A friend recently wrote that they wouldn't be attending a mask-optional contra dance weekend I'm playing at because it doesn't require vaccination. As an organizer of unrelated dance events, which also don't require vaccination, here's how I see the situation.

When the covid vaccines first came out they were a huge improvement over the status quo. Getting the vaccine reduced your risk of severe illness or death, reduced your chance of catching it, and reduced your chance of giving it to others after you were sick. Our events initially required vaccination, which I think was the right call.

At this point, however, there are a few different things you might mean if you say your event requires vaccination. Usually it's either:

  • A complete primary series counts, even if it's one shot of J&J from early 2021.

  • The most recent ("updated") booster is required.

The CDC used to call the first category "fully vaccinated", but no longer talks about it prominently. They've switched to focusing on the second one, which they call "up to date".

This change makes sense: from a perspective of avoiding getting infected and passing it on, only a recent booster does very much. A few months ago I wrote about the results in Menegale et. al (2023) where they saw that as an infection control measure vaccines wane quickly: effectiveness halves about three times a year.

Additionally, lots of people got sick this winter, which acts similarly to an a vaccination. Given how quickly the vaccine wears off, I'd be less concerned about risk from someone unvaccinated who'd had covid over Christmas than someone who got their booster in the early Fall and dodged the winter wave.

What does this mean for a dance event?

  • If you want a vaccination requirement to be doing anything useful, you need to require people be up-to-date with their vaccine. Risk from someone last boosted in Fall 2022 or before is not appreciably different from someone who was never vaccinated.

  • Requiring up-to-date vaccination rules out 80% of people 18+, though if they're excited enough about your event they could go get the shot.

  • Unless your event is in the fall or early winter, even requiring up-to-date vaccination doesn't help much: if someone got their "updated" booster in the fall they're down to ~35% of peak efficacy by the time of an April event.

If an event does want to reduce risk of transmission, I'd recommend considering some combination of ventilation, filters, tests, and masking.

Comment via: facebook, lesswrong, mastodon

New to LessWrong?

New Comment
21 comments, sorted by Click to highlight new comments since: Today at 9:30 AM

I have anecdotally heard of about 8 dances with roughly half not requiring masking and half requiring it, where it seemed like it would be possible to detect covid infections (scrupulous people attend, and test before and after, and either many get sick or none get sick).

My currently cached prior (a posterior over this data that predicts/summarizes all the events) is that if everyone is masking then no one gets sick, and if masking is optional there WILL be a wave of people reporting that they caught covid at the event. Vaccination requirements don't appear to be correlated at all.

We made a similar decision at my Zen center, although for different reasons.

When we re-opened to in-person sitting in 2022, we required both masking at all times indoors and that everyone be fully vaccinated.

Early in 2023, we dropped both requirements. We looked, and estimated that >90% of people likely to attend the Zen center would be vaccinated or have naturally immunity from prior infection. Additionally, mask requirements were rapidly disappearing across the Bay Area, and keeping one felt silly since we were all going out to places without masks on.

Here in early 2024, things are basically back to normal after a year of gradually ramping reduced mask usage. We don't ask or really think about vaccination status. We have an air purifier in the zendo, and people do occasionally wear masks if they want, and we added some text to our webpage encouraging people to stay home and join us on Zoom if they are sick or have recently been exposed to someone who was sick.

Given attendance is back up at or above pre-pandemic levels, I feel comfortable saying that the lack of restrictions is not in general deterring folks, and as far as I know we've not heard from anyone that they would attend in person if we had a mask or vaccine requirement.

I suspect there is some heavy overlap between people who attend the zen center and people who would attend a contra dance, although I'm guessing contra dance attendees skew younger than zen center attendees. If anything you'd expect old people to be more worried about COVID, so I consider this weak evidence that you're making the right choice.

I'm guessing contra dance attendees skew younger than zen center attendees

Nationally I suspect this isn't the case; I'd guess 80% of contra dancers are baby boomers. But some communities have much younger demographics.

Oh wow! Everyone I know who is into contra dancing in the Bay Area seems to be late 20s to early 30s. I don't go to contra dances myself, though, so seems I may have incorrectly extrapolated about the demographics.

The LW/EA overlap with contra is pretty much only in the younger crowd, driven mostly (I think) by LW/EA being younger overall.

If anything you'd expect old people to be more worried about COVID, so I consider this weak evidence that you're making the right choice.

That assumes that the motivating factor is more about health concerns than tribal signaling for those who care a lot about the vaccination status of other people. 

I expect that to be false, and a lot of people who are strongly in favor of keeping out unvaccinated people like the fact that it keeps out certain people with whom they politically disagree. If they could make a rule that disallows Trump voters from attending they would likely also be in favor of such a rule.

Sure, these are all relevant factors to why some people may want a vaccine requirement, but Jeff seems to mostly be concerned about making the choice that is best for the health of attendees and not about signaling tribal affiliation, and so it's in that context that I suggest this fact is weak evidence in support of his choice.

I'm looking forward to when we can produce cheap aerosolized vaccines and just routinely spray all public spaces. You aren't consenting to viruses being in the air, why should you need to consent to vaccines being in the air?

That seems pretty unlikely to be popular.

It also requires producing far more vaccine than if you just spray it in people's noses, and gives up control of the dosage and timing.

Yeah. Administering it in some way that is relatively controlled, such as a puff of mist in the face when boarding an international flight seems doable though. But yeah, a variety of downsides w current tech. Ten years from now? Hard to say what the world will look like.

Hopefully you're joking. Given the common (average?) reaction of two days of feeling flu-like and feverish after Covid vaccination, not to mention people who predictably react worse than average, I can think of a lot of times I'd rather schedule that sadly worthwhile experience than the first two days of an international trip.

Yes, the ideal tech would be configured to not have a noticeable effect on those already current on all recommended vaccines. So it would only be noticable at all if your neglected your vaccinations.

Is that sort of configuration even biologically possible (or realistic)? I have no deep immunology understanding, but I think bad reactions to vaccines have little to nothing to do with whether you're up-to-date on previous vaccines. So far, I'm not sure we're good at predicting who reacts with more severe side effects than average (and if we did, it's not like it's easy to tweak the vaccine, except for tradeoff-y things like lowering the vaccination dose). 

Honestly, I don't know. I'm just trying to be open minded about the weird stuff that might become possible due to rapid advances in AI speeding up tech progress generally. Areas like biotech, where the complexity of the problem and our limited understanding are the bottlenecks will be more likely to progress rapidly than, say, energy or mass constrained areas of tech.


When the covid vaccines first came out they were a huge improvement over the status quo. Getting the vaccine reduced your risk of severe illness or death, reduced your chance of catching it, and reduced your chance of giving it to others after you were sick. 

What makes you confident that the last of those three claims is true? It seems like the first two were backed by the clinical trials but the third not really?

reduced your chance of giving it to others after you were sick.

What makes you confident?

I've seen a lot of reporting of this over the years, which I think is going back to seeing lower viral loads for vaccinated people? Here's an example study, which I've only looked at the abstract of: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8982774/

I got the first three jabs out of a selfish desire to protect my health from a filthy world with no adequate public health systems.

I'm thinking of getting a fourth now that I've heard that the new vaccines are finally putting new strains (but not the old strains) into the new formulation...

...but my suspicion is that all of these are anti-social, because by widely deploying imperfect vaccines (that attenuate the symptoms of an enormously infectious airborne illness (that is already known to spread a lot at the very beginning of the illness at low doses and with low symptoms)) we're doing something that is probably (should probably be considered?) low-key "immoral". Because it probably encourages the evolution of a disease that hurts people with fewer medical resources.

Abstractly, it feels like defecting on an N-person prisoner's dilemma because it seems to be leading to a world where everyone has to get (imperfect?) vaccines in order to be safe from viruses that are harmful because everyone is getting imperfect vaccines.

Reading the paper you linked, the idea is that vaccination that doesn't result in eradication contributes to the evolution of the pathogen. They propose targeting vaccines toward virulence antigens (markers highly correlated to the pathogen being unpleasant) to influence the pathogen to evolve to lower virulence.

Their successful examples are all bacterial, which makes sense: it's much easier to separate the functions of a bacterial pathogen. Their viral example is HPV, where they discuss the idea of targeting vaccines to make it less likely to cause cancer. But, as they say, for most viruses we don't know what the virulence antigens are.

I may have missed this in the paper, but I don't see a reason to expect that randomly selected antigens would push virulence in one direction or another, so I don't know why you would expect the current covid vaccines to increase virulence?

Note that all of the above is just based on reading this one 2001 paper; I don't know if the field has made progress since then?

(But any costs or benefits here are quite low: in the US under 20% of people have the recent ("updated") booster, and globally it's surely lower. So the evolution of covid is overwhelmingly driven by escaping natural immunity from previous infections.)

This is an idea that feels "really really important if true" but that I'm not actually certain about and often bounce off of. Pushing on it a little more, this paper on Marek's Disease from 2015 sketches a theory of "hotness".

Hotness is a hypothetical "conflation of transmissibility and pathogenicity" that might sometimes occur as a spandrel at first, which then is found to be useful by some evolutionary systems, which optimize the spandrel "on purpose".

You could imagine a disease which has one "hotness level" with no vaccines at all (H0?), and a different "hotness level" (H1) in patients with an imperfect vaccine.

With no background knowledge at all H0 > H1 could be true on average regarding viruses (and that is consistent with the idea that vaccines are DESIGNED to help the patient by reducing pathogenicity from a patient-centric perspective).

However, we expect some amount of "hotness" might contribute (from a virus-centric perspective) to "transmissibility" as well... if your nose became so runny you die of dehydration before transmitting that would be "too hot" from a virus centric perspective, but if your nose is not runny at all in any way then maybe the virus isn't causing the host to shed as many viral particles as would maximize the total number of downstream infections.

The thing I'd suggest is that maybe "we as a collective herd" are LUCKY when only 20% of the population is defecting on the strategy that would tame any given virus?

Here's a hypothetical bad path, that probably only kicks in if almost everyone takes these imperfect vaccines, sketched as a possible future:

On step ZERO he first imperfect vaccine is deployed against a naive pathogen, with 60% uptake. H1_0 is kinder to the patient at first (and a reason to buy and take the vaccine, selfishly, for each patient) but H0_0 is tolerable and not (yet) a strong downside reason to take the vaccine to avoid the symptoms...

But then on step ONE the disease, which already had an optimized hotness level (and since 60% are imperfectly vaccinated that's the central case to optimize for), performs some evolutionary cycles so that H1_1 goes up to a higher (closer to optimal) level of hotness... a higher level of BOTH pathogenicity AND transmissibility. What happens to H0_1 is harder to say. It happens more "by accident" than "due to viral evolution". 

On step TWO, humans react by deploying a new imperfect vaccine to lower (pathogenic) hotness in newly vaccinated humans to H1_2. Just as before.

On step THREE the virus reacts by evolving to put H1_3 back up, to the level of hotness it prefers, with uncertain effects on H0_3, but in the battle between humans of viruses it seems like maybe a red queen race between science and evolution, and there's one thing NOT racing here: the naive immune system of naive humans.

On all subsequent even steps "science learns", and lowers "H1" (leaving H0 unconsidered) and if this leads to H0 becoming a large burden that might easily cause more humans (reacting to avoid serious pain) that is actually a nice thing from the perspective of the profit-seeking scientists: their market penetration is getting bigger!

On all subsequent odd steps "the virus learns" and raises "H1" again (not worrying too much about keeping H0 also close to the ideal hotness if the unvaccinated are very very rare, and so in general this could end up almost anywhere because it isn't being optimized by anyone or anything)?

(((Note that this model might be a BAD prediction of the future. It might be mechanistically false! The reason to think it might be false is a sort of "tails come apart" or "goodhart's law" issue, where, if we think that "hotness" is the only thing that exists (which subsumes both pathogencity and transmissibility) so that scientists vs evolution cause this one variable to go up and down over and over... but if the virus and the scientists could ask more specifiucally for exactly what they want then the virus could get very very high transmissibility and the scientists could get very very low pathogencity and they'd both be naively happy. However... this ignores the third party... the patients who the for-profit medical scientists are trying to extract payments from.)))

So overall, over time perhaps we see:

The virus becomes unconcerned if the 0.5% of patients  who lack an imperfect vaccine die from H0 being very very hot, and the for-profit private medical scientists become happy if H0 gets very very hot and kills anyone who doesn't buy their product. And the biology suggest that this might be a stable bioeconomic Red Queen Race... depending on how H0 fluctuates in (a loosely correlated?) response to the dynamic tensions to iteratively raise and lower H1.

A pattern similar to this sometimes "arises for some period of time" within normal evolution (without any intervention by scientists). For example, bay trees have unimportant symptoms when infected with sudden oak death, whereas oak trees are killed by the pathogen.

Bay trees thus have an evolutionary incentive to maintain their infections, which clear the area around them of competing trees, giving them access to the sunlight. Oak trees have an incentive to react to this combined attack, but if they don't learn to ALSO resist the sudden oak death pathogen very quickly they might simply be removed from the game board.

In this analogy, those who take imperfect vaccines would be like the bay trees, and the transition from "mostly oak forests" to "mostly bay forests" would be like what the vaccine-making for-profit scientists would want to cause, to maximize vaccine purchasing among the aggregate "herds of customer/victims" when they sell their products to individuals rather than selling to coordinated centralized (elected?) herd managers.

Something in my soul resonates with the idea of "doing what a benevolent herd manager would tell me to do" if any benevolent herd managers existed.

Since no benevolent-and-competent herd managers exist in the modern world, this is perhaps a silly yearning for me to have, yet I still think about it anyway, because I am a fool.

Separately, I'm not actually sure of the science here. Maybe "hotness" isn't a useful way to think about the relationship between pathogenicity and transmissibility and/or maybe H0 stays reasonably low no matter what, even when there's almost no optimization pressure on it?

I think what confuses me the most about your model is:

a. I'm pretty sure the immunity you get from having had an iteration of the virus and from having had the imperfect vaccine are similar.

b. Then the selection pressures of vaccination vs infection are quite similar.

(Separately, if it were actually the case that annual imperfect vaccination made things worse for the people not getting the vaccine, which in the US is a large majority, then it seems like the CDC should not be recommending them. What do they, and other public health authorities, think of these evolutionary arguments?)

That is all quite reasonable!

I. Regarding the CDC

I tried to write about the CDC taking hyperpathogenic evolution due to imperfect vaccines seriously at an object level (where the CDC was the object level thing being looked at).

It kept veering into, selectorate theory, first past the post voting, Solzhenitsyn, and so on. Best not to talk much about that when the OP is about dancing and voluntary association :-)

Treating imperfect diseases as the object level, and "going doubly meta", I'd point out that (1) argument screens off authority, and also (2) the best way for a group of umpires to get the right answer most reliably is for all of them to look ONLY at the object level: collecting the maximally feasible de-correlated observations using all the available eyes and then use good aggregation procedures to reach Bayesian Agreement over the totality of the observations.

Ideal umpires only give correlated answers through the intermediary of describing the same thing in the world (the actual ball landing in some actual place, and so on). This is why each additional umpire's voice means something extra, on an epistemic (rather than military/strategic) level.

If you want to talk politics, we can, but I think I'd rather talk "umpire to umpire", about "the thing in front of us".

(And also separately, if we get into politics, I don't think the CDC is anything like an ideal umpire, hence why I'd prefer to treat "politics" as a semantic stopsign for now. Why does the CDC say what it says? Politics. Does this answer help predict anything else about the CDC? Mostly not. Does it help keep other arguments clean and safe? Hopefully yes.)

II. Regarding Imperfect Vaccines And Imperfect Immune Operation

I think your "A" and "B" are roughly right, and a sign that I've communicated effectively and you've understood what I'm saying :-)

I think imperfect "endogenous immune responses" in one population would/should/could breed diseases that are unusually pathogenic in other populations.

The moral/deontic universalization argument against imperfect "exogenous immune responses" is just (1) it probably works the same way because biology is biology and evolution is evolution... and (2) we actually have a choice here because we can DO() a vaccine in a way that we cannot easily DO() an innate immune response programmed by our genome to happen in our bodies.

I think the logic I'm talking about is similar to the logic that explains why diseases tend to be especially virulent right after jumping from one species to the next.

It also might partly explain why a handful of endemic East Hemisphere diseases were so harmful to West Hemisphere populations during the genocides from ~1492 to ~1880.

A "maybe exceptional thing" here is that the natural immune system actually sometimes gives quite broad protection (equivalent to a perfect vaccine), as when a mild cowpox infection protects against cowpox and smallpox basically for life.

So "broad, perfect, endogenous, immune responses" exist.

If we had "broad, perfect, exogenous, immune responses", many existing pathogens might be eradicated!

It would push more pathogens into "counterfactual worlds" where they can be imagined, as what "would have happened if the infectious disease defense had not been adequate"... but they wouldn't be directly empirically observable. People would see this medical system, and they would see no diseases, and they might be confused.

There's already a bunch of diseases we don't have... like supermeasles and hyperrabies and sneeze-AIDS-herpes (which covid is kinda close to, but not as bad as, so far as I can tell), and so on... that we could hypothetically have if someone created them in a lab on purpose.

These are hard to count as "bayesian evidence" of "diseases that are only counterfactual and have, in some sense, been kept out of material reality due to no one performing the sequence of actions that would create and/or spread and/or not eradicate them".

Compared to all the hypothetically possible diseases, we've "successfully avoided" most of them! <3

If we "ban Gain-of-Function Outside BSL5s" then we could probably avoid nearly all of them forever.

We have a handful of cases of diseases at the edge of counterfactuality, like smallpox and polio and measles, which were diseases that basically didn't happen in the US back before US institutions fell into serious decline.

So those used to be "diseases that we could more easily 'count' because we used to be able to see them". Very long ago (before the germ theory of disease) they were quite common and very tragic, so we know they can exist. Then science and adequate medicine caused them to not ambiently exist to be counted. So their "absence now" is glaring when they are absent (and their return is (for measles) or would be (for worse ones) even more glaring).

In terms of why the immune system might sometimes internally do imperfect immune response already: it might just be that when it happens the species it happens to evolves to extinction, and this might be a way to use Gain-of-Function to kill all humans, if someone (like a hostile AI) wanted to do that. The modeling is very tricky. There are some known evolutionary systems (like hyperparasites) that can probably grow to a certain point and then catastrophically collapse to total extinction if there is a single well-mixed evolutionary compartment.

Also, arguably, it is "genocidally/evolutionarily correct strategy" to maintain a few "pet" diseases within your stable of "imperfect immune response diseases"? (Like earlier I mentioned "sudden oak death" being harborded by bay trees.)

With a "pet hyperpathogen" when you meet other similar animals after periods of long separation you have a decent chance to kill them without even really trying (as with the Europeans in North America), and so maybe this is a "good evolutionary strategy" even if it is wildly immoral. I don't think anyone who was all three of (1) sane, (2) reasonable, and (2) emotionally intact has ever claimed that evolution is stepwise continuously moral. It is at best "long run moral" and maybe not even that.

If my fears about the evolution of worse pathogens due to systematic exposure to imperfect vaccines is valid...

...then I guess "distant people (like future generations and people in other countries)" are just lucky right now that such a small percentage of current Americans are taking the new imperfect covid vaccines.

If my fears are right, then if we took imperfect vaccines very reliably across nearly the whole population, that might hurt distant people by making them either have to take the vaccine as well, or else suffering greatly.

But contrariwise, if my fears about the evolution of more pathogenic strains due to imperfect vaccines are not how things actually would or do or are working (which could be scientifically true as far as I know) then the low level of "personally effective even if imperfect" vaccine uptake is a minor tragedy. We're leaving health on the table for no reason, if that's the world we live in.

All my arguments here boil down to "if it hurts we shouldn't do it, but if it helps then we should do it, and I'm not sure which situation we're actually in, but almost no one is even looking at it very hard".

Knowing which thing is actually true, and convincing lots of people to believe the actual truth, has high aggregate Value of Information (VoI).

Millions of lives and lots of ill health are at stake considering the breadth and depth of time and space.

Answering this question properly is the sort of thing that a competent benevolent philosopher with a decent budget for important empirical efforts "would be interested in being able to do".

The ethics of it would be a little weird. The highest quality evidence would probably involve doing "random assignment challenge trials" on whole human societies, where isolated societies that want to ban imperfect vaccines "just in case" are randomly forced to use them anyway, to satisfy a scientific curiosity about whether that random assignment reliably makes their ambient diseases more harmful to people who haven't taken the imperfect vaccine yet.

With Marek's Disease we can just do this for chickens, since chicken death and illness isn't nearly as morally important as human death and illness. Like: we already torture chickens to death for the sake of Chicken McNuggets, and scientific truth about important questions is much more important than Chicken McNuggets, so I tentatively think it would be ethically OK to do that kind of research in the current wildly-non-utopian situation?

But my understanding is that we've already done that research, and it says "yeah, imperfect vaccines promote the evolution of diseases that are more virulent in the non-vaccinated, in chickens, with this one disease".

Maybe we should kill a lot more chickens with another disease?

Or kill a lot of ferrets with another disease? Or something?

To "prove it more broadly, and more generally, with slightly more data"?

Except I think that most humans simply don't have the patience to think about this stuff, and they won't understand or care about "why one particular vaccine might be net good but some other particular vaccine might be net bad based on <complex evidence and arguments>".

My current working model is that it is just "reasonably inferrable to anyone with the patience and interest in looking at the data and thinking properly" that taking an imperfect covid vaccine is not something a good Kantian would do, because universalizing the behavior among all people able to follow moral maxims (which includes all humans, right?) would be a net negative overall...

But also my current working model says that almost no one cares or wants to think about it very much, especially since the existing levels of imperfect vaccine uptake are already pretty low (quite a bit less than 50%), and therefore less likely to cause the evolutionary effects at the sociologically observed levels of default behavior.

So maybe we can use imperfect vaccines to protect the 5% of people who are most vulnerable, and just watch out for pathogenicity levels in the non-vaccinated, and then ban the imperfect vaccine based on live data? Or something?

Performing medical self-experiments is kind of heroic <3