exact idea you describe, tho the same issue might apply. Craig Gidney #37: They’re _not_ adding a (CNF) clause, but a DNF clause. Gerard #34: thanks for pointing to that section!

I understand well the non-quantum part of this paper, which #30 pointed to, but less of the quantum part (let’s say I took undergrad quantum computing years ago). And I think I found a specific non-quantum gap there that ensures BBBV is respected: to decide a problem in NP you must answer to problems which have _no_ solution, in which case their last search layer breaks down — in their notation on the last page, 𝜆_{n+1} = 0 instead of 𝜆_{n+1} >= 1/2. If a solution exists, for all I see, they might still find it faster, tho that still sounds unlikely.

In short, taking their claim at face value, they take an oracle f : {0,1}^n → {0,1} and searches solutions x such that f(x) = 1. Since Grover is faster when there are more solutions, they use n intermediate functions with extra solutions, called {f_k}, k=1..n+1, with f_{n+1} = f. Solutions of f_k must either solve f, or start with k bits equal to 0, or both.

Then, their algorithm starts from a state superposing all candidate solutions, and refine it by applying the composition of k layers of Grover search: the k-th Grover layer “cuts out” candidates that do not solve f_k, but the fraction of surviving candidates is 𝜆_k with 1/2 <= 𝜆_k = 1/2 makes the k-th Grover search layer fast (I double-checked that 𝜆_k >= 1/2 for k = 1..n). In the end, f_n has at most one extra solution, and f has no extra solutions.

Here’s the gap: if f had no solution, this clearly doesn’t help — indeed, in their formalism 𝜆_(n+1) = 0, so the minimum amount of queries for the n+1 layer diverges according to their formulas; if you terminate early, you should get nonsense. And if you skipped the last Grover step and measured, you’re likely to find an extra solution.

As anticipated, if a solution exists, maybe they do find it quickly. That sounds too good to be true, but it’s not even clear that this violates BBBV, so I’d leave that part to the expert.

41. Gerard Says:

@Bill Kaminsky #35

” run it first with an oracle just involving 1 clause. And then run it again with an oracle involving 2 clauses. And so on and so on, until you’ve done all $C$ clauses.”

That does not appear to be what they are doing however. Doing so would require introducing a notation that allows picking individual clauses out of the SAT formula. As I said in my previous comment all their expressions only involve a single variable E representing the entire SAT formula. Perhaps you are thinking that the layer expressions they give in the unlabelled equation following equation (16) introduce the clause expressions. However if you look closely you will see that each layer just contains E OR’d with another boolean expression which is constant (ie. it depends on the layer index but not on the SAT formula itself, E).

The only effect of the the extra expression in each layer function is to add additional solutions which are completely independent of the problem instance itself. So even without considering Scott’s objection related to relativization it would be quite surprising if such a method actually resulted in a more efficient solution (at least from the point of view of someone with no quantum computing intuition).

42. Scott Says:

Everyone: In some sense, there would be no point in proving lower bounds, if even after proving them, you still needed to individually examine each and every algorithm that was claimed to violate the lower bounds! 🙂 Feel free to use this comment section to continue trying to pinpoint the bug though.

43. Craig Gidney Says:

Gentzen #39 (with respect to error detection): Using error detection to extend a computation is not a very effective strategy. Even an absolutely perfect error detecting code, that detected every error with no overhead, would buy you at most one order of magnitude. If your per-operation error rate is 1/n with n > 100, then the chance of performing 10n error-free operations is roughly one in twenty thousand. Not so bad. But the chance of performing 100n error free operations is roughly one in ten million trillion trillion trillion. Not gonna happen.

Of course, in practice error detecting codes do have overheads and these eat into their benefits. Operations that were unitary suddenly require measurements (easily the noisiest operation on a superconducting device). Operations that used to just be a single physical rotation will not be native to the code and require a long series of approximating operations and/or magic state distillation.

I think the best uses of error detecting codes in the NISQ regime will be zero-overhead things like “I know my chemistry computation is supposed to preserve the number of electrons, which corresponds to the number of ones in the computational basis state, so I will post-select on the output having the correct number of ones.”.

Another way that error detection can be useful is for increasing the reliability of short repeatable computations, such as magic state distillation and entanglement distribution. But this is only really relevant in the context of a much larger computation, so it’s very relevant to an already fault tolerant computation but not so relevant to a NISQ computation.

44. gentzen Says:

Craig Gidney #43: Thanks for your detailed answer. I didn’t expect error detection to be useful to extend a computation. My surprise is more that demonstrating error detection seems to me like a natural step towards demonstrating error correction. Much of the overhead should be of a similar nature as the overhead encountered for error correction. But you can work with smaller codes, and you don’t need a high speed (and high quality) feedback loop between measurement outcomes and error correcting control action. And at least in theory, already the act of measurement itself (even without postselection) is supposed to somewhat reduce the error rate (quantum Zeno effect).

So it looks like an interesting experiment to me. Probably that experiment has already been done before with less powerful quantum computers, maybe even by Martinis’ group itself. So it might not be such a good experiment after all, since the publication of the results would just be old news.

45. Bill Kaminsky Says:

To everyone,

Bin, Yang, and Li have withdrawn their NP \in BQP claim. https://arxiv.org/abs/2004.11347 now takes you to a version 2 which doesn’t even have a PDF available, instead just being marked with comment “We found an error in the proof. The main conclusion does not hold.”

To Gerard (comment #41):

Thanks for the correction. I honestly read their paper in a very quick way, first trying to find a nice discussion to that’d be somehow equivalent to:

“Y’all not going to believe this, but there’s a simple-enough-to-prove-in-5-pages way to circumvent the famous no-go theorems about Grover’s algorithm being optimal for unstructured search. And amazingly, this circumvention involves running Grover’s algorithm recursively in many ‘layers.’ I know, I know… many of y’all might have secretly tried that in your spare time scribbling proofs NP \in BQP in the past. But here’s the intuition for the technical innovation that makes it work!”

Not finding that, I honestly then just jumped immediately to their summary pseudo-code for their algorithm and then skimmed the rest of the paper for all of 3 or 4 minutes trying figure out their their notation and novel ideas. After those 3 or 4 minutes, I must confess I then skipped exploring any further as to what their explicit approach was. Instead, I looked to their quick invocation of “fixed-point” recursive Grover’s algorithm in their Introduction and thought to myself, “I bet they blithely assumed these projectors $\ket{\varphi} \bra{\varphi}$ are efficiently implementable, and man… been there, done that… they are NOT. 🙄”

46. Bill Kaminsky Says:

AAAH! Now “mea culpa, mea culpa, mea maxima culpa” for real! I misremembered the author’s names as Bin, Yang, and Li. Their names are *Jin*, *Wang*, and Li.

47. Christian Says:

@gentzen #44:
Repeated quantum error detection was recently demonstrated in a small surface code architecture: https://arxiv.org/abs/1912.09410
(full disclosure, I am co-author on the paper)

We think this is an important step towards quantum fault-tolerance, but, as mentioned by Craig Gidney #43, using quantum error detection to actually run an algorithm better is hard as you will end up throwing away most of your data. Take a look at our fig 5 (c) for example. Of course if we would have better better fidelities, we would detect less error and keep more data etc, but in the end, you need to be able to correct your errors to really win the game of fault-tolerance.

48. E. Harding Says:

Yeah, so about that. You know Brian Kemp, the doofus governor of Georgia, who’s infamously announced plans to reopen the state right away, ignoring the pleading of public health experts—a act that will fill Georgia’s ICUs and morgues as surely as night follows day? And you know how Kemp defeated the Democrat, Stacey Abrams, by a razor-thin margin, in a 2018 election of which Kemp himself was the overseer? It turns out that Kemp’s office distributed defective memory cards to African-American and Democratic precincts, though not to white and Republican ones. There’s also striking statistical evidence that at least some voting machines were hacked, although because there was no paper trail it can never be proved.

Every bit of this is stupid and wrong. Georgia opening up will not lead to more deaths; there is no evidence to suggests lockdowns work at reducing the spread of the infection at all, given the current catch-and-release policy for the infected the states in lockdown have inexplicably adopted.

Likewise, every single conspiracy theory about Georgia’s 2018 election is wrong. Every one. They are blatant lies. 2018 was the first election held under universal voter registration in Georgia, it had among the highest turnouts of any state, and had the least White electorate in its history (not that elections are a good way to decide who the governor should be). Not merely its midterm history -its entire history. Abrams lost because she did much worse among working-class Whites and Blacks than Hillary Clinton, who herself did much worse among working-class Whites and Blacks than Barack Obama. Full stop. Georgia is not a majority-nonwhite state, nor does it have a majority-nonwhite voting age population (though, again, deciding who the governor should be by election is silly, and caring about tiny margins of fraud is even sillier -it honestly doesn’t matter who gets more votes if an election is very close).

n short, what The Plot Against America and Kill Chain have in common is that they would be desperately needed warnings about the ease with which democracy could collapse in the US, except for the detail that much of what they warn about has already happened, and now it’s not clear how we get back.

Why would you want democracy to survive in the United States? Why would you want this clown show to continue? I don’t want democracy; I want government by superforecasters. I think everyone should.

49. Michael Gogins Says:

E. Harding #48:

Democracy is the only system of government in which the allocation of political power to individuals is just.

In other words, if you want any share at all of political power, and you will be born as some person at random with respect to race, class, gender, and so forth, then as in John Rawls’ initial condition thought experiment, only democracy is just.

50. E. Harding Says:

“In other words, if you want any share at all of political power, and you will be born as some person at random with respect to race, class, gender, and so forth, then as in John Rawls’ initial condition thought experiment, only democracy is just.”

That’s called randomly selecting senators, representatives, judges, bureaucrats, governors, and the president from the general population. Which is much truer democracy than American-style “democracy”, in which only the media-famous and those reliant on political establishments (i.e., not you) can get elected, and congress, the executive, etc. still remain wildly unpopular. Heck, even Russian or Indian-style democracy is truer democracy than American-style “democracy”, because India’s and Russia’s leaders have much higher approval ratings than Trump does.

act that will fill Georgia’s ICUs and morgues as surely as night follows day

This is a testable prediction! Care to put some numbers on it so we can check back in a month or two and see how right you were?

52. Joshua B Zelinsky Says:

@ E. Harding #48

“Every bit of this is stupid and wrong. Georgia opening up will not lead to more deaths; there is no evidence to suggests lockdowns work at reducing the spread of the infection at all, given the current catch-and-release policy for the infected the states in lockdown have inexplicably adopted.”

Are you actually arguing that policies which prevent people from interacting are not going to reduce the spread of a disease? What is your logic here?

“Likewise, every single conspiracy theory about Georgia’s 2018 election is wrong. Every one. They are blatant lies. ”

Labeling things conspiracy theories or lies isn’t productive if you aren’t going to give any evidence for that. And higher non-white turnout isn’t evidence against any of the claims in question. That’s a complete non-sequitur.

53. Nick Says:

E. Harding #48

> Likewise, every single conspiracy theory about Georgia’s 2018 election is wrong. Every one. They are blatant lies.

Brian Kemp was Georgia Secretary of State while running for governor, overseeing the election in which he himself was a candidate. He did not resign, despite the flagrant conflict of interest. Is that a blatant lie?

54. Scott Says:

Everyone:

One of history’s darkest Jewish jokes is about a Jew in Berlin in 1936, sitting on a park bench reading Der Sturmer. When his friend sees, he runs up and says “what are you doing?? Don’t you realize that’s a Nazi paper?”

“Yeah, I know,” comes the reply, “but whenever I read the Jewish paper, it’s all too depressing. Whereas here the news is great! I’m learning that we control the banks, we control the media…”

I’ve noticed a similar phenomenon with myself and covid. Whenever I read the mainstream news, I’m overpowered with despair: all-cause mortality in NYC is up by a huge factor; ER docs are committing suicide; people are being turned away from hospitals and dying soon afterward; the virus attacks not only the lungs but the heart and brain, even in the survivors; lockdowns are ending even though there’s still no plan for mass testing, contact tracing, or isolation, the President’s solution is that people should drink bleach … all of which has given me an urge to switch to RealClearPolitics and the other hard-right sites, where I can read the wonderful news that the worldwide lockdowns were all a gigantic overreaction bordering on a hoax, the only people seriously affected were on death’s door anyway, we’re already close to herd immunity, life can go back to normal as soon as we snap our fingers… 😀

55. Edward Says:

I think the following analysis is useful for the Georgia situation — it presents both relevant links supporting Scott’s points and the best counter-arguments I’m aware of. (Even though I’m not usually a big fan of “fact checking.”)

https://www.politifact.com/article/2019/nov/21/no-proof-voter-suppression-kept-stacey-abrams-gove/

56. E. Harding Says:

“the worldwide lockdowns were all a gigantic overreaction bordering on a hoax”

They were. They were part of absolutely noone’s pandemic playbook prior to Communist China’s attempt at it (correctly dismissed at the time as political posturing). And even that didn’t get the R0 below 1 (the R0 in the U.S. is below 1 currently both in lockdown and non-lockdown states). States without lockdowns have shown no faster recorded COVID death growth than states with lockdowns. In contrast, R0 in Japan was, due to widespread pre-existing social distancing, low enough to keep deaths to well under 1000. Not under 1000 per day. Under 1000 total. The fact so many White countries failed to learn from Asian countries and adopted the most economically disruptive and least effective features of Communist China’s response is perhaps the strongest possible evidence in favor of the Social Justice narrative in regards to the reality of White racism. To millions of American Whites, both educated and not, Korea, Hong Kong and Taiwan might as well be Detroit, Philadelphia, and the Bronx.

“the only people seriously affected were on death’s door anyway, we’re already close to herd immunity, life can go back to normal as soon as we snap our fingers…”

All this is sadly untrue, of course, no matter how much Republicans might say so.

57. Joshua B Zelinsky Says:

@E. Harding # 56

“They were. They were part of absolutely noone’s pandemic playbook prior to Communist China’s attempt at it ”

Yeah, no one was discussing anything similar. That’s why it doesn’t show up when they quarantine entire cities in Contagion https://en.wikipedia.org/wiki/Contagion_(2011_film) or an even earlier film from 1995 https://en.wikipedia.org/wiki/Outbreak_(film) . Yeah, I guess this idea wasn’t already so well known that it had already entered the popular conscience in multiple Hollywood blockbusters.

“And even that didn’t get the R0 below 1 (the R0 in the U.S. is below 1 currently both in lockdown and non-lockdown states).”

I suspect you mean Rt here; this is a common confusion (and even some scientific papers don’t make the distinction). This is in any event, false https://rt.live/ . As you can see from the graphs there, Rt > 1 for 8 US states at present, including multiple states without substantial lockdowns (like Iowa). And many other states by their estimate have Rt extremely close to 1. For example, they estimate that Arkansas currently has an Rt of 0.99, and if you take that degree of precision in the data to be meaningful then we have other issues we need to discuss. But it is also important to look at the graphs there and their shape over time. By the estimates there, every single US state four weeks ago had an Rt >1, with some substantially higher. The data does not support your claim.

58. Bunsen Burner Says:

I wonder if there is a pragmatic way to accommodate the idiots who thing Lockdowns are a Chinese Plot Against The Bodily Fluids Of ‘Muricans. Suppose we let them out of Lockdown but only if they accept that if they get sick for any reason or suffer an accident neither they nor their families will receive hospital treatment? Furthermore, if they pass the virus onto someone who is in Lockdown or is an essential worker then they must pay for their treatment, and of course, if that person dies they are charged with manslaughter? If there is a pragmatic way of doing this it does look like a win-win scenario for everyone?

59. Gerard Says:

@ E. Harding #56

“(the R0 in the U.S. is below 1 currently both in lockdown and non-lockdown states”

I don’t know much about epidemiological models but if, as I suspect, that would imply that exponential growth is no longer occurring in the US then I don’t think the data support that conclusion.

I have been tracking covid death counts (I suspect they’re more reliable than infection counts, though I also think many covid deaths are not being reported in these statistics) in the US, several US states and a few European countries.

In order for exponential growth to be invalidated the daily growth rate (daily deaths/total deaths to date) must approach zero. I have yet to see that number fall significantly below 1 % (corresponding to a doubling time of about 70 days) in any of the locations I track (though in some places, like Italy, it has been close to that number for some time), For the US as a whole the minimum was 2% on April 26 and it is currently at about 4%.

As for drawing conclusions about the effectiveness of different policies from the extreme variations in death counts across different regions, I think that is very hazardous at this point. Much is still not known about the virus and there are many possible explanations for such differences besides or in addition to policy differences. In the case of Asia versus the Western World there could be differences in the genetics of the virus or host populations as well as in the prior immunological state and general health of the host populations, just to cite a few possible variables.

60. Scott Says:

Bunsen Burner #58: Interesting idea, but I think you and I both know perfectly well that the anti-lockdown protesters would never accept such a deal. For them, freedom from consequences for harm caused to others is the entire point!

Consider: Mitch McConnell’s priority right now is not testing, or contact tracing, or PPE for health workers, or vaccine research, or anything else like that, but indemnifying companies against any possible coronavirus-related lawsuits, to the point where he’s holding aid to the states hostage to that goal. And he’s moderate compared to the anti-lockdown activists. Do you imagine that anyone with such a mindset would agree to pay the medical costs of anyone they infected (even in the rare case that the chain of transmission could be proved)?

61. Bunsen Burner Says:

Scott #60

Oh yeah, I doubt they would accept those terms, though it might have the positive effect of humiliating them into silence. I also should have added that any business that insists on staying open must force the CEO and the entire senior management team to sit with the normal workers every single day.

But yes, it’s a pipe dream. The political system is so corrupt that human life really means absolutely nothing to these people. Profits, profits, and profits!

62. gentzen Says:

Christian #47: Thanks for mentioning your (co-authored) paper. Since I said that it would be an interesting experiment to me, I had to read your paper to satisfy my curiosity. And the paper provides exactly those postselected statistics that I wanted to see. It is nice to see the differences in fidelity and error rates for the logical |0>, |1> |+> and |-> states. I also really like that the experimental data was always compared to expectations from simulations.

However, one conclusion from reading your paper is that Google publishing results from experiments with error detecting surface code would definitively not “just be old news”. I learned now that there were previous demonstration of error correction and detection, but they only detected (and corrected?) a subset of the possible errors (for example by using repetition codes). On the other hand, your paper does not yet demonstrate operations on the encoded logical qubits. So I still hope that Google or another player with a quantum computer with many high quality qubits will demonstrate quantum operations on encoded qubits with error detection and postselection, before trying to demonstrate quantum operations on encoded qubits with error correction.

63. E. Harding Says:

@Joshua B Zelinsky

Why are you linking to movies rather than to governments’ pandemic playbooks? Really?

This is in any event, false

No; true. I didn’t say “all states”, I said “states” (as a group).

By the estimates there, every single US state four weeks ago had an Rt >1, with some substantially higher.

I’m not talking about the past, I am talking about the present.

I don’t think the data support that conclusion.

I do think the data support that conclusion. Recall, case growth was 28K yesterday; it was higher in early April with much fewer tests. Obviously, growth is still much too high due to governments’ refusal to institute quarantine (which actually was in governments’ pandemic playbooks prior to 2020) and failure to learn from East Asia in general.

I also should have added that any business that insists on staying open must force the CEO and the entire senior management team to sit with the normal workers every single day.

I don’t understand why people wish others to be poorer for no good reason. Does Japan have more coronavirus deaths than New York? Does Hong Kong have more coronavirus deaths than Kentucky? Does South Korea have more coronavirus deaths than Massachusetts? Does Taiwan have more coronavirus deaths than Washington? Learn from countries with effective policies that have not destroyed their economies, not those that have destroyed their economies for two months for absolutely no good reason.

64. E. Harding Says:

For the record, Scott, a fascinating account, completely impossible to believe today, of the West teaching the Chinese that yes, fighting infections using centralized quarantine is, in fact, possible:

At the beginning there was general disbelief in the necessity or usefulness of preventive measures. It was an absolute novelty to the Chinese mind to attempt to check the spread of any infection, and apathy naturally accompanied their fatalism. “This is the scourge of Heaven” said many. “All will die whose time has come, and no others. Then why take people away to isolation stations? Why burn good clothes and bedding?”

Interference with personal liberty was strongly resented, and still more the disturbance of trade and business. When a shop was forcibly closed and disinfected, and twenty-nine persons removed from it to an isolation station because of the death of a thirtieth, the merchants were highly incensed. The co-operation of the general public could thus hardly be expected. When the house-to-house visitation began it caused much fear. It was said that every sick person was to be removed, and those who had been ill for weeks struggled to rise and present a cheerful front to the unwelcome intruders. As days went on and no terrible results followed from the police inspection, it came to be welcomed by many as a kind of official certificate of health and protection from Plague.

I won’t link to it due to the spam filter, but you can find the book with a simple search.

65. Sniffnoy Says:

Michael Gogins #48:

The problem here is, what’s meant by “democracy”? Yes mechanisms must be symmetric — no kings or other hereditary positions — but beyond that…? Futarchy is democratic in this sense, but otherwise is very different from what’s currently known as democracy.

I think thinking of governance systems as if it’s a question of power distribution rather than aggregation mechanisms is a mistake. It’s all well to say “power to the people”, but “the people” is not a person, so you still need some aggregation mechanism, some mechanism that actually outputs the decisions. Direct democracy, representative democracy, liquid democracy; the use of FPTP vs cloneproof systems; futarchy vs having legislators directly make policy… these are all quite different things. Which one of these is truly the people speaking? None of them; in all cases they’re just inputs to the decision mechanism. Power to a good — and, as fairness requires, necessarily symmetric — decision mechanism!

66. Scott Says:

E. Harding: It’s true that in the US, our government seems determined to give us the worst of all possible worlds—a lockdown, followed by a premature end to the lockdown, followed by hundreds of thousands or millions of deaths that the lockdown simply delayed by a couple of months, thereby showing that the lockdown simply burned trillions of dollars for next to nothing. But (sigh) what was supposed to happen, what would’ve happened in a functioning country, was that the lockdown would be a way to buy time for a WWII-scale mobilization of testing, contact tracing, and away-from-home isolation of anyone infected.

To put it differently: we’re all trapped in a car with a drunk driver at the wheel. The medical professionals tackled the driver as he was about to veer off a cliff, a desperation move that bought a few extra minutes. If the driver ends up veering us off the next cliff anyway, then was the medical professionals’ act pointless and stupid?

67. Joshua B Zelinsky Says:

@ E. Harding # 63

“Why are you linking to movies rather than to governments’ pandemic playbooks? Really?”

The point here is how absolutely well known this sort of thing was a response type was that its even in Hollywood depictions. But if you prefer, one can look at different pandemic response exercises in the past. Crimson Contagion which was just last year included lockdowns as a possible response measure. Or if you prefer, look at this 2005 copy of the CDC’s pandemic response planning https://www.cdc.gov/flu/pandemic-resources/pdf/pandemic-influenza-strategy-2005.pdf which includes in bullet points both “Encourage all levels of government, domestically and globally, to take appropriate and lawful action to contain an outbreak within the borders of their community, province, state or nation.” and ” Where appropriate, use governmental authorities to limit non-essential movement of people, goods and services into and out of areas where an outbreak occurs. ”

I agree that other aspects have not been followed, and you are correct to identify a lack of following some quarantine procedures as one of the major differences between the current response and prior plans.

“No; true. I didn’t say “all states”, I said “states” (as a group).”

My apologies; I didn’t read your statement as charitably as maybe I should have. It is true that most states have Rt below zero at this point, but it certainly isn’t all. It also is the case that estimates for Rt are generally lower in states with more strict lockdowns. (Roughly speaking; the data here is very noisy.)

68. E. Harding Says:

Interregional travel restrictions are not a lockdown, as people can still generally go to work instead of staying at home. U.S. states have rarely, if at all, implemented interstate travel restrictions in regards to the pandemic, though they have widely implemented lockdowns. China restricted internal travel. The U.S., for whatever reason, mostly never did.

69. Christian Says:

gentzen #62: thank you for the comment and I am happy to hear that you enjoyed the paper. I completely agree with you, that there is still plenty of things to explore within a quantum error detection code especially if you can implement a larger distance code. I expect that such experiments will be explored in the next couple of years mostly by academic labs

70. asdf Says:

71. Ethan Says:

Martinis explains. It seems to me that this is the n-th example of Google’s tendency to attract arrogant technical people backfiring. I see a lot of humility (and class) in Martinis’ comments. Knowing what you don’t know is one of the best qualities I have seen in the great people I have worked with over the years (once one gets past things like intelligence, competence, etc that I always see as necessary but not sufficient conditions for me when I consider close collaborations).

“PSG:

Yes, I completely understand. What was competing for control of the group at the time? Was it anything in particular or was it just that the group wanted more freedom?

Professor Martinis:

Yes. I’ll give you a specific example that happened last year. It was very difficult for me at the time.

When wiring a qubit system, it is very important to figure it out in a scalable, cost-sensitive way. I had been developing some new technology that was quite far along. I was quite proud of various inventions, as solutions were somewhat tricky, and you had to think very carefully about how to build a system. Wiring is funny as it is something that everyone thinks they can do, but it’s subtle. But one theorist, who did not have experimental experience, felt that he knew how to do wiring better than me. We talked and I said look, this isn’t going to work, and you should not be doing it. But he didn’t accept no and kept working on it.

This went on and on, so I started talking to Hartmut about it, explaining that it didn’t make sense. We have a relatively small group and we don’t have an abundance of resources. And I have it covered, so we should stop this other program. But Hartmut didn’t back me up and wanted to go ahead and try both.

I guess I just didn’t handle that very well. I was the lead of the hardware group, but I couldn’t stop a project that didn’t make any sense to me.

PSG:

Did you feel like you were losing control?

Professor Martinis:

More precisely, I had already lost control. I wasn’t really leading the hardware group, but of course still felt responsibility, like getting the quantum supremacy experiment to work. In the end, Hartmut and I just disagreed on what to do. But he’s the boss, so it’s his decision. I think this whole year-long process made me very uncomfortable, and I just couldn’t handle it because I have been making hardware decisions for a very long time. I think I know the best way to make technology work, in a definite direction. “

72. Andrei Says:

Scott:

1) Eliezer Yudkowsky has some justified optimism here:

2) Regarding the recent retracted proof: I wonder why so many legitimate people (not crackpots) tend to get duped into thinking they solved P/NP, BQP/NP, NP/PSPACE etc. My hypothesis is that it is a field in which usually proofs aren’t written in explicit detail (unlike, say, algebra) because they would be unbearably long and almost incomprehensible, and this allows errors to easily creep in.

73. Gerard Says:

From reading the interview transcript I suspect that Martinis may not be the ideal person to be leading that particular group. However given his degree of confidence and track record I hope he has the opportunity to pursue his vision somewhere. It sounds like what he needs is more a team of engineers and technologists willing to follow his lead rather than one with a lot of PhD’s and former academics all with their own ideas to pursue.

74. Anonymous Says:

But (sigh) what was supposed to happen, what would’ve happened in a functioning country, was that the lockdown would be a way to buy time for a WWII-scale mobilization of testing, contact tracing, and away-from-home isolation of anyone infected.

This was only supposed to happen in some peoples technocratic fantasy – China with its unitary government structure, totalitarian obedience culture, mass surveillance infrastructure and giant manufacturing capacity barely managed to get a grip on one hotspot after having been aware of the problem as early as November 2019 and we only have their word for how successful that was.

The stated purpose of the lockdowns was always to flatten the curve just enough to avoid overwhelming medical services so as not to be able to provide essential response not just to covid19 cases but to – child birth, heart attacks, traumas etc.

It was also done in expectation of a much higher CFR than what was eventually seen as everyone were looking at the numbers from Lombardy.

75. fred Says:

I’ve completed my experimental 3d printed mask from https://www.makethemasks.com/ (I did a few small adjustments to make it easier to wear protective goggles with it)

It’s easy to clean.
The filter is made by cutting a square from a regular facial mask, the ones with loops for the ears (so it’s a more efficient use of them).
It also has a perfect seal along the edges (the white rubber thing is made from self-adhesive window sealers), and breathing in creates a differential of pressure between the inside and outside that seals it even more. It’s way more difficult to breathe than with a regular facial mask… it made me realized that with those you’re really breathing through the air gaps on the sides.

I wore it yesterday for 6 hours non-stop while visiting a NYC hospital. Not too bad.

76. Scott Says:

Anonymous #74:

This was only supposed to happen in some peoples technocratic fantasy…

I agree that it’s now just a technocratic fantasy. What’s so painful to many of us is that it was once the United States’ reality when great crises hit, and it’s also the current reality of other countries, like New Zealand and Taiwan.

77. Bunsen Burner Says:

Anon #74

The idea of a Lockdown was to “flatten the curve” for the reasons you gave but only until either a vaccine was developed or you reduced the number of infected individuals toa point where a test, trace and isolate regime could be effective. No idea why you think this is a technocratic fantasy given that we have empirical confirmation that several countries have managed this.

78. Ethan Says:

Anonymous #74

To me the most surprising fact, for which I haven’t seen a satisfactory answer yet, is why China didn’t share with the rest of the world that covid19 is a virus with an estimated mortality rate of at most 0.5%. Within a month of the epidemic starting to hit hard in the US we had 3 different independent seroprevalence studies (Stanford’s for Santa Clara County -the results of this study were initially questions but they were later confirmed by the other two-, USC’s for Los Angeles County and Andrew Cuomo’s for New York State) that all gave a similar result: the true mortality rate of covid19 is between 0.1% and 0.5%, of the same order of magnitude as the seasonal flu. Granted, we don’t have a vaccine or a cure, but it’s not the 5% mortality rate virus that some feared.

China didn’t have the ability to conduct seroprevalence tests or, if it did, it didn’t share the information with the rest of the world. Either way, China is a very unreliable partner that should not be trusted as the factory of the world, or at least the US. I hope that one of the consequences of the crisis is that, at least in a US context, our elites begin to understand that you can’t buy a cheap iPhone if you are dead because you lack manufacturing capabilities for things like PPE, ventilators, etc.

79. Ethan Says:

More on the Chinese government’s mismanagement of the covid19 crisis,

With the caveat that none should never, ever blindly believe any intelligence agency from any government -since these agencies, as any other bureaucracy, are first and foremost in the business of self-preservation- what is described here is consistent with the Chinese government’s traditional ways.

Why people continue to think that China can be redeemed and become a trusted international partner like Western powers are with each other is beyond me.

80. Douglas Knight Says:

Joshua,
The reason that the website uses Rt and papers use R0 is that the website is using technical language wrong. The time in the term refers to the SIR model where the population does not respond to the epidemic and the only change is immunity. Since no place on earth (probably not even Lombardy and NYC) has appreciable immunity, the correct term is R0, reflecting almost pure changes in behavior.

You could say that behavior will always change and thus Rt is a useless term for empirical measurements that might as well be picked up for a new use, but that is very different from saying that people using the consensus language are wrong!

81. Anonymous Says:

Bunsen Burner #77

No idea why you think this is a technocratic fantasy given that we have empirical confirmation that several countries have managed this.

Not really we don’t, we have a couple of relatively small, centralized and highly urbanized countries that reacted early and fast enough to treat and do contact-tracing so that they could avoid the need for war-style mobilizations. They didn’t “reduce the number of infected individuals to a point where a test, trace and isolate regime could be effective” they prevented it from reaching an ineffective stage in the first place.

Should the US have been reasonably expected to have done that given its size, federal structure and the purported role of government in society ? I can’t think of anything about its history to suggest that. WW2 was four generations ago, a long-foreseen world war in times where mandatory draft existed and it still took three years to commit and involved a Japanese sneak attack. The Manhattan project took between four to six years and during all that time the US had the prospect of becoming the world’s unprecendented superpower and recoup its sacrifice for decades to come to motivate the powers that be.

The realistic goal should have been more like Germany not Korea or Singapore, so socialized single-payer medical system anyone ?

82. Anonymous Says:

Ethan #78

What China did or did not know is impossible for anyone without access to intelligence reports (not the public statements) to ascertain – and if they did announce a 0.5% IFR and it would later turn out to be much higher because for some reason COVID19 really hates Belgians or something everybody would blame them for downplaying the severity.

The only thing that looks like China can be blamed for at this point on is not disclosing the epidemic itself earlier than they did and possibly for lying about its origin in a food market although the later would not really matter in terms of response.

Btw a 0.5% IFR is still pretty terrible as far as attaining herd-immunity is concerned, it is only the fact that it is concentrated in 70s+ age group that makes it potentially manageable without a vaccine or a cure in the near term, also we still don’t know anything about the duration of immunity conferred by an infection, this actually might be something that China could have a lead on having some of the earliest-recovering patients.

83. Ethan Says:

Anonymous #82

We are largely in agreement, particularly that all data I am aware of points to the 70+ age group being the most vulnerable. I would even add, males who are 70+. For reasons not well understood, men are being worse hit than women even in that age group.

In any case, I think that China has a lot to answer for. Whether it’s a deliberate attempt to mislead other countries or incompetence.

The lavish praise China received from so many in a US context prior to the crisis reminds me of the 1989 statement by Paul Samuelson on the Soviet Union’s economy “Contrary to what many skeptics had earlier believed, the Soviet economy is proof that … a socialist, command economy can function and even thrive.”.

The ability of Communist regimes to consistently charm a significant portion of American intellectuals amazes me.

84. Anonymous Says:

Ethan #82

For reasons not well understood, men are being worse hit than women even in that age group.

Smoking prevalence ?

The ability of Communist regimes to consistently charm a significant portion of American intellectuals amazes me.

China is hyper-capitalist economically in most aspects and socially socialist, “communism” is largely there for historical continuity of the party. I’m not aware of many American intellectuals being “charmed” with China like some were with the USSR back in its days.

85. Ethan Says:

Anonymous #84

It could be smoking is a factor but probably it’s not the only one since according to different statistics I have seen, the ratio of men smokers to women smokers is something like 5 to 1 whereas for covid19, the ratio of men dead to women dead is more like 1.5, ie, around 50% more men than women have died of covid19.

China’s influence is softer and more insidious that the Soviet Union’s but no less dangerous. I will give you two examples:

– Confucius Institutes . On paper, they are language institutes promoting the Chinese language around the world. In a US context, everybody knows that in practice they are propaganda institutes of the Chinese government. And yet, there several of America’s most prestigious universities hosting them.

– The funding of researchers such as Harvard’s Lieber. While I think that the Lieber case is definitely an overreach by the US government to make an example of him, funding prestigious academics is one way the Chinese government influences American intellectuals to look the other way when it commits abuses such as in the case of “reeducating” the Uighurs.

Some are calling the US vs China confrontation the Cold War of the 21st century. I think the cold war is a 20th century way of thinking because the threat was about he obliteration of the land where people live. Whatever is going on now between the US and China is more about which country is going to leverage the world’s high interconnectedness better. Co-inventor of public key cryptography Whitfield Diffie wrote this insightful commentary that covid19 ” is a harbinger of biological warfare” https://www.mercurynews.com/2020/04/13/diffie-covid-19-crisis-is-a-harbinger-of-biological-warfare/ .

In my mind, those who continue to see China as largely a friend of the United States are very misguided. China is not even a frenemy. Due to bad decision making in the last 40 years by government and business leaders in the US, they became America’s factory but they don’t make it a secret they have a world domination agenda. After the covid19 crisis it would be unforgivable if these leaders don’t learn the lesson.

86. Anonymous Says:

A good chunk of COVID19 deaths are said to be cardiovascular rather than pulmonary, basically the heart goes out trying to supply oxygen to the body before organs fail from lack of oxygen
And men are known to have worse cardiovascular health than women, more so as they get older.

All major countries have “language institutes” and various academic and cultural establishments whose purpose is to exercise cultural influence and so called 2nd and 3rd track diplomacy and there will always be people buying into those at face value or at least profess to.

America’s primary cultural influence establishment recommends its products to be consumed with popcorn&soda in theater settings and pizza&beer in domestic ones.

I don’t think anyone in US leadership (Harvard professors are not it, as much as they like to believe otherwise) has any misgivings about China, I’d say Russia’s influence should be much more worrying particularly because those who appear to be under it are not the Ivory Tower academics, have actual political and economic power, and are far more unwitting of promoting its agenda.

87. Ethan Says:

Anonymous #86

What you say makes a lot of sense. I agree that my own contacts who work or have worked in American national security are clear about the threat that China poses. At the same time, the intellectuals, who also have a role shaping American public opinion, are more of a mixed bag. I have heard very influential people making the argument that it is better to engage China -despite their horrendous human rights record with the Uighurs but also in other areas- to try to change them than to boycott them. The argument went something along the lines that given the trade exchanges between our two countries, China wouldn’t do anything to damage the US because that would go against their own interests. That argument clearly failed in the case of the covid19 crisis. Everything points to China engaging in a cover up that damaged them and damaged us to save face rather than admitting that they did something wrong.

With respect to the Lieber case, from the facts known in the public domain, it seems to me a classic “perjury trap” to make an example of him and to scare other academics who might be considering taking money from the Chinese government.

88. Anonymous Says:

Ethan 87

Treating COVID19 as a deliberate attempt by China to damage the US is just silly and if only because the US is far from being the worst-hit country, saving face on the world stage and more importantly domestically for the CCP is more than good enough a reason for a coverup on its own.

I don’t know anything about the Lieber case so won’t opine.

89. Ethan Says:

Anonymous #88

This is the latest. It seems plausible,

https://www.msn.com/en-us/news/world/report-china-hid-coronavirus-severity-to-hoard-supplies/ar-BB13yha5?li=BBnb7Kz

Still, the covid19 crisis underscores bad decision making by US leaders. The people who sold the notion of using China as the factory of the world for this kind of stuff don’t understand human nature. In a situation such as this, any country -including our own- would have probably done the same thing: hoard supplies for your people before you tell others about the event that will push them to ask you for supplies. The dream of a single, global supply chain with China as its bottleneck became a nightmare. We have the large technology companies, particularly Apple, to thank for this line of thinking.

Had the US been the global epicenter of the covid19 epidemic, the big difference, of course, would have been that if the US government had tried to do anything of this sort, there would have been reports from the free press leaking information. China doesn’t have a free press. Having a robust free press is one way we are different from them.

90. Anonymous Says:

Ethan 89

There was no dream or plan – just obeying the simple principle of sourcing your stuff at the cheapest cost available and selling it for what the market will bear without too much foresight because both Poltiicans and CEOs have reward horizons measured at best in few years.

And the supply chains still hold remarkably well given the situation, no efficient supply chain can be planned to accommodate panic hoarding – not maintaining emergency PPE and equipment stockpiles was an explicit decision made by the Obama Administration somewhere in 2013 iirc.

Yet two months into the crisis there are no acute shortages of essential supplies anywhere in the developed world, some people briefly had to wait for eggs, chicken etc. due to local supply chains disruptions while having many staple alternatives available. We are still playing this on easy mode.

Had the US been the global epicenter of the COVID19 epidemic, the big difference, of course, would have been that if the US government had tried to do anything of this sort, there would have been reports from the free press leaking information. China doesn’t have a free press. Having a robust free press is one way we are different from them.

The notion of the US being exceptionally free for speech is largely outdated at this point and relies on an increasingly nonsensical freedom vs. freedom-from-consequences distinction.

While nominally free from state action, media outlets and professinals are are controlled almost entirely by private interests that can and do have state and party interests willingly (and since 2001 not so willingly) practiced through them. Robust ? Not particularly.

And as we speak, a foundation for overt national censorship and communications surveillance mechanism is being legislated in the Senate under a tissue-thin guise of fighting child abuse, with bipartisan support.

91. Ethan Says:

Anonymous #90

I do think that changes in the supply chain are still probably in the horizon for things such as PPE. The US established its Strategic Petroleum Reserve as a result of the 1970s oil crisis and I think we are headed to a similar situation with respect to items deemed of high national security value in the context of healthcare, irrespective of who wins in November.

We will have to agree to disagree with respect to your diagnosis about the status of free media in the US. By “free” I don’t mean “perfectly free” . We don’t have anything of this sort here https://www.theguardian.com/news/2018/jun/29/the-great-firewall-of-china-xi-jinpings-internet-shutdown . Or consider the Edward Snowden revelations. In my opinion he is one of the heroes of our time. China doesn’t have the mechanisms at large to expose these things to their own population as we do. The free media is more than the official big three (ABC, NBC, CBS), the NY Times, the WaPo or CNN/Fox News. It’s a very diverse set of outlets that can go directly to the people with their message largely unfiltered.

I am very happy that the free media is still one of our country’s most distinctive features not only with respect to totalitarian countries like China but also European ones. In the UK, for example, you can be charged for dishonoring the Royal Family. British courts are also known for being one of the best venues in the free world to sue people for libel.

92. E. Harding Says:

Had the US been the global epicenter of the covid19 epidemic, the big difference, of course, would have been that if the US government had tried to do anything of this sort, there would have been reports from the free press leaking information. China doesn’t have a free press. Having a robust free press is one way we are different from them.

We have a press that is “free” to publish articles like this:

The notion of the US being exceptionally free for speech is largely outdated at this point and relies on an increasingly nonsensical freedom vs. freedom-from-consequences distinction.

Half-true. Legal harassment in the U.S. is far less prevalent than in Europe/Asia. However, there is no question VKontakte is freer than Facebook.

The people who sold the notion of using China as the factory of the world for this kind of stuff don’t understand human nature.

Much better have China as the factory of the world than a country that can’t even control a simple outbreak because “lots of states rewrote their quarantine laws in the wake of the AIDS epidemic providing for mandatory judicial review of isolation orders”.

93. Ethan Says:

E. Harding #92

We will have to do something as American as agreeing to disagreeing. When it comes to free speech, I find the protections afforded under the 1st amendment -and the American tradition of free exchange of ideas behind it that goes beyond the technicalities of the of the US Supreme Court rulings based on the 1st amendment- a national treasure. There is no other country in the world I am aware of that has stronger free speech protections. Thus, I am of the opinion that the answer to speech someone doesn’t like is more speech, not censorship, prior restrain or government bodies that determine “official truth”.

When it comes to the covid19 epidemic, I am am actually quite happy with the way the US has confronted it. Curtailing civil liberties is always a balancing act. The lockdown measures have also caused stress in people unrelated to covid19. It’s not lives vs the economy but some lives vs others lives. The lives of those whose medical procedures have been delayed -potentially resulting in death- in those places with the strictest measures have also been affected by the lockdown. I am of the opinion that restricting civil liberties should be done only in the most limited way, for the shortest period of time possible and lifted immediately after the compelling government interest -such as minimizing the number of deaths due to covid19- is no longer served by them. And the final decision should always go to go government officials accountable to voters, never to self appointed guardians of society regardless of their expertise or credentials. This notion that scientists are immune to human fallibility and are not prone to the so called “law of the instrument” has no empirical basis. There is an entire -very rigorous- website dedicated to exposing scientific malpractice and fraud, this one https://retractionwatch.com/ . You should check it out and see how a lot of the retracted work comes from the life sciences.

Ultimately, what matters is saving lives. In a US context, you have all sorts of states adopting all sorts of different lockdown measures giving all sorts of different results in terms of rate of death per 100000. The voters of each state will have an opportunity to hold their governors accountable. And when it comes to the federal response, voters will have an opportunity to give their opinion this coming November. I am quite sure that Trump’s management of the crisis will be if not the top concern of Americans when they go to the ballot box.

94. Anonymous Says:

Ethan, E. Harding

Yes the US is still on an entirely different level of censorship and intimidation than China – comparison to Europe is complicated and the UK is a yet different kind of mess. And yet ‘one of the heroes of our times’ had to flee to Hong Kong, meet with British journalists and is now taking refuge in Russia of all places although not by choice.

The reason VK is freer than Facebook btw is not very indicative of freedom or speech being valued for its sake, rather it is because the Russian government is using it as a social and political monitoring honeypot in a far more direct way than the US does with Facebook.

So much like with COVID19 it is at least about the trends as it is about the status quo and the trends are very disturbing.

95. Ethan Says:

Anonymous #94

“And yet ‘one of the heroes of our times’ had to flee to Hong Kong, meet with British journalists and is now taking refuge in Russia of all places although not by choice.”

Touché ! America is also a land of contradictions :-). One thing I have to say about this is that Edward Snowden is a product of the American system too. America is bigger than the federal government and its many unjust laws ( https://www.amazon.com/Three-Felonies-Day-Target-Innocent/dp/1594035229 ). It’s a mindset and a state of mind as well. We haven’t seen -or at least I don’t know of anyone- similar whistleblowers produced by other European/Western countries.

In the context of covid19, I can’t see a situation like the intimidation Li Wenliang endured happening going largely unreported or underreported in the United States.

The United States has a long tradition of elevating whistleblowers to hero status, particularly those who break unjust laws to bring truth and justice to the American public.

96. Ethan Says:

What did I say about scientists being flawed human beings too…

“Prof Neil Ferguson, the epidemiologist whose modelling helped shape Britain’s coronavirus lockdown strategy, has quit as a government adviser after flouting the rules by receiving visits from his lover at his home.

Ferguson runs the group of scientists at Imperial College London whose projections helped persuade ministers of the need to impose stringent physical distancing rules, or risk the NHS being overwhelmed.”

97. Scott Says:

Ethan #96: What a travesty. Not only was Ferguson immune (having already recovered from covid), but if you just defined him and his lover to be a “family,” inhabiting two houses, then there’d be no problem. Meanwhile Trump, Pence, Boris Johnson… openly flouted the rules in more significant ways with no negative political consequences (in Boris’s case, a negative health consequence).

98. Ethan Says:

Scott #97

For the record, I am personally against strict lock down measures to combat covid19. Sweden and Japan are showing empirically that when you trust people to behave like adults, they generally behave like adults. Those two countries have had reasonable fatality rates without wrecking their respective economies. As I mention above, restricting civil liberties to accomplish one goal -such as minimizing the number of covid19 fatalities- always has other unintended consequences, so restricting basic liberties is always a balancing act. We know enough about covid19 to conclude that most strict lockdown measures are irrational. The raw mortality rate, 3 independent studies in the US show, is at most 0.5% and men older than 70 are being disproportionately affected. Keeping everybody lockdown, doesn’t make any sense. People should be given the choice of staying home or getting out with the commitment of social distancing themselves from others. And if government wants to do anything, it should focus on assisting -not restricting the freedoms of- vulnerable populations -such as men older than 70- who want the extra help.

This said, with Neil Ferguson we have a case of Enlightened absolutism. That kind of thinking goes all the way back to the Enlightenment era. It assumes that the rulers are enlightened but the ruled are stupid and thus such a situation calls for rules that apply to the ruled and not to the rulers. Such a proposition is antithetical to the values on which the US was founded -government of the people, by the people and for the people. I have met my share of modern day “enlightened despots” and in my experience they rarely see any kind of hypocrisy in their demanding a set of rules for the common people and another set of rules for them. Those in government typically forget that they are two centuries late for that kind of thinking to be tolerated by voters in Western democratic countries.

Your plea for Neil Ferguson would make sense if that kind of reasoned approach is what Neil had asked from everyone (such as the governments of Sweden and Japan did of their citizens). Alas, that’s not what happened, so calling him a hypocrite -whether he feels any remorse- is apt. So is asking him to resign.

99. Ethan Says:

Scott #97

One more thing. The other people you mention are accountable to voters. Neil Ferguson is not. That’s a huge difference! If voters (American or British) don’t like the way those leaders have done things, they will have a chance to kick them out -in a US context pretty soon, actually.

PS: please remove my other comment with content equal to this one in which I erroneously put “Scott” in the Name field.

act that will fill Georgia’s ICUs and morgues as surely as night follows day

This is a testable prediction! Care to put some numbers on it so we can check back in a month or two and see how right you were?

We’re two weeks in now, which is about the earliest we could expect to see effects.

“Florida’s new cases have actually declined by 14% compared to the previous week, and Georgia’s fell by 12%.”

https://www.axios.com/coronavirus-cases-map-high-risk-states-8ceeaa05-cc07-4e8b-b9f4-df3a3315f143.html

101. Scott Says:

We’re in the Calm before a New Storm of COVID-19 Infections and Deaths

If this is not what happens next—i.e., the same thing that’s happening to numerous other nations, a few weeks or a month after they eased restrictions far more competently than we are—then the universe is magic. It seems obvious enough that, rather than argue it, I’d rather just post the link and then let events do the arguing.

The peak demand for hospital beds could range from 11,000 to 25,000; for ICU beds from 1,800 to 4,000; and for ventilators from 900 to 1,900. These demands are likely to far exceed the capacity that may be available for COVID-19 patients in the state.

Our research suggests shortages across 14 coordinating hospital regions in Georgia

Is it fair for me to pull this out as the metric, then? If, by August, 10 hospital regions run out out of space for COVID-19 patients for several weeks, this model was accurate and Georgia was wrong, but if it’s less than that, Georgia was right, and Scientific America was wrong.

Is there a state you think did it right that we should keep handy as a control?

103. Thapsus Says:

“Brian Kemp, the doofus governor of Georgia, who’s infamously announced plans to reopen the state right away, ignoring the pleading of public health experts—a act that will fill Georgia’s ICUs and morgues”

I realize that this was a throw-away slur at a conservative and not your main point, nevertheless you made it and there is now data to evaluate it. So let’s do that.

Go here: http://91-divoc.com/pages/covid-visualization/ then go to the last graph in the page. Now select from the options at the bottom: None, Total Deaths, Linear, Days Since Threshold and Fixed. We now have a graph that reflects population size adjusted rankings of the various states’ fatality rate. Can you find Georgia?

Now you can see that the worst states for Covid-19 results are New York, New Jersey, Massachusetts, Michigan, Illinois, Connecticut, California, Pennsylvania, Louisiana and then things get crowded. Only one of those states have Republican governors; can you guess which one? Georgia is near the top of the crowd but below some more Democrat governor states.

104. Mark Says:

Scott #101 – it’s now closing in on 7 weeks since Georgia’s reopening, and there’s still no sign of an increase in the rate of new cases or deaths there.

I’m not bringing this up to try to dunk on you. Like you, I thought for sure that new cases and deaths would go up. Why wouldn’t they?

Well, they haven’t, not in Georgia and not in several other places, and this has led me to the conclusion that _we do not understand what is happening_. The experts and the models _do not have predictive power_.

I believe this ignorance is far more dangerous than any individual risky policy decision. If we can’t successfully make predictions as basic as “if shelter-in-place restrictions are lifted, disease transmission will increase”, we must be completely in the dark. What are your thoughts?

105. Scott Says:

Mark #104: I agree,
(1) Georgia has not yet had a second wave, and
(2) neither you nor I know anything that could possibly have led a reasonable person to predict that back in April.
Regarding the possible explanations on the table—e.g., heat, informal social distancing, luck, data manipulation—I’ll just refer you to this Vox article, which (despite Vox’s earlier missteps) I found to be pretty detailed and good.

106. Filip Says:

0x4 (c) ™ ‘. ‘ ‘3 3’ ‘ 3’ — —