On the question of lineage B preceding A, it's not clear who believes that any more. Crits-Christoph 2024 phylogenies came out 90% in favor of A as ancestral to B. Figure 1B is much clearer than the text on this point. (The text is also a bit odd around the relative frequencies of lineage A and B in market samples, which they describe as "the presence of both lineages A and B at the market".)
Maybe it's a cheap point to make, but the CIA, the FBI, and intelligence agencies writ large aren't largely comprised of rebels, self-styled or otherwise.
There are intel lab leakers, data-oriented lab leakers, academic ones, activist ones, crazy ones, bat-shit crazy ones, MAGA ones, and Jamie Metzl. Conversely, the other side is a small academic niche and the rationalist autists they've persuaded.
Xinfadi is a particularly odd choice as an example for a reason to be skeptical of a market as the origin of an outbreak. Published evidence and analysis that I'm aware concludes it's an animal industry source at the market -- albeit a frozen one that wasn't infected itself.
I'm a little skeptical of that conclusion but I haven't looked closely enough at the data to say more than that.
In general, these examples are problematic because they almost all (maybe all?) take place during the first wave of the pandemic and its response, in countries were much of the economy and other parts of society had a dramatically reduced density and when infection control was dramatically increased in healthcare. The food industry isn't capable of reducing density to the same degree as other industries; people need to eat and food is cheap. Lots of food market anecdotes after early 2020; few market anecdotes in the beginning of 2020 when SARS-CoV-2 was first introduced around the world.
It's not a status quo situation like Wuhan was at the end of 2019. For that, you need to look at large SARS and MERS outbreaks for a comparison of where somewhat similar viruses spread disproportionately. The answer is that beyond the earliest days of the outbreak, spread is most rapid in the healthcare system.
I particularly like how you began this post by saying, "there’s no reliable substitute for really following the arguments."
And then you later went on to say, "the participants in the debate you watched seem to have just brushed past this required step, at least in the few hours that I could put up with watching."
Are you arguing in favor of thoroughness, or not?
You can call me crazy, too, but I think you should probably actually watch a debate and understand it before critiquing it.
It's easier to follow and check on written versions. I read the write-ups by both judges and by Scott carefully, in detail. At certain points in the video, I lost patience. E.g. where you faked the t-axis on a plot and where you made a big point about N501Y incorrectly attributing its rapid evolution to culture on humanized mice rather than regular mice. There's a reason that committing to written work is preferred. In this case, it saved me from a tedious critique of your dishonest tangents and allowed a focus on the more solid parts of the argument as written up by multiple viewers strongly favorable to your conclusions.
if you were to write up your Bayesian analysis it would be harder to brush past the enormous problems with the non-hierarchical extreme factors used. If you think the analysis here is wrong, it's written down for everyone to share and for you to respond to.
It's okay for a person to not have the patience to review an argument. If you make that choice, however, you should not profess to have reviewed or understood the content, nor should you criticize others for not reviewing arguments in sufficient detail.
I have likewise not read every version of your bayesian analysis, after sensing from a subset of text as well as from the summaries that you have no valuable insight into the subject of Covid origins, nor any arguments better than those put forth by Rootclaim. Scott likewise claims to have read none of your substack posts.
I think it's fine that I have not reviewed your work and think it lacks value. That said, it would be dishonest for me to write a detailed critique of your longest substack posts without reading them. I would also have to admit that there exists a possibility that there is some valuable argument put forth in one of your posts that would lead Scott or I to conclude that Covid has a lab origin. My bayesian assessment, however, holds that the odds that such an argument is present are very low.
I watched the key part from which the huge Bayes factor was derived, the part about how unlikely it would be to get the initial ascertained outbreak at HSM if the spill was at WIV. I watched parts of that more than once. The whole bit about inferring the reliability of an exponential extrapolation using the high value of R^2, etc, stuff to give stats types giggle fits. I wrote down in detail above what was wrong with that type of argument.
Scott needn't read my whole long-winded analysis, just the very brief response to his own analysis.
Rather than rehash old process stuff, why not specifically reply to the actual technical points?
The specific reason for not giving an itemized response, every time someone on the internet makes a new list of reasons why they think Covid is a lab leak, is that such lists have been created thousands of times, and will continue to be created.
The Rootclaim debate was very long, not because the case for zoonosis is long and complicated, but because the number of lab leak theories and objections to that case is nearly limitless. The fact that DRASTIC cofounder Yuri Deigin, multimillionaire Saar Wilf, and Saar's research team were unable to find any convincing argument against that case, after 18 hours of trying, is itself strong evidence that there is no good argument for the lab leak theory.
The "key part from which the huge Bayes factor was derived" is the fact that all epidemiological evidence in Wuhan clearly points to the pandemic starting at Huanan market, that starting point is very unlikely if SARS-CoV-2 was created at WIV, and 18 hours of objections to that theory can not produce a single valid reason why it's not true. There is, by contrast, no convincing genetic evidence that SARS-CoV-2 was engineered, and the odds of an epidemic starting in the same city with a lab are quite high. Bayesian analyses are toy models at best. Scott and both Rootclaim judges expressed great skepticism as to how well such models could solve a complicated question like this. That said, any unbiased analysis of that type would side towards zoonosis, just as the scientific consensus favors a zoonotic origin.
I've reviewed many lab leak arguments raised since the debate ended. I have found none convincing. I've debunked many of them on Twitter, including several claims on your new list.
I could do yet another debate rebutting the new, longer list. Perhaps that would stretch to 24 hours, instead of 18. After I won, the next round of conspiracy theorists could argue that debate is too long to watch, and then make new lists of itemized objections, which will also be wrong.
This phenomenon is the same with many other conspiracies and forms of pseudoscience. Anti-vaxxers have spent years writing different reasons why Covid vaccines are extremely dangerous or deadly. One could easily spend over 18 hours debating every version of these theories. And, in fact, Rootclaim has been doing just that, by taking the pro-vaccine side of a debate against Steve Kirsch. This time they're holding most of the discussion in writing, and the back and forth has been going on for over 6 months now. Steve Kirsch does not tire easily, and I'd imagine he could spend another 6 months making false claims about the danger of Covid vaccines, based on obscure data sources. I've long since stopped reading Steve's claims, and don't need to know what he's discovered in his latest secret data from New Zealand, or whatever, but I can see why Rootclaim would take the time to wade through all that nonsense for a million dollars.
If you'd like to put up 100k as an incentive for me to have another Covid origins debate, I'll gladly show why your "new and improved" list of lab leak claims is just as wrong as the previous one. You're also welcome to publish your claims and hope that scientists respond. Otherwise, I don't feel much need to reply to every new list of recycled lab leak claims some conspiracy theorist puts up on Substack or Twitter.
I have only a few points that seem reliable enough to give significant Bayes factors. They're quite similar to the first version, although the CGGCGG factor is substantially reduced and new data has led to me including the RE site factor, which I had initially rejected.
Ah, yes, the restriction sites paper. I also proposed that Valentin Bruttel try to defend his idea in a judged debate, but he refused, saying that any judges would not be able to understand its brilliance because he is Galileo, Einstein, and Semmelweis combined:
Virologists aren't convinced by the restriction sites argument. Saar and Yuri thought it wasn't reliable enough to even try presenting it at the Rootclaim debate. Even Guy Gadboit has done an analysis arguing against it, giving a bayes factor that leans towards a natural origin.
If this is the kind of stuff that underpins your Bayesian analysis, then it's just garbage in, garbage out.
On the question of lineage B preceding A, it's not clear who believes that any more. Crits-Christoph 2024 phylogenies came out 90% in favor of A as ancestral to B. Figure 1B is much clearer than the text on this point. (The text is also a bit odd around the relative frequencies of lineage A and B in market samples, which they describe as "the presence of both lineages A and B at the market".)
Maybe it's a cheap point to make, but the CIA, the FBI, and intelligence agencies writ large aren't largely comprised of rebels, self-styled or otherwise.
There are intel lab leakers, data-oriented lab leakers, academic ones, activist ones, crazy ones, bat-shit crazy ones, MAGA ones, and Jamie Metzl. Conversely, the other side is a small academic niche and the rationalist autists they've persuaded.
The other side also comprises the stenographer scientific press and some big scientific institutions and journals.
I'm old-fashioned. I think the proper word would be "composed", not "comprised" in your Comment.
Xinfadi is a particularly odd choice as an example for a reason to be skeptical of a market as the origin of an outbreak. Published evidence and analysis that I'm aware concludes it's an animal industry source at the market -- albeit a frozen one that wasn't infected itself.
I'm a little skeptical of that conclusion but I haven't looked closely enough at the data to say more than that.
In general, these examples are problematic because they almost all (maybe all?) take place during the first wave of the pandemic and its response, in countries were much of the economy and other parts of society had a dramatically reduced density and when infection control was dramatically increased in healthcare. The food industry isn't capable of reducing density to the same degree as other industries; people need to eat and food is cheap. Lots of food market anecdotes after early 2020; few market anecdotes in the beginning of 2020 when SARS-CoV-2 was first introduced around the world.
It's not a status quo situation like Wuhan was at the end of 2019. For that, you need to look at large SARS and MERS outbreaks for a comparison of where somewhat similar viruses spread disproportionately. The answer is that beyond the earliest days of the outbreak, spread is most rapid in the healthcare system.
I particularly like how you began this post by saying, "there’s no reliable substitute for really following the arguments."
And then you later went on to say, "the participants in the debate you watched seem to have just brushed past this required step, at least in the few hours that I could put up with watching."
Are you arguing in favor of thoroughness, or not?
You can call me crazy, too, but I think you should probably actually watch a debate and understand it before critiquing it.
It's easier to follow and check on written versions. I read the write-ups by both judges and by Scott carefully, in detail. At certain points in the video, I lost patience. E.g. where you faked the t-axis on a plot and where you made a big point about N501Y incorrectly attributing its rapid evolution to culture on humanized mice rather than regular mice. There's a reason that committing to written work is preferred. In this case, it saved me from a tedious critique of your dishonest tangents and allowed a focus on the more solid parts of the argument as written up by multiple viewers strongly favorable to your conclusions.
if you were to write up your Bayesian analysis it would be harder to brush past the enormous problems with the non-hierarchical extreme factors used. If you think the analysis here is wrong, it's written down for everyone to share and for you to respond to.
It's okay for a person to not have the patience to review an argument. If you make that choice, however, you should not profess to have reviewed or understood the content, nor should you criticize others for not reviewing arguments in sufficient detail.
I have likewise not read every version of your bayesian analysis, after sensing from a subset of text as well as from the summaries that you have no valuable insight into the subject of Covid origins, nor any arguments better than those put forth by Rootclaim. Scott likewise claims to have read none of your substack posts.
I think it's fine that I have not reviewed your work and think it lacks value. That said, it would be dishonest for me to write a detailed critique of your longest substack posts without reading them. I would also have to admit that there exists a possibility that there is some valuable argument put forth in one of your posts that would lead Scott or I to conclude that Covid has a lab origin. My bayesian assessment, however, holds that the odds that such an argument is present are very low.
I watched the key part from which the huge Bayes factor was derived, the part about how unlikely it would be to get the initial ascertained outbreak at HSM if the spill was at WIV. I watched parts of that more than once. The whole bit about inferring the reliability of an exponential extrapolation using the high value of R^2, etc, stuff to give stats types giggle fits. I wrote down in detail above what was wrong with that type of argument.
Scott needn't read my whole long-winded analysis, just the very brief response to his own analysis.
Rather than rehash old process stuff, why not specifically reply to the actual technical points?
The specific reason for not giving an itemized response, every time someone on the internet makes a new list of reasons why they think Covid is a lab leak, is that such lists have been created thousands of times, and will continue to be created.
The Rootclaim debate was very long, not because the case for zoonosis is long and complicated, but because the number of lab leak theories and objections to that case is nearly limitless. The fact that DRASTIC cofounder Yuri Deigin, multimillionaire Saar Wilf, and Saar's research team were unable to find any convincing argument against that case, after 18 hours of trying, is itself strong evidence that there is no good argument for the lab leak theory.
The "key part from which the huge Bayes factor was derived" is the fact that all epidemiological evidence in Wuhan clearly points to the pandemic starting at Huanan market, that starting point is very unlikely if SARS-CoV-2 was created at WIV, and 18 hours of objections to that theory can not produce a single valid reason why it's not true. There is, by contrast, no convincing genetic evidence that SARS-CoV-2 was engineered, and the odds of an epidemic starting in the same city with a lab are quite high. Bayesian analyses are toy models at best. Scott and both Rootclaim judges expressed great skepticism as to how well such models could solve a complicated question like this. That said, any unbiased analysis of that type would side towards zoonosis, just as the scientific consensus favors a zoonotic origin.
I've reviewed many lab leak arguments raised since the debate ended. I have found none convincing. I've debunked many of them on Twitter, including several claims on your new list.
I've responded to Levin:
https://x.com/tgof137/status/1886597906897232256
I've responded to Nod:
https://x.com/tgof137/status/1896008441317245055
I've even responded to your inane argument for ascertainment bias:
https://x.com/tgof137/status/1887674605403578638
I could do yet another debate rebutting the new, longer list. Perhaps that would stretch to 24 hours, instead of 18. After I won, the next round of conspiracy theorists could argue that debate is too long to watch, and then make new lists of itemized objections, which will also be wrong.
This phenomenon is the same with many other conspiracies and forms of pseudoscience. Anti-vaxxers have spent years writing different reasons why Covid vaccines are extremely dangerous or deadly. One could easily spend over 18 hours debating every version of these theories. And, in fact, Rootclaim has been doing just that, by taking the pro-vaccine side of a debate against Steve Kirsch. This time they're holding most of the discussion in writing, and the back and forth has been going on for over 6 months now. Steve Kirsch does not tire easily, and I'd imagine he could spend another 6 months making false claims about the danger of Covid vaccines, based on obscure data sources. I've long since stopped reading Steve's claims, and don't need to know what he's discovered in his latest secret data from New Zealand, or whatever, but I can see why Rootclaim would take the time to wade through all that nonsense for a million dollars.
If you'd like to put up 100k as an incentive for me to have another Covid origins debate, I'll gladly show why your "new and improved" list of lab leak claims is just as wrong as the previous one. You're also welcome to publish your claims and hope that scientists respond. Otherwise, I don't feel much need to reply to every new list of recycled lab leak claims some conspiracy theorist puts up on Substack or Twitter.
I have only a few points that seem reliable enough to give significant Bayes factors. They're quite similar to the first version, although the CGGCGG factor is substantially reduced and new data has led to me including the RE site factor, which I had initially rejected.
Ah, yes, the restriction sites paper. I also proposed that Valentin Bruttel try to defend his idea in a judged debate, but he refused, saying that any judges would not be able to understand its brilliance because he is Galileo, Einstein, and Semmelweis combined:
https://x.com/VBruttel/status/1913498526543127007
Virologists aren't convinced by the restriction sites argument. Saar and Yuri thought it wasn't reliable enough to even try presenting it at the Rootclaim debate. Even Guy Gadboit has done an analysis arguing against it, giving a bayes factor that leans towards a natural origin.
If this is the kind of stuff that underpins your Bayesian analysis, then it's just garbage in, garbage out.