Over at National Review, Jim Geraghty has a series of articles suggesting that the Covid virus escaped from a research lab in Wuhan, China. Today, he has a story with the following headline:
Guess Where the Possibly Nuclear-Fuel-Leaking Sunken Chinese Submarine Is?
I didn’t have much trouble guessing—it was Wuhan. What did surprise me is the way he spun the story:
You probably remember that one, on account of the fact that it completely disrupted your life for a year or two and caused 27 million or so “excess deaths” around the world. But I’ll bet you don’t remember the Wuhan University researchers who allowed artificial intelligence to control an Earth-observation satellite, which led the satellite to start looking at Indian military bases and a Japanese port used by the U.S. Navy. Lead researcher Wang Mi boasted, “This approach breaks the existing rules in mission planning.” Yes, and we all know all the great things that happen when scientific researchers in Wuhan break the existing rules. First the Andromeda Strain, then SkyNet.
What other kinds of experiments are they doing over there in Wuhan these days? Summoning demons? Reaching out to say “hi” to some hostile alien empire in outer space? Are they just flipping through old Marvel comics, reading about the villains’ plots, and thinking, “Hey, that would make a cool experiment”? All the troubles in the world apparently lead back to Wuhan.
That final paragraph—especially the final sentence—is the sort of thing I’d expect from a conspiracy theory skeptic, someone who wished to make fun of the idea that certain coincidences are suspicious. I could imagine someone mocking the claim that, “Wuhan has only about 1% of China’s population, so how likely is it that the submarine would happen to sink in the same city where Covid started?” In other words, making fun of someone for not understanding Bayesian reasoning.
To see the problem consider how the final sentence of the first quoted paragraph could be re-written:
Yes, and we all know all the great things that happen when wild animal wholesalers in Wuhan break the existing rules. First a repeat of what happened with SARS-1, then SkyNet.
Yes, I understand that Geraghty is mostly just being humorous here. But if you treat the column as humor, then he’s poking fun at his own views on Covid. Thus I wonder if he’s being at least slightly serious. At some level he seems to be assuming that digging up more dirt about Wuhan makes it somehow more likely that readers will believe (if only subconsciously) that something bad happened there back in late 2019. But we already know that something bad happened in Wuhan—a man was selling raccoon dogs in the food market.
What this example actually shows is that weird coincidences happen all the time, and it would be foolish to make any causal claims based on their existence.
Here’s another coincidence. For the first time in 36 meetings, the Fed cut its fed funds rate target. What are the chances that politics had nothing to do with a rate cut occurring at the final meeting before the November election?
I’d say the chances are pretty good. (BTW, the previous rate cuts were also in an election year.)
Here’s another interesting pattern: There has never been a time when the 3-month average of the unemployment rate rose by more than 0.5% without a recession. What are the chances that the recent increase in the unemployment rate over that threshold will not lead to a recession?
I’d say the chances are pretty good.
If you seek out patterns, you will find them. Lots of them. But the world is full of unusual events.
READER COMMENTS
Scott H.
Sep 28 2024 at 8:22am
If you give someone a six-sided die and tell them to roll it 10 times, the exact chance of getting whatever they rolled is 1 in 60,466,176. That’s incredibly unlikely. So, to the mathematically challenged, the question arises… what are the odds the person really did roll those numbers?
Robert EV
Oct 1 2024 at 2:49pm
The odds of them rolling those numbers are the odds of them completing the task.
I think you’re conflating combinations with permutations, though.
Trump tried to coo
Sep 28 2024 at 9:56am
This is why I don’t believe the Great Stagnation happened. The GDP data are noisey so it always looks like averages are calculated with too little data, measurement methods evolve, and a person’s intuition that there will be flying cars doesn’t give insight into when. Even if GDP stagnation is believable, and it isn’t for me, we should care more about stagnation of what people try to maximize and we know it’s not income because people have some other priorities.
Scott Sumner
Sep 28 2024 at 3:14pm
To me, there is no definitive answer to this question. How are we measuring stagnation? Quantity of stuff? Quality? Happiness? How much weight do we put on stuff like better restaurants, better TVs, better cell phones, etc. It all seem very subjective.
But I do think 1900-1973 was especially transformative. I suppose that’s because I find things like cars, indoor plumbing, and home appliances to be more impressive than iPhones and the internet. It’s very subjective.
Jim Glass
Sep 28 2024 at 2:21pm
Great correlations.
Scott Sumner
Sep 28 2024 at 3:14pm
Funny.
Stephen
Sep 28 2024 at 11:22pm
“But I do think 1900-1973 was especially transformative. I suppose that’s because I find things like cars, indoor plumbing, and home appliances to be more impressive than iPhones and the internet. It’s very subjective.”
Because you list five things, it’s not too hard to imagine how much worse life would be without any one of them. I did without smartphones and the internet for more than half my life, and I could do so again if I had to. But no hot water, washers and dryers, toilets, or horseless carriages? “Let’s always take a moment to also see what we have achieved thus far, while we clearly see the moment that we are presently in.”
Market Fiscalist
Sep 29 2024 at 12:41am
Coincidences do happen but in most cases investigation shows that things thought to be coincidences are in fact not coincidences at all.
The fact that a new virus emerges in a city which is a major global center of research into new viruses may be a coincidence – and claims about people selling raccoon dogs in the food market may be evidence in favor of the coincidence theory – but I don’t quite get why Scott is so opposed to the non-coincidence explanation when the evidence for the coincidence view is circumstantial at best.
Jon Murphy
Sep 29 2024 at 9:56am
You have the burden of proof backward on this. At least from a scientific perspective, one needs to show that something is not a coincidence, not that something is a coincidence.
Scott H.
Sep 29 2024 at 10:54am
Actually, Prof Sumner is taking into account the most important coincidence — the fact that the outbreak happened right next to a wet market atmosphere famous for causing such epidemics. It’s THAT coincidence that will turn out to be no coincidence at all.
So, the COVID outbreak is not a coincidence versus non-coincidence dichotomy. It’s which coincidence is more likely to produce this event?
Market Fiscalist
Sep 29 2024 at 12:18pm
I think the coincidence/non-coincidence thing is largely irrelevant if one is interested in establishing the truth. The fact that a virus lab is in the same city as an outbreak is certainly a fact that would lead one to examine the evidence supporting the lab leak theory very carefully as well as looking at evidence supporting other theories.
BTW: I see that I probably mis-stated Scott’s views in claiming he is opposed to the lab leak theory – I see (thanks to ChatTMI) that his view is actually more nuanced on this issue.
Jon Murphy
Sep 29 2024 at 12:29pm
Quite the contrary, mon ami. It is of utmost primacy. Determining what is coincidence and what is causal is the primary concern in establishing truth.
Market Fiscalist
Sep 29 2024 at 1:29pm
OK, I will concede that point.
However I still object to your original comment about backward causality as nowhere do I suggest that anyone has a requirement to show “that something is a coincidence?”
I said ‘evidence for the coincidence view is circumstantial at best’. In context ‘the coincidence view’ is the wet market origin theory (a theory that if proven would lead to the conclusion that the presence of the lab was indeed a coincidence)
So I was claiming that ‘evidence for the wet market origin theory is circumstantial at best’ and your statement that ‘one needs to show that something is not a coincidence, not that something is a coincidence’ is irrelevant.
Jon Murphy
Sep 29 2024 at 1:51pm
Indeed, you do. You highlight so in the very next sentence. You write:
The full sentence you wrote was:
Market Fiscalist
Sep 29 2024 at 2:02pm
So with relevant substitution I am saying “I don’t quite get why Scott is so opposed to the lab-leak theory when the evidence for the wet market theory is circumstantial at best.”. How is your comment relevant to that statement?
Warren Platts
Sep 29 2024 at 12:21pm
That is backwards. The burden of proof lies on the scientists conducting risky research to show that (a) they know what they’re doing; (b) there’s a good reason for doing the research; (c) they can conduct the research safely; and (d) when something terrible happens, they are presumed responsible; it’s their job to prove their innocence. At Wuhan, none of these conditions were met.
Warren Platts
Sep 29 2024 at 12:27pm
Regarding proof that SARS-CoV-2 is a synthetic virus, that’s easy: (1) SARS-CoV-2 contains a furin cleavage site than none of the other viruses in its family have that radically increases the virus’ virulence; (2) the SARS-CoV-2 genome is separated into six equally spaced pieces by restriction enzyme sites.
Re: (2), the SARS-CoV-2 genome is like a pearl necklace with five evenly spaced clasps. Commercially available endonucleases (enzymes that bind to restriction sites and split the genome there) viz. Bsal and BsmBI, can split the SARS2 virus into six pieces. The reason virologists like to do this is that it is easier to manipulate smaller strands of DNA rather than the entire genome at once. The restriction site consists of a sequence of several base pairs, and these can occur naturally; but if the sites are evenly spaced, that’s a typical fingerprint of a synthetic virus.
But I suppose one could still claim all that’s a million to one coincidence. However, the smoking gun is that DEFUSE grant proposal that was submitted by American virologists to DARPA before the pandemic. In it, they proposed to take a coronavirus, re-engineer the virus so that it could be split into six manageable chunks in order to facilitate the insertion of a furin cleavage site. Guess what commercially available endonucleases they wanted to use? Yep: Bsal & BsmBI. Indeed, they had drawn up purchase orders for these endonucleases! And since DARPA refused to sanction the research, they moved the project to the scientific Wild West of Wuhan.
It’s like a murder mystery where a person dies apparently by a bizarre accident but that required a suspicious number of coincidences to happen. The person on trial says million-to-one coincidences happen all the time! Reasonable doubt right there! But then the prosecution produces a document written by the defendant prior to the death that contains a detailed, step-by-step plan on how to carry out the murder such that it looked like a bizarre accident. In that case, if you’re a juror, you have to vote guilty as charged.
That’s what’s happened with covid, except that 26 million people died. There is no question that SARS-CoV-2 is a synthetic virus manufactured in Wuhan. It’s highly irresponsible to suggest otherwise imho. The sick part is that SARS-CoV-2 is one of those “Designed in America — Made in China” products. It was a joint effort…
Scott Sumner
Sep 29 2024 at 9:04pm
“Regarding proof that SARS-CoV-2 is a synthetic virus, that’s easy”
This claim has been thoroughly discredited by scientists that understand the subject.
Warren Platts
Sep 30 2024 at 6:23am
And at least as many knowledgeable scientists who have examined the evidence say SARS-CoV-2 is a synthetic virus created in a Wuhan lab…
Cf. There is now very little doubt that Covid leaked from a lab. Opponents of the lab-leak theory have comprehensively lost the argument.
Scott Sumner
Oct 1 2024 at 10:57am
Would it surprise you that I’m not interested in conspiracy theories from Spiked magazine? Consider your sources.
Warren Platts
Oct 1 2024 at 2:17pm
Scott, the author of that article was Matt Ridley, noted science writer, libertarian, Hayek Lecturer, former PM in the House of Lords, and author of Viral: The Search for the Origin of COVID-19! I’m honestly not trying to be a troll here — and I admit there is uncertainty in any empirical claim — but here is a formal debate between Ridley and Stephen Goldstein, University of Utah virologist, a bona fide member of zoonotic origin theorists you cite. This is the closest thing you’ll find to a civil trial of the Wuhan Institute of Virology where anyone can judge where the preponderance of the evidence lies. Worth anyone’s time imho:
Did COVID come from a lab? Matt Ridley vs. Stephen Goldstein (youtube.com)
Robert EV
Oct 1 2024 at 3:51pm
1) The furin cleavage site is non-canonical.
2) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7836551/
Robert EV
Oct 1 2024 at 3:58pm
Also, as a person who has literally published on synthetic biology, BsaI and BsmBI are the two most common enzymes for Golden Gate cloning. Golden Gate cloning is one of the most used methods for assembling smaller pieces of DNA (synthetic, or otherwise). Typically, when using Golden Gate cloning, you fully cure the DNA parts of interest of all internal BsaI or BsmBI (or SapI, AarI, what have you) restriction enzyme recognition sites. You don’t have to cure all of these sites, but to make assembly easier you typically do. Which means that had SARS-Cov-2 been manufactured in a lab using BsaI or BsmBI based Golden Gate cloning, it is more likely that the BsaI or BsmBI restriction recognition sites would be fully absent from the SARS-Cov-2 genome, not present at 6 places.
If you don’t understand why, I suggest looking up a description of Golden Gate cloning, maybe include the search term “domesticate” in your query.
Robert EV
Oct 1 2024 at 4:00pm
Note that I’m neutral on the lab leak theories (engineered or natural) versus the wet market theory.
Robert EV
Oct 6 2024 at 12:35pm
You aren’t going to read this, but I downloaded the Wuhan SARS-CoV-2 genome and analyzed it in ApE (A Plasmid Editor). Those 3 BsmBI sites and 2 BsaI sites are not evenly spaced. They would digest a DNA version of the SARS-CoV-2 genome up into the following sizes: 7578 bp, 7558 bp, 6130 bp, 5802 bp, 2192 bp, and 643 bp.
Furthermore they yield the following 4-nucleotide overhangs for Golden Gate ligation: TCTT, CTAA, TGCC, TGAT, and TGCT. The last overhang differs by only a single nucleotide with the third and fourth. Prior to NEB’s profiling of overhangs it was rule-of-thumb to have at least two nucleotide differences between each overhang in order to minimize mis-ligation events between similar, but not identical, overhangs. NEB’s ligase fidelity calculator shows that misligation events will occur an estimated 5% – 24% of the time with these overhangs, depending on the assembly conditions.
To sum up, there are a number of reasons why the restriction site patterns of SARS-CoV-2 have not been broadly cited as evidence of a lab-made virus. What I’ve told you hear are not all of the reasons, but I hope you can see that they are sufficient to fully contradict the reasoning you’re making here.
Robert EV
Oct 6 2024 at 12:35pm
here! here! here! Not hear.
JoeF
Oct 7 2024 at 7:25am
That’s interesting, Robert. If you have downloaded the genome and sincerely believe you are an expert on restriction sites, you should engage with Alex Washburne on X. He is very open to honest discussion.
Warren Platts
Oct 1 2024 at 2:18pm
Link here.