30
votes
"I watched fifteen hours of COVID origins arguments so you don't have to"
Link information
This data is scraped automatically and may be incorrect.
- Title
- Practically-A-Book Review: Rootclaim $100,000 Lab Leak Debate
- Authors
- Scott Alexander
- Word count
- 12 171 words
I haven't followed "lab leak" arguments very closely and I'm not sure it's all that important, but I thought this was a decent summary of an in-depth lab leak debate, and I didn't have to watch any videos.
It gets very far into the weeds. Here's a quote from Peter Miller who is on the anti-lab-leak side:
Edit: long article is very long. I thought I was almost done, but there's a lot more after summarizing the debate.
Edit 2: Much of it isn't about Covid, it's about meta-issues. I guess I should stop making the argument that Rationalist are mostly using math as metaphor, because they're straight up admitting it now. From the beginning of the article:
And later:
...
Off topic, but this is my first time learning about raccoon dogs. They're so damn cute!
The Japanese have some wild legends about them.
Nice username.
Admission: I didn't read the article yet.
If he is anti-lab leak, isn't that kind of a pro-lab leak stance or am I reading that wrong?
He's arguing against the claim that the virus was designed for humans in "gain of function" research. Instead, he thinks it came from the racoon-dogs. (As an intermediate animal carrier.) If it weren't suited for racoon-dogs then it wouldn't have come from them.
Ah, thanks for the explanation. I read it as using them as test subjects.
I read it the same way at first, that the implication was the lab used raccoon dogs as test subjects. But I think what's actually meant is that it evolved in wild populations of raccoon dogs then jumped to humans.
The strongest reason I have for believing that Covid-19 was a zoonosis event is basically covered in the debate, but I'll quickly paraphrase it: when something is man-made like this, there are tell-tale markers. No such markers were found in the original Covid strains (Note that I am NOT a microbiologist, but I did read a lot of that during the various debates online, and it's believable because I do have a scientific background).
In addition... if it were a lab leak, who would've benefited? We saw how many people died during the height of the pandemic, and while obviously the Chinese gov't would want to silence any fault as much as possible- or more accurately, possibly try to keep any "We're sorry, we majorly fucked up the world" apologies hidden- it seems like the scale on which this happened should be too much for just one country to cover up. Rather, it seems to me like most of the "The Chinese are to blame for unleashing this!" talk comes from the Western hemisphere, where resentment for the Chinese gov't is already strong and this would just be a further way to blame China for something.
There’s an alternative scenario where the virus wasn’t engineered. It’s natural, but it was in the lab before it escaped. By accident. I don’t know why nobody really argued in favor of that.
You may be right about some people’s motivations, but it’s a dead end. Nothing gets settled that way, and we don’t learn anything from it, because it’s an excuse not to look for evidence.It seems like the debate was refreshingly free of talk about motivations.
Nobody needs an excuse to avoid spending time on this, but at the same time, I think it’s better to leave it as an open question.
I guess we should be happy that only 36% of the public picked “definitely true” or “definitely false.”
Probably the biggest reason why no one would try that line is because that already implicitly answers the question of its origin- that it already existed before it was in the lab. And its origin is the focus of the debate. Plus, I would add that if we went down that route, it's highly unlikely the lab would've had the only existing sample of it- while probably not 100% impossible, isolating it so that it exists only in the lab after coming from the natural world seems very, very difficult to do.
The WIV was well-known for its work on coronaviruses, so it wouldn’t be surprising to see it isolating the first samples of as-yet undiscovered viruses. And it wouldn’t be the first time a facility handling viruses has accidentally leaked biohazardous material. This was China’s first BSL-4 lab, and a point of pride for them. A leak would be very embarrassing, so it doesn’t surprise me they would try to cover it up.
I’m on the fence as to purely natural origin versus an accidental leak (I’m discounting secret bioweapons entirely). I don’t think we’ll be able to fully disprove a leak unless we get our hands on internal WIV or CPC communications (especially the later, they are usually a bit more forthright in their discussions that aren’t for public dissemination).
Perhaps I misunderstood, the debate is on the origin of the pandemic, or perhaps of the first human infection, and not necessarily of the virus itself.
There's also a third possibility, that it evolved "naturally" while being studied in the lab, that is, no engineering was done, but simply observation trhough multiple generations that led to its current iteration that was leaked by the lab.
Actually good post, I had a friend ask me about this recently, and I kind of dismissed the lab leak theory since to me it felt like too many coincidences needed to happen, and from what we know of how Covid spread it seemed unlikely it would have become endemic just from a single or extremely small number of lab workers, vs an alternative zoonotic origin, and I didn't think much of him asking me about it. I was floored that according to this 66% of people thought it was a lab leak? 4:1 ratio over zoonotic? that just seems outlandish, when I was talking to him I figured the Wuhan lab might make it I dunno, 5x more likely, 10x more likely to be lab leak? but my prior was probably 99:1 in absence of the lab, so 90:10 I'd still have said zoonotic, either way, what a cool debate and I guess idea, I learned more about the origin of covid than I did prior. Hope I can watch some highlights from the vids but I'm not gonna go through 15 hours worth.
edit: I meant actually good post (from ACXD) btw, I mean also from OP, but I wasn't trying to disparage OP, I just generally don't put a lot of stock in ACXD, but this article is well put together.
I enjoyed this post, I learned so much about COVID. I was expecting the usual skeptics approach from Peter of just playing defense against lab leak theories. But instead he took what I thought was a harder route of defending the alternate wet market hypothesis. I think he mainly won because he just had incredible knowledge and extremely strong arguments. But this defense also was interesting for the Bayesian scoring rules. You take a strong point of the lab leak (what are the odds it appears first in a city with a coronavirus lab?) and turns it into a weakness (what are the odds that we first find it at a wet market, given it was a leak).
Saar strikes me as the traditional "person who is smart in some areas but sure they're right in all of them, so they've been working backwards on how to prove that to everyone." rather than, you know, actually trying to apply good research and math to discover what's right in the first place. As they say, garbage in, garbage out and this sure feels like a case of that given the quality of evidence Saar proposes.
Some related quotes:
This is all I think anyone really needs to see. You can verify, in many of these claims, if Peter is lying. He is not. The fact Saar thought to include these claims in his argument, at all, is evidence of at best shoddy research and at worst, massive bias.
This is such amateur hour nonsense that it feels like Alex Jones dressed up in Math. I've also included the quotes where Peter, repeatedly, had to say again and again that it was NOT a super spreader event.
See, this is the thing. He's not a debate savant. He's just actually good at it. Most debate isn't hard, it's just that the internet, and specifically this weird subculture of self identifying "rationalists" are more often than not pretty fucking awful at it, and the standards are now very very low. Saar and his stand in are frankly pathetically awful at it, to the point that it looks like their opponent is amazing.
Finally-
And this, to me, is the key point. For whatever reason, there's this ego thing often tied to these "super smart" people, so incidents like COVID are the perfect trap for them to fall into. Early on, it's hard to tell. It should be unlikely, to be a lab leak, but yeah it could be a thing, so they pick a side.
Then, as more and more evidence develops, rather than say "yeah hmm...looks like at this point that's just not likely" they only dig in the heels more and more and more. Just as he's doing with this debate loss (it wasn't me, it was the FORMAT!).
This ego fragility isn't actually just limited to "super smart" people, and its the same issue you see all over the place, especially with modern politics. Sure trump seemed reasonable at the start, but now you know he's a fucking lunatic but you'll be damned if all those people who you argued with (and hell, may very well have treated you like trash about it) get to know they're right. It's just more obvious with successful people who can signal boost, and have the extra reinforcement of "well i've done all this other stuff right, so clearly i'm not wrong now".
If you're going to pick on the Rationalists because Saar turned out to bad at it, I think you should also admit that Peter is good at it, and that counts too.
This debate happening in such depth seems like one of the better things happening in the Rationalist community. It seems that a fair number of people changed their minds because of it - a rare victory. If you know of other debates that are this good, please share!
Sure. I'd just be willing to bet Peter (or how peter even handles this debate, not maybe all of them) is a minority in such a community. It's vaguely similar to the whole "not golfer" community argument by Tyson (who also does in fact fit the mold of smart guy who thinks they're smart at everything yes).
I think how good it is will depend mostly on the outcome from this. Ultimately the problem with rationalists are not unique to rationalists. Communities using traits as strong identifiers, often to ignore their flaws, is nothing new.
I do agree that the debate itself was a good thing, but rootclaims reaction is frankly awful. I worry that we'll get even more working backwards from the solution behavior as people rally around rootclaim because they want them to be right. Literally what Saar is doing.
I didn't get the "not golfer" reference. Tyson who? I agree that the Wilf's reaction is disappointing.
Talking about communities is rather weird sometimes because we don't have a clear idea who's a member. This is particularly true of online communities. Apparently Peter Miller and Saar Wilf met in the comment section of Astral Codex Ten.
By that criteria, am I a member? I don't consider myself one, more of a observer. Sometimes I've been a fan and certainly I've read quite a lot. Other times I've been more critical. I don't think reading a blog or leaving comments makes you a member. I've never met anyone in person who identifies as a Rationalist.
I can think of a lot of common themes and tropes. Having this debate, in this particular way, seems like a rather Rationalist thing to do. Being a fan of Bayesian probability is a Rationalist thing. But so is being critical of that? As Scott Alexander is, apparently.
So I would struggle to identify any particular membership traits in a rigorous way. There are some obvious central figures. This makes it hard to say who's a minority.
Neil deGrasse Tyson: "It's odd that the word atheist even exists. I don't play golf, is there a word for non-golf players? Do they gather and strategize? I can't do that, I can't gather around and talk about how much everyone doesn't believe in God. I don't have the energy for that."
Someone else already gave the answer, but in case you want the source/full context:
https://www.youtube.com/watch?v=CzSMC5rWvos
As silly as it is, I think that's basically all that matters. The problem is that you can identify as anything even if you're "doing it wrong". There are specific rules on say, being jewish or catholic. I can still adamantly claim to be one even though I meet literally none of the criteria.
This is more obvious with such beliefs due to how the criteria are laid out, but there's a lot of people who are rational, and then there are people who use rational thinking as a crutch to not need to be introspective and grow. It's hardly unique to rationalism, but I think it's signal boosted in certain circles, especially the internet, because it attracts a lot of coders and the like (because of course it would, no matter where on the scale you are)
Personally I've found the people who will strongly identify as rationalists are the least rational of the group. Perhaps my view is flawed, but I've seen a lot of "i must be right so let me figure out why" arguing because I think it's a group that is extra good at attracting that mindset.
I tend to favor answers to controversial questions which support a core bias I have, namely, that capitalism is bad for the world and provides a lot of freedom for powerful actors to get away with a lot of awful things. As such, I tend to draw conclusions and fill gaps based on my answers to the question, "Who benefits?" As a result, I'm persuaded that JFK's assassination involved a lot more than just one guy with emotional challenges, that 9/11 involved more than a few muslim fanatics, and, prior to today, that covid was probably the result of a lab leak.
Small point of clarification, the Cretan scripts are known as Linea_r_ A and B, not Linea_ge_. Perhaps this was a joke?
Anyway, based on this summary alone*, I've moved the needle from somewhat persuaded about lab leak to a little skeptical. Saar was especially weak in responding to Peter's arguments about Banal-52 and the credibility of identity of early cases.
Still, opacity breeds suspicion, and there was, and as far as I know still is, a fair amount of opacity and PR from both China and WHO; and what communications there are tend to be obfuscatory and Bernays-ian. If I had the power to limit and change what data were made available, the wet-market story is the one I would try to fabricate. Again, as far as I know, there are significant evidence and data gaps that appear to be politically driven.
*Side note, I'm highly skeptical of ultra-rationalism and trying to decide truth and reality based on mathematical probability. I say this as someone who studied epidemiology at UNC and aced biostatistics there. As the saying goes, there are lies, damn lies, and statistics. Math and discrete logic are very good at answering specific, very tailored questions. For example, "What is the geographical center of the residences of all covid positive human patients listed in available hospital records for the period from 12/1/2019 to 1/1/2020?" Determning bigger questions requires more abstract, divergent thinking. Chess is sometimes usefully illustrative, but by itself, is not a valid model for real geopolitics.
Yep, the image captions are jokes.
Asking “who benefits” gives us clues about where to look, but the problem is when people use it as a substitute for evidence. While incentives are pretty powerful in aggregate, people may disagree with us about what their interests are. Own goals are still possible. Personal sacrifice is still possible. Mental illness still happens. Overestimating incentives makes people seem more predictable than they are.
I thought this debate was pretty impressive and better than anything I’m likely to do on my own, but I still don’t see a real need to have an opinion on this. I leaned towards “not a lab leak” even before, but nothing I care about hinges on taking a side.
“Who benefits” and “follow the money” are good starting places for finding the starting points of conclusions and finding evidence but they work terribly as evidence themselves. The thing those questions answer is possible intent and motives, which are great, but also lead to wild conspiracy theories when used without evidence or, worse, when used to find evidence that only agrees with one’s predetermined answers to those questions.
China's MO here has not been to favor either origin hypothesis; rather, they want people to think that COVID didn't originate in China. Neither lab-leak nor zoonosis makes them look particularly good. Actually, I'd argue that zoonosis is slightly worse for them: a lab leak is something that, on a long enough timescale, will happen in any country performing gain-of-function research, whereas zoonosis is how the original SARS happened, and it's kind of embarrassing for the exact same thing to have happened twice.
This has happened far more than twice. China has been the source of a lot of novel and emerging pathogens. The year prior (roughly) there was an emerging encephalitic tick borne disease as well, they happen about/at least once a year from my memory. SARS-CoV-2 was just very successful, it was a respiratory virus with a great degree of transmission. The use of wet markets and consumption of bush meat is a high risk event. They are common across China and you see a lot of emergence due to things like this. Prior to the pandemic there had been a culling of pigs in China due to a swine fever outbreak and people were using wet market and bush meat more often. The region is well known for coronaviruses already which is part of why the virology institute was there. The year prior to the pandemic we had a virologist give a lecture basically saying this was bound to happen soon in about that same region.
Speaking as someone who worked in pathogen surveillance prior to and during the pandemic, the zoonosis origin always made sense. I started to wonder about the lab leak hypothesis when there were news reports about security agencies in the government that seems to give it more credence, because I trust the CIA has more information than I did, but I always trusted the zoonosis origin the most just because it is so unbelievably common and expected.
Apologies if these were covered in the article/videos.
As someone who has here previously defended lab-leak as plausible (albeit less likely than zoonosis), this really shifted my opinion. Based on what I read here, zoonosis seems significantly more likely, to the point that I'd call lab-leak only a remote possibility.
Tinfoil Hat Information including spoilers for one of the Three Body Problem books
I read the books late in the pandemic. One of the plotlines in the books involves an assassination attempt utilizing a virus that is very easily transmissable and non-fatal to everyone except one person's specific genes. While this doesn't perfectly map onto coronavirus, there are some striking similarities. Not that I needed any more cause for paranoia at the time.I haven’t got the patience to read through all of it but I gather they came to the conclusion that COVID is indeed of natural origin.
I just wanted to drop this link for anyone who is interested to see an in depth analysis that strongly supports the lab leak theory: https://cards.rootclaim.com/analysis/65892f07d97bdb00199d4366
Anecdotally I’ve read an in depth piece about the public letters to the government (apparently that’s a thing for public organizations in China) that hint at there being issues with keeping the lab sanitized due to the specific stainless steel they are using that’s corroding from the sanitation agents they use. Additionally there seemed to have been an incident somewhere late 2019 that caused a government official to visit the site. The letters involving the incident have then disappeared a couple months later. As I can’t find where I’ve read it take this with a grain of salt but maybe someone else has read the same thing and can provide a link.
Edit: I think this was it: https://www.propublica.org/article/senate-report-covid-19-origin-wuhan-lab
You just dropped a link from the guy arguing the lab leak in the debate covered in OP. It's not an independent analysis that agrees with one of the people in the debate, it is one of the arguments in the debate.
First, it's shocking to me that they would claim 94% confidence. Looking into how the site works, I'm shocked again to see that this site puts so much stock in their mathematical formula that they put it on their articles even though they may not have all the right inputs.
More importantly, just googling the name of the website gives me this:
https://www.reddit.com/r/slatestarcodex/comments/1au49wh/peter_miller_has_won_his_debate_with_rootclaim/
That 94% confidence has a GIGO problem. How do you even go about assigning probability in the face of amplified, manufactured, and selectively truthful sources? If your sources are biased, of course your math is going to show greater certainty in the direction of the bias.
The debate was between Saar Wilf, who founded rootclaim, and Peter Milller. Wilf offered to bet $100,000 that he would win against anyone in a debate, and Peter Miller took that bet and won the debate. I only read the summary in the blog post, but it’s clear that he’s a very impressive debater.
It’s a long blog post, but I recommend reading the first part for Peter’s arguments.
Thanks I didn’t catch that they literally got the guys from the link I posted into the debate. I’ll definitely give it a read now once I find the time.