After the mainstream media and tech giants tried to censor the lab leak hypothesis for an entire year, I did a deep dive on the arguments its proponents make. Unfortunately, the more I read, the less I was convinced. I now put the chances of a zoological origin at 85%, and the chances of a lab leak or deliberate release at 15%. I decided to organize my thoughts about proponents’ arguments and write them down. Are my analyses sound? Are there important arguments I missed that should swing a reasonable person’s judgment? Please let me know.
Argument: Three WIV staff got sick in November 2019, and this is significant.
Is this above the expected rate of seasonal illnesses? The WIV website (http://www.whiov.cas.cn/rcdw/yjy_160246/) lists 62 researchers, but doesn’t list administrators, janitors, IT people, or other staff. Let’s assume that in total, 100 people work at the WIV. If 15% of people get the flu every year, and the flu season lasts 3 months, you’d naively expect 5 people to get the flu in November 2019. Of course, flu is contagious, so if one person gets it, the chances of 8 people getting it become much higher.
The WSJ article says three people “sought hospital care”. It doesn’t say they were hospitalized. Here, the WSJ misunderstands Chinese culture. I love Chinese culture, but there’s no denying that Chinese people are hypochondriacs. They go to the hospital for every minor thing they experience, and aren’t satisfied until something is done for them. At the hospital, they sometimes get traditional Chinese medicine (a good approximation to placebo), or an IV drip (https://www.reddit.com/r/China/comments/2w5jy9/eli5_why_so_many_iv_drips_in_china/)--even when they just have a cold. Ask any Chinese friend who grew up in China and I’m sure that even if he’s never gotten IV for a cold or flu himself, someone in his immediate family has.
Even if the intelligence report explicitly said the WIV researchers had COVID in November, which it definitely didn’t, there would still be reasons to be skeptical. First, the intelligence document is never provided. Second, the WSJ article itself says “Current and former officials familiar with the intelligence about the lab researchers expressed differing views about the strength of the supporting evidence for the assessment. One person said that it was provided by an international partner and was potentially significant but still in need of further investigation and additional corroboration. Another person described the intelligence as stronger.” Third, how much do you trust the CIA, the same agency that lied Americans into war in Iraq, that orchestrated coups around the world, and that’s currently doing mass surveillance on Americans?
Argument: Wuhan has China’s only BSL-4 lab, and this coincidence is significant.
Analysis: China actually has two BSL-4 labs, the Wuhan Institute of Virology and the Harbin Veterinary Research Institute at Harbin. It is the Harbin institute that attracted controversy in 2013 (https://www.nature.com/articles/nature.2013.12925) for gain of function research with influenza.
Proponents of the lab leak hypothesis argue that safety standards in China are lax, and that coronavirus research is often done in BSL-3 or even BSL-2 labs. This is true. In fact, all SARS research was done in BSL-3 or lower labs before China’s first BSL-4 facility, the one in Wuhan, was finished in 2014. However, these two arguments are in tension with each other. There are at least 100 (https://zenodo.org/record/4067919) BSL-3 labs in the country, so if Chinese scientists don’t care about safety, the coincidence that one of China’s two BSL-4 labs is in Wuhan says little about the probability of a lab leak. Notably, the 4 Chinese SARS-CoV-1 escapes on record have all been from a Beijing lab, not the WIV.
Argument: we still haven’t found the natural animal reservoir COVID-19.
Analysis: In order for this to mean anything, it has to be remarkable. A look at other recent epidemics shows that scientists almost never identify the natural animal reservoir within 1.5 years of an epidemic outbreak. I went down Wikipedia’s List of Epidemics (https://en.wikipedia.org/wiki/List_of_epidemics#Major_epidemics_and_pandemics_by_death_toll) and focused only on those caused by new viruses:
2020 Congo, Ebola: Ebola has caused frequent outbreaks in Africa for decades, including in 2004, 2007, 2013-2016, and 2020. Only in 2019 did scientists discover the first bat carrying ebolavirus antibodies (https://www.sciencedaily.com/releases/2019/01/190124095156.htm), 43 years after the virus was first described.
2019 Nigeria, Lassa fever: This virus was first discovered in 1969. In 1972, the multimammate rat, Mastomys natalensis, was found to be the main animal reservoir.
2008, swine flu pandemic: Even though this started in Mexico, the geographic origin was thought to be Asia until 2016, when this paper (https://elifesciences.org/articles/16777) reported detecting several segments of genetic material in Mexican swine viruses that were, until then, unknown in the Americas. As far as I can tell, no exact match to the human virus has been found among pigs, nor is one expected due to the high mutation rate of influenza.
It’s even more illuminating to look at the other two coronaviruses that caused epidemics, MERS and SARS. MERS was discovered in 2012. In 2013, antibodies were found in camels, showing that camels are the natural reservoir. This is the only case, among all the epidemics I looked at, where scientists discovered the animal origin within 1.5 years. SARS caused widespread fears of a pandemic in 2002, but was successfully contained. It took 15 years (https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndrome) for scientists to track down the bats that it came from: “Around late 2017, Chinese scientists traced the virus through the intermediary of Asian palm civets to cave-dwelling horseshoe bats in Xiyang Yi Ethnic Township, Yunnan.[3]” The team that tracked down SARS’ origin was led by Shi Zhengli, director of the Center for Emerging Infectious Diseases at the Wuhan Institute of Virology, in collaboration with EcoHealth Alliance (https://www.ecohealthalliance.org/wp-content/uploads/2018/03/Virologica-Sinica-SARSr.pdf). If you don’t trust Shi or EcoHealth Alliance, then the origin of SARS is still unknown, 19 years later.
Why did it take so long to track down the origin of SARS, if the origins of MERS and Nipah were tracked down so quickly? MERS has a 30% fatality rate and spreads very poorly among humans, with R0 of around 0.3. Nipah has an even higher fatality rate of 50-75%, and spreads poorly because it can only spread through bodily fluids. The combination of severe symptoms and low human-to-human spread means MERS/Nipah is easy to detect early, and anyone who has it is only 1-2 degrees of separation away from the original animal. At that point, it’s just a matter of asking patients which animals they’ve been in contact with recently and testing them. SARS has a fatality rate of 10% (1% for people younger than 25) and a R0 of 2-3, meaning that SARS can spread much farther before anyone even realizes a new virus is afoot, making it much harder to identify the original animal. SARS-CoV-2 has a fatality rate of 1% (0.01% for young adults) and a similar R0 to SARs-CoV-1, meaning it can spread unnoticed for much longer before anyone notices.
We shouldn’t expect it to be as easy to track down the origin of SARS-CoV-2 as it was for MERS, which took 1 year. We shouldn’t even expect it to be as easy as it was for SARS, which took 14 years. We should expect it to be an order of magnitude harder than it was for SARS, and many orders of magnitude harder than it was for MERS.
Argument: How could the virus have jumped from Location X to Wuhan, Y km away, without anyone noticing in between?
Analysis: SARS originated from bats in Yunnan, but the outbreak was first detected in Guangdong. The two are 1300 km apart. If the proposed “Y” isn’t much larger than 1300 km, there’s no mystery. Even if it is much larger, COVID-19 has a much lower fatality rate than SARS, meaning it can spread undetected a lot farther than SARS. Also, 2019 China was a lot more interconnected by bus, rail, and plane than 2003 China. Wuhan is China’s ninth largest city and a major transport hub (https://upload.wikimedia.org/wikipedia/commons/4/4a/Rail_map_of_PRC.svg), so SARS-CoV-2 could really have jumped to humans anywhere in China.
Is there a good overview and assessment about the potential uses of ivermectin for COVID treatment? I have been trying to find one, but there is so much noise on both sides it's hard to make sense of it all. I was hoping Scott or Zvi would write an article about it. But until then, is there an alternative good and reputable source putting things in context and predicting what we will likely find out when the dust settles?
I replied to someone's comment a few weeks ago as a true believer in Zembrin. Well, I want to announce that I am now a true believer in double-blind clinical trials. I don't notice any effect anymore and I am willing to chalk up how I felt initially to a strong placebo effect. But I am stopping for a while now and will see if its perceived effects come back.
I'm looking for good research or discussions of intertemporal trade-offs for utilitarians. It seems reasonable to suspect that there will be fluctuations in the amount of money required to save a life. For simplicities sake, imagine if the number were to remain relatively fixed, then would it make sense to actually save money I would've donated to save a life and save more in the future? If the price of a life saved were to grow at 2% per year and I could get an investment return of 7% wouldn't it do the most good to wait until the end of my life? If the price to save lives was falling than this would make the case for waiting even stronger. I'm looking for discussions of this nature. Thanks
Has any thorough research been done yet on the possible effects that SSRIs have on the sexual development of pre-adolescents?
I read Scott's blog post on SSRIs and he mentioned (as a piece of anecdotal anthropic evidence that the sexual side effects are being underblown) that he took SSRIs when he was very young and now considers himself asexual, and is suspicious that the antidepressants are part of the explanation.
My friend and I also noticed that almost all of our acquaintances who identify as asexual also took SSRIs when they were children, which is what made us suspicious in the first place.
Since the existence of sexual side-effects from SSRIs is already well-established and is much more serious than initially believed (more anecdotal evidence: the friend that I mentioned above took SSRIs for thirty days under the prescription of his psychiatrist, and experienced serious sexual dysfunction during that month and during the month that followed), I can't help but wonder if there's a causal link here. SSRIs are known to seriously inhibit the sexual functioning of many people who take it. SSRIs are also diagnosed routinely to children as they are going through puberty, where they're supposed to be developing sexual interests for the first time. It doesn't seem unreasonable to posit that SSRIs are permanently inhibiting sexual development in a number of children.
I feel like this would be something that's hard to detect unless you're actually looking for it. For adults, you're able to actually ask them and they'll tell you if they're experiencing any sexual side effects. But you can't do that with adolescents, because (1) asking kids about their sexual activities isn't highly ethical, and (2) there's no A/B testing going on because the SSRIs are being prescribed before the patients are old enough to develop sexual interests in the first place. I'd like to see research that deliberately tries to address this concern, but I haven't found anything so far.
I've been wondering if it's worth having a strong opinion on this: if it's true that SSRIs can permanently inhibit sexual growth in adolescents and cause asexuality, and the risk is significant, then it means that children really shouldn't be prescribed SSRIs at all. On the other hand, I am not a physician or a pediatrician and I know virtually nothing about sexual development, so I can't help but be cautious when I find myself getting alarmed over something like this.
I've been working in software for 16 years now, and tech more broadly for 24. I'm starting to feel burned out. I won't bore you with a lengthy personal narrative but I will link this tweet...
...and summarize my frame of mind as "6 out of 7".
In any event, I'm interested to hear what other people have done. Did taking a year off help? Changing specialties? Starting your own thing? Finding a seat in a faceless megacorp and phoning it in?
I'm especially interested to hear from people that quit the business entirely. What do you do now? I struggle to come up with any kind of mental picture of Doing Something Else, and I'm hoping some stories will light my imagination.
I think some polyamorous people believe that (1) jealousy is entirely or mostly a socially constructed phenomena and that it can be overcome through reason. Alternatively, some believe that (2) some polyamorists have a different psychological profile and just do not experience jealousy. The final position I could imagine is that (3) jealousy is in almost everyone's genetic wiring but polyamorous people find a method to suppress it or keep it manageable. I am under the impression that most poly people probably believe (1), some probably (2) and almost nobody (3) but the critics of polyamory probably think the true state of affairs is (3) until it goes wrong.
Am I wrong to think that many polyamorous people are social constructivists with regards to jealousy? Anyone take the stance that it is in our nature and cannot be reasoned away except for some people? What are some other positions that I am missing? Thanks in advance.
Could the US function as a multiparty democracy and also a Presidential system? I've been increasingly onboard with the idea that the US should (somehow) transition to using a multiparty system and away from the Democrat-Republican dichotomy. Increasing partisan polarization seems self-evidently bad, and also somewhat undemocratic in that a number of people seem to want political options that just aren't on the menu in America. (Socially conservative, fiscally liberal seems to be the biggest unmet need).
With that being said- multiparty Presidential systems don't seem to have a great track record. This is the norm in the rest of the Americas, and supposedly Mexico, Brazil, and other Latin American countries 'prove' that it mostly doesn't work- the legislature is gridlocked and unable to pass bills, the President increasingly takes on more & more power to bypass a Congress that can't get anything done, he eventually becomes authoritarian, etc. Sounds like the road the US is already on, but worse.
The only other actually functional Presidential country is South Korea, but they seem to have two major parties making up 90% of the legislature. Unless you want to count semi-presidential systems like France, which just confuse me.
Does anyone want to argue for the US to have multiple parties, while retaining a Presidency? Could Congress really pass a budget or major legislation with 4-6 squabbling parties? I'm guessing the major US parties would be Center Left/Pro Business, America 1st/Populist Right, Green/Far Left, Country Club Republicans/Center Right, and Libertarians. Could legislators from these 5 parties really pass a budget every year? I dunno. Or is '2 major parties and a few really minor ones' actually a good system? This is like South Korea mentioned above, but also Australia, Britain, and I believe Canada (the last three obviously being parliamentary)
Did taking a social anxiety prescription medication improve anyone's overall courage level? I know this seems like a bit of a random question, so let me explain- I not only have generalized anxiety (especially in social situations), but I'm just not a particularly brave person in general. My CNS is very quick to the fight or flight response- I am just frightened by a lot of things. If someone confronts me verbally, I experience such an adrenaline rush that I find speaking difficult. I am jumpy and easily spooked by little things, a car backfiring, etc. Public speaking is completely off the table. I have always been physically pretty cowardly (afraid to jump off a cliff into a lake when all my friends were doing it, backed down from fistfights, and so on). I am now approaching middle age, so this seems to be a pretty fixed psychological trait. Oddly I am pretty physically durable and have done some fairly challenging things, so the 'deal with fatigue & pain' part of my brain seems to work fine- just not with fear.
When you are diagnosed with 'anxiety', supposedly there are a variety of prescription drugs that reduce or manage it. I've never taken any of them, but I was curious to hear from people who have- does a drug that (I suppose) makes you less anxious also make you less fearful if, say, someone comes up and angrily confronts you about an issue? Does your general fear response lessen? Is this a permanent effect? For example, I read that the North Hollywood bank robbers took phenobarbital before their robbery & massive shootout to calm their nerves, but I'm not clear if one can take that every day regularly. I will say that alcohol absolutely has this effect on me, both reducing anxiety & increasing courage, but obviously I would prefer not to drink heavily every day.
Maybe some mix of anti-anxiety meds and also testosterone would do the trick? Anyways, I'd be curious to hear people's experiences with various prescriptions. (I'm also going to experiment with magnesium supplementation, which I've heard can reduce anxiety)
After more than a year of corona, do we have any indication whether there is a dose effect? I.e., whether people who are infected through a higher virus dose get more sick?
It used to be a big question at some point, but I haven't heard anything. Did people realize that it's not a thing, or did they just give up?
Given the assumption that there'll be a nuclear war between the West and the PRC at a uniformly-distributed time in the next 15 years, what actions would best preserve one's life and livelihood?
Obvious #1 for "life" is "live somewhere that is not likely to be nuked or invaded". I'm thinking this resolves to "not the PRC, not Taiwan/Ryukyus/South Korea, not a large Western city".
Obvious #2 for "life" - at least, in a Western country - is "maintain sufficient food and water supplies to outlast a supply chain failure" (given that most supply chains are controlled from large cities).
But some of the details on #2 aren't so obvious - I'm not sure how long it'd take to re-activate supply chains. I'm also not sure how important personal-safety and anti-theft considerations are in this instance - local police exist and are obviously capable of acting on their own initiative, but I'm not sure of how well they'd cope with a starving populace. And I've got no solid ideas for "livelihood" (besides "don't invest your life savings in somewhere likely to get nuked or invaded") and I might be missing something for "life".
Over at TheMotte I had a debate about lockdowns where to TL;DR the other point of view:
Lockdowns are popular among the professional managerial class but unpopular among the working class. Polls showing that lockdowns were popular everywhere were run by the PMC and paid for by the PMC, and thus cannot be trusted.
My actual question is, assuming that is true and there is effort to make lockdowns look more popular than they really are. What hard evidence could we use to see through the spin, or to prove that the polls are accurate? I asked in that discussion but didn't get any answers, I thought I'd ask here because it is an interesting question.
Okay, I'm reading this Noah Smith article and the graph about doctors has me confuzzled. There are more *Copts*, relatively speaking, than the other ethnic group? Copts? As in "Coptic Christians from Egypt" who are a persecuted minority? Let me just try and find some population figures to give me a foothold here:
"Egyptian Copts are the biggest Christian community in the Arab world. Estimates of their numbers vary, but generally range between 4.7 and 7.1 million. They are proportionately most numerous in Upper Egypt."
"As of 2018, there were some 500,000 Copts living in the United States."
The short Wikipedia article seems to imply that Coptic immigration (prior to the Egyptian revolution of 2011) were the educated middle-class, with the poor(er), rural Copts only fleeing to America after the turmoil and upheaval.
So I'd expect more professionals to be the ones emigrating to the US, including doctors. And of course, once here, I'd expect families to steer their kids into the same professions as their parents, hence (possibly) more doctors.
How many Copts back in Egypt are/were doctors, and how did that stack up as a representative sample is, I think, the question that should be asked here. If every single accountant in Nowherestan emigrated to the USA, then I think the graph about "population demographics of accountants amongst ethnic groups" would be skewed so as to make it seem that Nowherestanis were disproportionately accountants by comparison with other ethnic groups.
Or maybe I'm an idiot who can't read a graph might be the simpler answer.
How does one get better at verbalizing? (writing / explaining / conveying ideas)
Practice is good, but deliberate practice is better. So far I've treated my verbalizing as a black box, where the more time I put into writing something, the more palatable it gets. But while it's gradually gotten better over the years, it's still agonizingly slow. Too slow for e.g. fluid IRL conversations.
Any tips for a more deliberate, maybe less blackboxy approach to improvement?
The stimulus for this comment is the banning of Bret Weinstein on youtube. The bigger picture is that we've given up our freedom of speech to corporations. This has happened slowly enough that most of us have gone along with it. Ban Alex Jones... that's easy, Ban Trump, He's telling lies that may cost lives, doesn't matter that he's the Pres. Ban all messages that we think may be wrong... It's like 1984, but with big business, not big brother. I'm sick of the comments that these are private corporations. I'm not sure what to do, any ideas, comments are welcome.
A project to find out whether trends in fiction can predict conflicts. The project was promising, but ended after five years. It's probably worth continuing with the premise.
"In 2018, weeks after the Bundeswehr officers had travelled to Tübingen, Wertheimer presented his initial findings at the defence ministry in Berlin. He drew attention to a literary scandal around Jovan Radulović’s 1983 play Dove Hole, about an Ustashe massacre against their Serbian neighbours, and the expulsion of non-Serbian writers from the Serbian Writers’ Association in 1986. In the years that followed, he showed, there was an absence of tales about Albanian-Serbian friendships or love stories, and a rise in revisionist historical novels. Literature and literary institutions, he told the military men, had “paved the way for war” a good decade before the start of the bloodshed of the Kosovo war in 1998.
"Carlo Masala was at the presentation. “At the beginning, I thought: this is crazy shit,” he recalls. “It won’t fly.” But Masala, who had spent a part of his academic career studying the conflict in Bosnia, remembered how the hardening tensions in the regions had been preceded by a decline in interfaith marriages. “In Kosovo, it seemed, you could detect similar early warning signs in the literary scene.”"
A recent meta-analysis of vitamin D supplementation studies found that vitamin D is probably effective against Covid. Have you changed your mind on this?
To provide a precise summary and collate the hitherto available clinical evidence on the effect of vitamin D supplementation on clinical outcomes in COVID-19 patients.
Methods
PubMed/MEDLINE, Scopus, and Web of Science databases were systematically searched using appropriate keywords till June 8, 2021, to identify observational studies and randomized controlled trials (RCTs) reporting adverse clinical outcomes (ICU admission and/or mortality) in COVID-19 patients receiving vitamin D supplementation vs. those not receiving the same. Both prior use and use of vitamin D after COVID-19 diagnosis were considered. Unadjusted/adjusted pooled odds ratio (OR) with 95% confidence intervals (CI) were calculated (PROSPERO registration number CRD42021248488).
Results
We identified 13 studies (10 observational, 3 RCTs) pooling data retrieved from 2933 COVID-19 patients. Pooled analysis of unadjusted data showed that vitamin D use in COVID-19 was significantly associated with reduced ICU admission/mortality (OR 0.41, 95% CI: 0.20, 0.81, p = 0.01, I2 = 66%, random-effects model). Similarly, on pooling adjusted risk estimates, vitamin D was also found to reduce the risk of adverse outcomes (pooled OR 0.27, 95% CI: 0.08, 0.91, p = 0.03, I2 = 80%, random-effects model). Subgroup analysis showed that vitamin D supplementation was associated with improved clinical outcomes only in patients receiving the drug post-COVID-19 diagnosis and not in those who had received vitamin D before diagnosis.
Conclusions
Vitamin D supplementation might be associated with improved clinical outcomes, especially when administered after the diagnosis of COVID-19. However, issues regarding the appropriate dose, duration, and mode of administration of vitamin D remain unanswered and need further research.
John Schilling brought up the question of offsets for murder in a no-politics thread, and trade-offs (if not offsets) for killing are exactly the issue for how aggressive police should be permitted to be.
Do more aggressive police result in less crime? Do the police become criminals themselves if they're cut too much slack? Does restricting police aggressiveness dispirit the police so much that they quit doing useful work? (Sidetrack: A lot of the police seem to have no faith in the justice system if it's being applied to themselves.)
There's a common libertarian refrain that "if it is moral to perform some action for free, then it is moral to perform that action for money." I currently believe this statement, but I want to challenge that belief and see if it survives. I'm willing to bite this bullet when to comes to both prostitution and blackmail. Does anyone have any good counter-examples?
An old post by Scott https://slatestarcodex.com/2013/09/12/the-life-cycle-of-medical-ideas/ mentions minocyclyne, an antibiotic that was found to help schizophrenia patients with treatment-resistent symptoms but even after some research confirming the findings no one seemed to want to prescribe it. Is this still the situation with the drug today? (From the wikipedia article it seems that it is.)
Regarding the delta variant, I decided to plot the Covid-19 infection vs death rate of the UK to see whether the recent increase in cases corresponds to an increase in deaths.
The recent increase due to the delta variant is shown at the bottom right end of the line. This shows a lower CFR (c. 0.3%) for the latest wave than for previous waves.
Two friends - let's call them Yvonne and Zachary, for a change, to let Alice and Bob have some time off - have, once again, been captured by a mad jailer.
He's going to show Yvonne a chessboard full of 64 coins (perfectly ordinary coins, no double-headers or anything like that), each showing either heads or tails in some pattern of his choice. He's going to hide the key to freedom under one coin, letting Yvonne see which one it is.
Yvonne gets to turn over exactly one coin, and will then be escorted back to her cell. The jailer will then show Zachary the board. Zachary gets to look under one coin, and if the key isn't there then they're both executed.
If they get to communicate beforehand, what is their best strategy? Does it make a difference if the jailer is listening in on their conversation?
I've done a huge number of podcasts (Future Strategist Podcast) with Greg Cochran, a co-author of the paper Noah Smith links to in his "Genetic superpowers" section. Based on what Greg has told me there is a massive attempt in academia to silence any discussion of the genetics of Jewish intelligence. The attitude Noah takes in his reply is that while he doesn't find the issue interesting, he would have no objection to others looking into it. This is analogous to someone in the Soviet Union in 1950 saying "while I can't think of anything about Stalin to criticize, you should feel free to honestly critique him if you feel so inclined." As Greg wrote in his blog about a co-author of his Jewish intelligence paper "A certain person we will not name, famous for the discovery of the double helix, once said 'Harpending must really have balls of steel, in order to take a genetic look at Jewish intelligence.'” Scott, you once criticized me for saying that you are brave, and because of that I'm not going to comment on your balls.
I've got an idiot thought experiment re. trust to get out there. Don't take it too seriously, just gotta post it so I can forget it.
An omnipotent space bat has come to US, and presents you with a situation:
The population has been split perfectly in half by political alignment. You have been selected as the median representative of your alignment. You have a button. If your press the button, the other wing is instantly disenfranchised without any other harm being done to them.
If you don't press the button, the bat will offer the same choice to the median representative of the other wing.
1. Do you press the button?
2. What if it was a random member of the other wing that gets a shot at the button?
3. What if n random members got a chance at the button? What number n would get you to press?
4. What if the most extreme member of the opposite wing got the button?
Now, the same again, but the button causes some level of harm in addition to disenfranchisement.
What level of harm (you lose 100$ ----> You are a second class citizen ------> death) causes you to the press the button/not press the button?
I remember reading on SSC that all psychotherapy schools are equivalent, and that the result only depends on the therapist's IQ (or something like that, I don't remember well). Can someone point me to a good discussion about this? (Or clear up my confusion?) I said this to a psychology student and he replied that this was totally false.
Are there any inner circle LW / REACH rationalists on the ACX these days? By inner circle I mean someone who's active in the community offline. Well besides Scott himself obviously.
My son is interested in Transcendental Meditation and just got back from a weekend retreat out of state. Is there any organizations I can point him toward that are more legit than others?
I worry about any group discovered over the internet :).
Anyone any suggestion on a good book on history of the worker movement (from industrial revolution up to now; world-wide, not US-specific)? While it's generally easy to find a book on any topic I am having a hard time with this one. Maybe I am using the wrong search words? Or is it really that historians don't care much about the topic?
AI risk question: what's the plausible path from GPT3-esque AI to a AI X-risk doom scenario? I get the classic "paperclip maximizer" argument of why an unfriendly AI can arise unintentionally, but I don't see how it can happen with a design like GPT-3. Specifically, my understanding is that GPT-3, like many of the current breed of machine learning, is trained on a dataset and then later is run on the novel input *without further training*. So the minimization process is done entirely offline at the start with a fixed dataset and the goal is to minimize errors the AI makes at a task like "complete this piece of the dataset given just the start".
How can this ever get to an unfriendly AI? It's not "in a box", the AI is sorta trying to maximize how likely you are to 'accept' it's predicted output and so it could theoretically start outputting things like "Accept this text or you'll never see your children again" and go rogue. But there's no maximization process bringing it to that result - such a threat would never work on the training dataset since it's already fixed and so threats cannot be learned by the AI. Is there some other plausible path for it have unfriendly behavior?
This contrasts to the situation of online reinforcement learning where I can imagine things going out of control. I can't imagine that scenario for this style of pre-trained AI, yet that seems like too simple a solution to the friendly AGI problem. Thoughts?
(And what's the right terminology for this distinction between kinds of AI? Or am I mistaken about there being such a distinction?)
It will only open "actual" posts, so no open threads, meetups, or links posts. Developed it originally because it was a hassle to scroll through SSC's archive every time I wanted to read a post — now it works for most other blogs as well, although I haven't found a way to randomize ACX yet.
I read a post on what I believe was somebody's blog a few months back, and it had to do with the oil shock, and I believe it argued it preceded Bretton Woods. Does this sound familiar to anyone? I'm trying to find that post!
One data point that sticks out in my mind is the comparison between India and China. Both countries have populations of approximately 1.4 billion yet India has had just under 31 million cases and China only 92,000? Most would surmise that the Chinese data is incorrect, but what if it is not? What could account for such an unheard of discrepancy? After reading your article on lockdown effectiveness, is there any combination of actions that could have resulted in this? It makes me wonder if China already had a vaccine and had administered it.
After the mainstream media and tech giants tried to censor the lab leak hypothesis for an entire year, I did a deep dive on the arguments its proponents make. Unfortunately, the more I read, the less I was convinced. I now put the chances of a zoological origin at 85%, and the chances of a lab leak or deliberate release at 15%. I decided to organize my thoughts about proponents’ arguments and write them down. Are my analyses sound? Are there important arguments I missed that should swing a reasonable person’s judgment? Please let me know.
Argument: Three WIV staff got sick in November 2019, and this is significant.
Analysis: This WSJ report (https://www.wsj.com/articles/intelligence-on-sick-staff-at-wuhan-lab-fuels-debate-on-covid-19-origin-11621796228) claims that according to “a previously undisclosed US intelligence report”, “three researchers from China’s Wuhan Institute of Virology became sick enough in November 2019 that they sought hospital care”. A State Department fact sheet (https://ge.usembassy.gov/fact-sheet-activity-at-the-wuhan-institute-of-virology/) says “The U.S. government has reason to believe that several researchers inside the WIV became sick in autumn 2019, before the first identified case of the outbreak, with symptoms consistent with both COVID-19 and common seasonal illnesses.”
Is this above the expected rate of seasonal illnesses? The WIV website (http://www.whiov.cas.cn/rcdw/yjy_160246/) lists 62 researchers, but doesn’t list administrators, janitors, IT people, or other staff. Let’s assume that in total, 100 people work at the WIV. If 15% of people get the flu every year, and the flu season lasts 3 months, you’d naively expect 5 people to get the flu in November 2019. Of course, flu is contagious, so if one person gets it, the chances of 8 people getting it become much higher.
The WSJ article says three people “sought hospital care”. It doesn’t say they were hospitalized. Here, the WSJ misunderstands Chinese culture. I love Chinese culture, but there’s no denying that Chinese people are hypochondriacs. They go to the hospital for every minor thing they experience, and aren’t satisfied until something is done for them. At the hospital, they sometimes get traditional Chinese medicine (a good approximation to placebo), or an IV drip (https://www.reddit.com/r/China/comments/2w5jy9/eli5_why_so_many_iv_drips_in_china/)--even when they just have a cold. Ask any Chinese friend who grew up in China and I’m sure that even if he’s never gotten IV for a cold or flu himself, someone in his immediate family has.
Even if the intelligence report explicitly said the WIV researchers had COVID in November, which it definitely didn’t, there would still be reasons to be skeptical. First, the intelligence document is never provided. Second, the WSJ article itself says “Current and former officials familiar with the intelligence about the lab researchers expressed differing views about the strength of the supporting evidence for the assessment. One person said that it was provided by an international partner and was potentially significant but still in need of further investigation and additional corroboration. Another person described the intelligence as stronger.” Third, how much do you trust the CIA, the same agency that lied Americans into war in Iraq, that orchestrated coups around the world, and that’s currently doing mass surveillance on Americans?
Argument: Wuhan has China’s only BSL-4 lab, and this coincidence is significant.
Analysis: China actually has two BSL-4 labs, the Wuhan Institute of Virology and the Harbin Veterinary Research Institute at Harbin. It is the Harbin institute that attracted controversy in 2013 (https://www.nature.com/articles/nature.2013.12925) for gain of function research with influenza.
Proponents of the lab leak hypothesis argue that safety standards in China are lax, and that coronavirus research is often done in BSL-3 or even BSL-2 labs. This is true. In fact, all SARS research was done in BSL-3 or lower labs before China’s first BSL-4 facility, the one in Wuhan, was finished in 2014. However, these two arguments are in tension with each other. There are at least 100 (https://zenodo.org/record/4067919) BSL-3 labs in the country, so if Chinese scientists don’t care about safety, the coincidence that one of China’s two BSL-4 labs is in Wuhan says little about the probability of a lab leak. Notably, the 4 Chinese SARS-CoV-1 escapes on record have all been from a Beijing lab, not the WIV.
Argument: we still haven’t found the natural animal reservoir COVID-19.
Analysis: In order for this to mean anything, it has to be remarkable. A look at other recent epidemics shows that scientists almost never identify the natural animal reservoir within 1.5 years of an epidemic outbreak. I went down Wikipedia’s List of Epidemics (https://en.wikipedia.org/wiki/List_of_epidemics#Major_epidemics_and_pandemics_by_death_toll) and focused only on those caused by new viruses:
2020 Congo, Ebola: Ebola has caused frequent outbreaks in Africa for decades, including in 2004, 2007, 2013-2016, and 2020. Only in 2019 did scientists discover the first bat carrying ebolavirus antibodies (https://www.sciencedaily.com/releases/2019/01/190124095156.htm), 43 years after the virus was first described.
2019 Nigeria, Lassa fever: This virus was first discovered in 1969. In 1972, the multimammate rat, Mastomys natalensis, was found to be the main animal reservoir.
2018 India, Nipah virus: First isolated in 1999. Fruit bats were identified (https://sci-hub.do/https://pubmed.ncbi.nlm.nih.gov/16847084/) as the natural reservoir in 2001.
2008, swine flu pandemic: Even though this started in Mexico, the geographic origin was thought to be Asia until 2016, when this paper (https://elifesciences.org/articles/16777) reported detecting several segments of genetic material in Mexican swine viruses that were, until then, unknown in the Americas. As far as I can tell, no exact match to the human virus has been found among pigs, nor is one expected due to the high mutation rate of influenza.
It’s even more illuminating to look at the other two coronaviruses that caused epidemics, MERS and SARS. MERS was discovered in 2012. In 2013, antibodies were found in camels, showing that camels are the natural reservoir. This is the only case, among all the epidemics I looked at, where scientists discovered the animal origin within 1.5 years. SARS caused widespread fears of a pandemic in 2002, but was successfully contained. It took 15 years (https://en.wikipedia.org/wiki/Severe_acute_respiratory_syndrome) for scientists to track down the bats that it came from: “Around late 2017, Chinese scientists traced the virus through the intermediary of Asian palm civets to cave-dwelling horseshoe bats in Xiyang Yi Ethnic Township, Yunnan.[3]” The team that tracked down SARS’ origin was led by Shi Zhengli, director of the Center for Emerging Infectious Diseases at the Wuhan Institute of Virology, in collaboration with EcoHealth Alliance (https://www.ecohealthalliance.org/wp-content/uploads/2018/03/Virologica-Sinica-SARSr.pdf). If you don’t trust Shi or EcoHealth Alliance, then the origin of SARS is still unknown, 19 years later.
Why did it take so long to track down the origin of SARS, if the origins of MERS and Nipah were tracked down so quickly? MERS has a 30% fatality rate and spreads very poorly among humans, with R0 of around 0.3. Nipah has an even higher fatality rate of 50-75%, and spreads poorly because it can only spread through bodily fluids. The combination of severe symptoms and low human-to-human spread means MERS/Nipah is easy to detect early, and anyone who has it is only 1-2 degrees of separation away from the original animal. At that point, it’s just a matter of asking patients which animals they’ve been in contact with recently and testing them. SARS has a fatality rate of 10% (1% for people younger than 25) and a R0 of 2-3, meaning that SARS can spread much farther before anyone even realizes a new virus is afoot, making it much harder to identify the original animal. SARS-CoV-2 has a fatality rate of 1% (0.01% for young adults) and a similar R0 to SARs-CoV-1, meaning it can spread unnoticed for much longer before anyone notices.
We shouldn’t expect it to be as easy to track down the origin of SARS-CoV-2 as it was for MERS, which took 1 year. We shouldn’t even expect it to be as easy as it was for SARS, which took 14 years. We should expect it to be an order of magnitude harder than it was for SARS, and many orders of magnitude harder than it was for MERS.
Argument: How could the virus have jumped from Location X to Wuhan, Y km away, without anyone noticing in between?
Analysis: SARS originated from bats in Yunnan, but the outbreak was first detected in Guangdong. The two are 1300 km apart. If the proposed “Y” isn’t much larger than 1300 km, there’s no mystery. Even if it is much larger, COVID-19 has a much lower fatality rate than SARS, meaning it can spread undetected a lot farther than SARS. Also, 2019 China was a lot more interconnected by bus, rail, and plane than 2003 China. Wuhan is China’s ninth largest city and a major transport hub (https://upload.wikimedia.org/wikipedia/commons/4/4a/Rail_map_of_PRC.svg), so SARS-CoV-2 could really have jumped to humans anywhere in China.
Is there a good overview and assessment about the potential uses of ivermectin for COVID treatment? I have been trying to find one, but there is so much noise on both sides it's hard to make sense of it all. I was hoping Scott or Zvi would write an article about it. But until then, is there an alternative good and reputable source putting things in context and predicting what we will likely find out when the dust settles?
Wrote about Heidegger and what we can learn from mood https://whatiscalledthinking.substack.com/p/what-can-we-learn-from-mood
I replied to someone's comment a few weeks ago as a true believer in Zembrin. Well, I want to announce that I am now a true believer in double-blind clinical trials. I don't notice any effect anymore and I am willing to chalk up how I felt initially to a strong placebo effect. But I am stopping for a while now and will see if its perceived effects come back.
I'm looking for good research or discussions of intertemporal trade-offs for utilitarians. It seems reasonable to suspect that there will be fluctuations in the amount of money required to save a life. For simplicities sake, imagine if the number were to remain relatively fixed, then would it make sense to actually save money I would've donated to save a life and save more in the future? If the price of a life saved were to grow at 2% per year and I could get an investment return of 7% wouldn't it do the most good to wait until the end of my life? If the price to save lives was falling than this would make the case for waiting even stronger. I'm looking for discussions of this nature. Thanks
> some people might need to run questions by an IRB first
Isn't this nightmare stuff? https://slatestarcodex.com/2017/08/29/my-irb-nightmare/
Has any thorough research been done yet on the possible effects that SSRIs have on the sexual development of pre-adolescents?
I read Scott's blog post on SSRIs and he mentioned (as a piece of anecdotal anthropic evidence that the sexual side effects are being underblown) that he took SSRIs when he was very young and now considers himself asexual, and is suspicious that the antidepressants are part of the explanation.
My friend and I also noticed that almost all of our acquaintances who identify as asexual also took SSRIs when they were children, which is what made us suspicious in the first place.
Since the existence of sexual side-effects from SSRIs is already well-established and is much more serious than initially believed (more anecdotal evidence: the friend that I mentioned above took SSRIs for thirty days under the prescription of his psychiatrist, and experienced serious sexual dysfunction during that month and during the month that followed), I can't help but wonder if there's a causal link here. SSRIs are known to seriously inhibit the sexual functioning of many people who take it. SSRIs are also diagnosed routinely to children as they are going through puberty, where they're supposed to be developing sexual interests for the first time. It doesn't seem unreasonable to posit that SSRIs are permanently inhibiting sexual development in a number of children.
I feel like this would be something that's hard to detect unless you're actually looking for it. For adults, you're able to actually ask them and they'll tell you if they're experiencing any sexual side effects. But you can't do that with adolescents, because (1) asking kids about their sexual activities isn't highly ethical, and (2) there's no A/B testing going on because the SSRIs are being prescribed before the patients are old enough to develop sexual interests in the first place. I'd like to see research that deliberately tries to address this concern, but I haven't found anything so far.
I've been wondering if it's worth having a strong opinion on this: if it's true that SSRIs can permanently inhibit sexual growth in adolescents and cause asexuality, and the risk is significant, then it means that children really shouldn't be prescribed SSRIs at all. On the other hand, I am not a physician or a pediatrician and I know virtually nothing about sexual development, so I can't help but be cautious when I find myself getting alarmed over something like this.
I've been working in software for 16 years now, and tech more broadly for 24. I'm starting to feel burned out. I won't bore you with a lengthy personal narrative but I will link this tweet...
https://mobile.twitter.com/john_blum/status/1407765390797590530/photo/1
...and summarize my frame of mind as "6 out of 7".
In any event, I'm interested to hear what other people have done. Did taking a year off help? Changing specialties? Starting your own thing? Finding a seat in a faceless megacorp and phoning it in?
I'm especially interested to hear from people that quit the business entirely. What do you do now? I struggle to come up with any kind of mental picture of Doing Something Else, and I'm hoping some stories will light my imagination.
I think some polyamorous people believe that (1) jealousy is entirely or mostly a socially constructed phenomena and that it can be overcome through reason. Alternatively, some believe that (2) some polyamorists have a different psychological profile and just do not experience jealousy. The final position I could imagine is that (3) jealousy is in almost everyone's genetic wiring but polyamorous people find a method to suppress it or keep it manageable. I am under the impression that most poly people probably believe (1), some probably (2) and almost nobody (3) but the critics of polyamory probably think the true state of affairs is (3) until it goes wrong.
Am I wrong to think that many polyamorous people are social constructivists with regards to jealousy? Anyone take the stance that it is in our nature and cannot be reasoned away except for some people? What are some other positions that I am missing? Thanks in advance.
Could the US function as a multiparty democracy and also a Presidential system? I've been increasingly onboard with the idea that the US should (somehow) transition to using a multiparty system and away from the Democrat-Republican dichotomy. Increasing partisan polarization seems self-evidently bad, and also somewhat undemocratic in that a number of people seem to want political options that just aren't on the menu in America. (Socially conservative, fiscally liberal seems to be the biggest unmet need).
With that being said- multiparty Presidential systems don't seem to have a great track record. This is the norm in the rest of the Americas, and supposedly Mexico, Brazil, and other Latin American countries 'prove' that it mostly doesn't work- the legislature is gridlocked and unable to pass bills, the President increasingly takes on more & more power to bypass a Congress that can't get anything done, he eventually becomes authoritarian, etc. Sounds like the road the US is already on, but worse.
The only other actually functional Presidential country is South Korea, but they seem to have two major parties making up 90% of the legislature. Unless you want to count semi-presidential systems like France, which just confuse me.
Does anyone want to argue for the US to have multiple parties, while retaining a Presidency? Could Congress really pass a budget or major legislation with 4-6 squabbling parties? I'm guessing the major US parties would be Center Left/Pro Business, America 1st/Populist Right, Green/Far Left, Country Club Republicans/Center Right, and Libertarians. Could legislators from these 5 parties really pass a budget every year? I dunno. Or is '2 major parties and a few really minor ones' actually a good system? This is like South Korea mentioned above, but also Australia, Britain, and I believe Canada (the last three obviously being parliamentary)
Did taking a social anxiety prescription medication improve anyone's overall courage level? I know this seems like a bit of a random question, so let me explain- I not only have generalized anxiety (especially in social situations), but I'm just not a particularly brave person in general. My CNS is very quick to the fight or flight response- I am just frightened by a lot of things. If someone confronts me verbally, I experience such an adrenaline rush that I find speaking difficult. I am jumpy and easily spooked by little things, a car backfiring, etc. Public speaking is completely off the table. I have always been physically pretty cowardly (afraid to jump off a cliff into a lake when all my friends were doing it, backed down from fistfights, and so on). I am now approaching middle age, so this seems to be a pretty fixed psychological trait. Oddly I am pretty physically durable and have done some fairly challenging things, so the 'deal with fatigue & pain' part of my brain seems to work fine- just not with fear.
When you are diagnosed with 'anxiety', supposedly there are a variety of prescription drugs that reduce or manage it. I've never taken any of them, but I was curious to hear from people who have- does a drug that (I suppose) makes you less anxious also make you less fearful if, say, someone comes up and angrily confronts you about an issue? Does your general fear response lessen? Is this a permanent effect? For example, I read that the North Hollywood bank robbers took phenobarbital before their robbery & massive shootout to calm their nerves, but I'm not clear if one can take that every day regularly. I will say that alcohol absolutely has this effect on me, both reducing anxiety & increasing courage, but obviously I would prefer not to drink heavily every day.
Maybe some mix of anti-anxiety meds and also testosterone would do the trick? Anyways, I'd be curious to hear people's experiences with various prescriptions. (I'm also going to experiment with magnesium supplementation, which I've heard can reduce anxiety)
After more than a year of corona, do we have any indication whether there is a dose effect? I.e., whether people who are infected through a higher virus dose get more sick?
It used to be a big question at some point, but I haven't heard anything. Did people realize that it's not a thing, or did they just give up?
Given the assumption that there'll be a nuclear war between the West and the PRC at a uniformly-distributed time in the next 15 years, what actions would best preserve one's life and livelihood?
Obvious #1 for "life" is "live somewhere that is not likely to be nuked or invaded". I'm thinking this resolves to "not the PRC, not Taiwan/Ryukyus/South Korea, not a large Western city".
Obvious #2 for "life" - at least, in a Western country - is "maintain sufficient food and water supplies to outlast a supply chain failure" (given that most supply chains are controlled from large cities).
But some of the details on #2 aren't so obvious - I'm not sure how long it'd take to re-activate supply chains. I'm also not sure how important personal-safety and anti-theft considerations are in this instance - local police exist and are obviously capable of acting on their own initiative, but I'm not sure of how well they'd cope with a starving populace. And I've got no solid ideas for "livelihood" (besides "don't invest your life savings in somewhere likely to get nuked or invaded") and I might be missing something for "life".
Thoughts?
Over at TheMotte I had a debate about lockdowns where to TL;DR the other point of view:
Lockdowns are popular among the professional managerial class but unpopular among the working class. Polls showing that lockdowns were popular everywhere were run by the PMC and paid for by the PMC, and thus cannot be trusted.
My actual question is, assuming that is true and there is effort to make lockdowns look more popular than they really are. What hard evidence could we use to see through the spin, or to prove that the polls are accurate? I asked in that discussion but didn't get any answers, I thought I'd ask here because it is an interesting question.
Okay, I'm reading this Noah Smith article and the graph about doctors has me confuzzled. There are more *Copts*, relatively speaking, than the other ethnic group? Copts? As in "Coptic Christians from Egypt" who are a persecuted minority? Let me just try and find some population figures to give me a foothold here:
https://minorityrights.org/minorities/copts/
"Egyptian Copts are the biggest Christian community in the Arab world. Estimates of their numbers vary, but generally range between 4.7 and 7.1 million. They are proportionately most numerous in Upper Egypt."
https://en.wikipedia.org/wiki/Coptic_Americans
"As of 2018, there were some 500,000 Copts living in the United States."
The short Wikipedia article seems to imply that Coptic immigration (prior to the Egyptian revolution of 2011) were the educated middle-class, with the poor(er), rural Copts only fleeing to America after the turmoil and upheaval.
So I'd expect more professionals to be the ones emigrating to the US, including doctors. And of course, once here, I'd expect families to steer their kids into the same professions as their parents, hence (possibly) more doctors.
How many Copts back in Egypt are/were doctors, and how did that stack up as a representative sample is, I think, the question that should be asked here. If every single accountant in Nowherestan emigrated to the USA, then I think the graph about "population demographics of accountants amongst ethnic groups" would be skewed so as to make it seem that Nowherestanis were disproportionately accountants by comparison with other ethnic groups.
Or maybe I'm an idiot who can't read a graph might be the simpler answer.
How does one get better at verbalizing? (writing / explaining / conveying ideas)
Practice is good, but deliberate practice is better. So far I've treated my verbalizing as a black box, where the more time I put into writing something, the more palatable it gets. But while it's gradually gotten better over the years, it's still agonizingly slow. Too slow for e.g. fluid IRL conversations.
Any tips for a more deliberate, maybe less blackboxy approach to improvement?
The stimulus for this comment is the banning of Bret Weinstein on youtube. The bigger picture is that we've given up our freedom of speech to corporations. This has happened slowly enough that most of us have gone along with it. Ban Alex Jones... that's easy, Ban Trump, He's telling lies that may cost lives, doesn't matter that he's the Pres. Ban all messages that we think may be wrong... It's like 1984, but with big business, not big brother. I'm sick of the comments that these are private corporations. I'm not sure what to do, any ideas, comments are welcome.
https://www.theguardian.com/lifeandstyle/2021/jun/26/project-cassandra-plan-to-use-novels-to-predict-next-war
A project to find out whether trends in fiction can predict conflicts. The project was promising, but ended after five years. It's probably worth continuing with the premise.
"In 2018, weeks after the Bundeswehr officers had travelled to Tübingen, Wertheimer presented his initial findings at the defence ministry in Berlin. He drew attention to a literary scandal around Jovan Radulović’s 1983 play Dove Hole, about an Ustashe massacre against their Serbian neighbours, and the expulsion of non-Serbian writers from the Serbian Writers’ Association in 1986. In the years that followed, he showed, there was an absence of tales about Albanian-Serbian friendships or love stories, and a rise in revisionist historical novels. Literature and literary institutions, he told the military men, had “paved the way for war” a good decade before the start of the bloodshed of the Kosovo war in 1998.
"Carlo Masala was at the presentation. “At the beginning, I thought: this is crazy shit,” he recalls. “It won’t fly.” But Masala, who had spent a part of his academic career studying the conflict in Bosnia, remembered how the hardening tensions in the regions had been preceded by a decline in interfaith marriages. “In Kosovo, it seemed, you could detect similar early warning signs in the literary scene.”"
A recent meta-analysis of vitamin D supplementation studies found that vitamin D is probably effective against Covid. Have you changed your mind on this?
https://link.springer.com/article/10.1007/s40618-021-01614-4
Purpose
To provide a precise summary and collate the hitherto available clinical evidence on the effect of vitamin D supplementation on clinical outcomes in COVID-19 patients.
Methods
PubMed/MEDLINE, Scopus, and Web of Science databases were systematically searched using appropriate keywords till June 8, 2021, to identify observational studies and randomized controlled trials (RCTs) reporting adverse clinical outcomes (ICU admission and/or mortality) in COVID-19 patients receiving vitamin D supplementation vs. those not receiving the same. Both prior use and use of vitamin D after COVID-19 diagnosis were considered. Unadjusted/adjusted pooled odds ratio (OR) with 95% confidence intervals (CI) were calculated (PROSPERO registration number CRD42021248488).
Results
We identified 13 studies (10 observational, 3 RCTs) pooling data retrieved from 2933 COVID-19 patients. Pooled analysis of unadjusted data showed that vitamin D use in COVID-19 was significantly associated with reduced ICU admission/mortality (OR 0.41, 95% CI: 0.20, 0.81, p = 0.01, I2 = 66%, random-effects model). Similarly, on pooling adjusted risk estimates, vitamin D was also found to reduce the risk of adverse outcomes (pooled OR 0.27, 95% CI: 0.08, 0.91, p = 0.03, I2 = 80%, random-effects model). Subgroup analysis showed that vitamin D supplementation was associated with improved clinical outcomes only in patients receiving the drug post-COVID-19 diagnosis and not in those who had received vitamin D before diagnosis.
Conclusions
Vitamin D supplementation might be associated with improved clinical outcomes, especially when administered after the diagnosis of COVID-19. However, issues regarding the appropriate dose, duration, and mode of administration of vitamin D remain unanswered and need further research.
https://astralcodexten.substack.com/p/open-thread-177#comment-2233892
John Schilling brought up the question of offsets for murder in a no-politics thread, and trade-offs (if not offsets) for killing are exactly the issue for how aggressive police should be permitted to be.
Do more aggressive police result in less crime? Do the police become criminals themselves if they're cut too much slack? Does restricting police aggressiveness dispirit the police so much that they quit doing useful work? (Sidetrack: A lot of the police seem to have no faith in the justice system if it's being applied to themselves.)
There's a common libertarian refrain that "if it is moral to perform some action for free, then it is moral to perform that action for money." I currently believe this statement, but I want to challenge that belief and see if it survives. I'm willing to bite this bullet when to comes to both prostitution and blackmail. Does anyone have any good counter-examples?
An old post by Scott https://slatestarcodex.com/2013/09/12/the-life-cycle-of-medical-ideas/ mentions minocyclyne, an antibiotic that was found to help schizophrenia patients with treatment-resistent symptoms but even after some research confirming the findings no one seemed to want to prescribe it. Is this still the situation with the drug today? (From the wikipedia article it seems that it is.)
Regarding the delta variant, I decided to plot the Covid-19 infection vs death rate of the UK to see whether the recent increase in cases corresponds to an increase in deaths.
The result is here https://imgur.com/a/aPuNo9q
The recent increase due to the delta variant is shown at the bottom right end of the line. This shows a lower CFR (c. 0.3%) for the latest wave than for previous waves.
Fig 5 here shows an age breakdown of cases: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/996369/Weekly_Flu_and_COVID-19_report_w25.pdf
This wave is being driven by young people, especially 10-29 year olds, in a way that previous ones were not.
Here's another puzzle:
Two friends - let's call them Yvonne and Zachary, for a change, to let Alice and Bob have some time off - have, once again, been captured by a mad jailer.
He's going to show Yvonne a chessboard full of 64 coins (perfectly ordinary coins, no double-headers or anything like that), each showing either heads or tails in some pattern of his choice. He's going to hide the key to freedom under one coin, letting Yvonne see which one it is.
Yvonne gets to turn over exactly one coin, and will then be escorted back to her cell. The jailer will then show Zachary the board. Zachary gets to look under one coin, and if the key isn't there then they're both executed.
If they get to communicate beforehand, what is their best strategy? Does it make a difference if the jailer is listening in on their conversation?
I've done a huge number of podcasts (Future Strategist Podcast) with Greg Cochran, a co-author of the paper Noah Smith links to in his "Genetic superpowers" section. Based on what Greg has told me there is a massive attempt in academia to silence any discussion of the genetics of Jewish intelligence. The attitude Noah takes in his reply is that while he doesn't find the issue interesting, he would have no objection to others looking into it. This is analogous to someone in the Soviet Union in 1950 saying "while I can't think of anything about Stalin to criticize, you should feel free to honestly critique him if you feel so inclined." As Greg wrote in his blog about a co-author of his Jewish intelligence paper "A certain person we will not name, famous for the discovery of the double helix, once said 'Harpending must really have balls of steel, in order to take a genetic look at Jewish intelligence.'” Scott, you once criticized me for saying that you are brave, and because of that I'm not going to comment on your balls.
I've got an idiot thought experiment re. trust to get out there. Don't take it too seriously, just gotta post it so I can forget it.
An omnipotent space bat has come to US, and presents you with a situation:
The population has been split perfectly in half by political alignment. You have been selected as the median representative of your alignment. You have a button. If your press the button, the other wing is instantly disenfranchised without any other harm being done to them.
If you don't press the button, the bat will offer the same choice to the median representative of the other wing.
1. Do you press the button?
2. What if it was a random member of the other wing that gets a shot at the button?
3. What if n random members got a chance at the button? What number n would get you to press?
4. What if the most extreme member of the opposite wing got the button?
Now, the same again, but the button causes some level of harm in addition to disenfranchisement.
What level of harm (you lose 100$ ----> You are a second class citizen ------> death) causes you to the press the button/not press the button?
I remember reading on SSC that all psychotherapy schools are equivalent, and that the result only depends on the therapist's IQ (or something like that, I don't remember well). Can someone point me to a good discussion about this? (Or clear up my confusion?) I said this to a psychology student and he replied that this was totally false.
Are there any inner circle LW / REACH rationalists on the ACX these days? By inner circle I mean someone who's active in the community offline. Well besides Scott himself obviously.
My son is interested in Transcendental Meditation and just got back from a weekend retreat out of state. Is there any organizations I can point him toward that are more legit than others?
I worry about any group discovered over the internet :).
Anyone any suggestion on a good book on history of the worker movement (from industrial revolution up to now; world-wide, not US-specific)? While it's generally easy to find a book on any topic I am having a hard time with this one. Maybe I am using the wrong search words? Or is it really that historians don't care much about the topic?
AI risk question: what's the plausible path from GPT3-esque AI to a AI X-risk doom scenario? I get the classic "paperclip maximizer" argument of why an unfriendly AI can arise unintentionally, but I don't see how it can happen with a design like GPT-3. Specifically, my understanding is that GPT-3, like many of the current breed of machine learning, is trained on a dataset and then later is run on the novel input *without further training*. So the minimization process is done entirely offline at the start with a fixed dataset and the goal is to minimize errors the AI makes at a task like "complete this piece of the dataset given just the start".
How can this ever get to an unfriendly AI? It's not "in a box", the AI is sorta trying to maximize how likely you are to 'accept' it's predicted output and so it could theoretically start outputting things like "Accept this text or you'll never see your children again" and go rogue. But there's no maximization process bringing it to that result - such a threat would never work on the training dataset since it's already fixed and so threats cannot be learned by the AI. Is there some other plausible path for it have unfriendly behavior?
This contrasts to the situation of online reinforcement learning where I can imagine things going out of control. I can't imagine that scenario for this style of pre-trained AI, yet that seems like too simple a solution to the friendly AGI problem. Thoughts?
(And what's the right terminology for this distinction between kinds of AI? Or am I mistaken about there being such a distinction?)
https://www.youtube.com/watch?v=_7VAnAgmshA
5D chess with multiuniverse time travel.
I admit I lost track, but this seems to be legitimate and would probably be of interest here.
I made a "random SSC post" button: https://unirandom.com/slatestarcodex.
It will only open "actual" posts, so no open threads, meetups, or links posts. Developed it originally because it was a hassle to scroll through SSC's archive every time I wanted to read a post — now it works for most other blogs as well, although I haven't found a way to randomize ACX yet.
I read a post on what I believe was somebody's blog a few months back, and it had to do with the oil shock, and I believe it argued it preceded Bretton Woods. Does this sound familiar to anyone? I'm trying to find that post!
Is there a "Cool people street" in Albany, NY? The sort of place where there are young adults, good restaurants, bars, etc.
One data point that sticks out in my mind is the comparison between India and China. Both countries have populations of approximately 1.4 billion yet India has had just under 31 million cases and China only 92,000? Most would surmise that the Chinese data is incorrect, but what if it is not? What could account for such an unheard of discrepancy? After reading your article on lockdown effectiveness, is there any combination of actions that could have resulted in this? It makes me wonder if China already had a vaccine and had administered it.