It was a quiet week on the Covid front. That is a very good thing.

If anything, the theme of this week was looking backwards to how things were handled. The CDC is (at least half-heartedly) looking to fix things. There was an offer of ‘pandemic amnesty’ by someone who thinks they don’t need amnesty. There was more consideration of counterfactual actions on the libertarian roads not taken.

There was, however, a rise in cases and deaths. This raises the chances of a substantial winter wave.

Executive Summary

  1. Cases and deaths rise modestly, raising chance of winter wave.
  2. Forgiveness is often offered where it needs to also be requested.
  3. Things are the largely the good kind of quiet.

Let’s run the numbers.

The Numbers


Predictions from Last Week: 218k cases (+0%) and 2,150 deaths (-5%)

Results: 240k cases (+10%) and 2,618 deaths (+8%).

Predictions for Next Week:

Last week I missed that North Dakota reported -151 deaths. I have now corrected that to zero. If I had caught it I would have then revised this week’s prediction up somewhat, still would have missed low. We are far enough away from the major declines in cases that deaths no longer declining is plausible.

We continue to see deaths and cases move together even when there is no obvious change in reporting that would seem to be causing this.

Despite that, I do think the rise in cases is likely real. Given the timing and how things are going with variants, that means it is likely to continue for a bit. I would interpret one more week of rise as reasonably strong evidence of a future substantial winter wave, which was always a possibility. The regional patterns all match what one would expect to see here. I wouldn’t expect it to get out of hand, certainly not anything like we saw last year.



Booster Boosting

No noteworthy new attempts.


No news, which is by default good news. New variants are still taking over.

Physical World Modeling

A bunch of people are tweeting the same chart with very different ideas about what it means.

The usual Long Covid brigade has one interpretation, because it is their interpretation for everything.

Those who read economic data for a living have a different one.

The second group seems clearly right here, or at least citing the far bigger effect. This is not ‘millions of workers now disabled.’ It is instead ‘millions of disabled now workers.’

A good point that I’d forgotten, the most dangerous part of flying, in terms of infection, is when you are still on the ground because the filtration systems aren’t in full effect. So you might want to mask until you take off, remember that Covid-19 is not the only virus out there.

Masking in the waiting room, no masks in the doctors office that often doubles as a hot box with zero ventilation.

That’s right. Doctors are fine ending the mandate. Admin said no.

Summary of the Boston University Gain of Function incident from a reader, matches my previous views.

Promising signs of moving towards a nasal vaccine. No, nothing urgent, why?

US Senate minority report says Covid likely originated from a research incident (aka a lab leak). Report seems to contain at least one factual error, as we have observed cases of Covid ‘crossing back over’ into humans later in the pandemic. There are details here I did not know that I presume are correct, yet it represents clearly motivated reasoning, and it seems some important details are importantly wrong. I’m not sure if this is stronger or weaker evidence than I would have expected, given that? I do know that ‘they decided to issue a motivated report at all’ is news.

Incidentally, how about this other lab leak of the 1918 flu virus? Perhaps our safety protocols are not so great?

Bob Wachter updates on recent data, mostly little has changed in his model. He does not expect a major surge. The big thing to notice here is the shift where the abundance of caution (not his term here) shifts actions. Thus, at some point you go from ‘the new booster is new so use caution’ to ‘the new booster might be better so use caution and get it’ despite the new data suggesting it isn’t substantially better, because it still might be.

Which is right. If you definitely want a booster and have to choose which booster to get, you 100% would prefer the updated ones. Note that supplies are still limited, as a sign in my local CVS pointed out when I visited the pharmacy.

In Other Covid News

What would have happened if we had tried a Libertarian approach to Covid vaccines? Tyler Cowen seems to affirm in his response as Scott Sumner notes that a true libertarian solution would have been great.

However, Tyler says, we are not legally set up for one in terms of liability without an FDA approval to fall back upon, and there is no way the government would have allowed the vaccine to remain expensive, so the rewards had to come up front.

I see this as a strange choice of counterfactual. As in, it’s super weird. You let Pfizer sell its vaccine without FDA approval, while also letting people sue Pfizer over it and not let them waive that liability. One must presume that the libertarian solution would have to include the ability to waive liability, and then people would get to choose either wait until you have stronger evidence of safety or get the shot now accepting the risks.

In terms of what to charge for the vaccine, that’s odd too. The whole idea is that the government won’t allow X, so it’s strange to object to ‘what if we did allow X?’ by saying the government wouldn’t allow X. What would be a reasonable version of this? Perhaps let us say that Pfizer and Moderna can charge market-clearing prices as they ramp up production, which people are happy to pay, then once there is enough supply they have to charge reasonable prices to insurance companies (and for those without insurance, the government) the same as everyone else. That seems like a least implausible version of the implausible hypothetical? And it seems great. I would have loved to see ‘the early doses get de facto auctioned off, then the rest get bought the standard way.’


Emily Oster proposed what she calls an ‘amnesty’ and says she ‘forgives’ everyone for what they did during Covid ‘while they were in the dark.’

Some people thought this was a good idea.

Most responded about how you would expect.

In some situations, it is generous to offer people forgiveness. In others, it is a claim that you are the one who gets to do the forgiving.

There are similar responses from the other direction, talking about Oster and others trying to coerce them into vaccinations. Or simply things like this.

There is typically little appetite for ‘forgiveness’ when everyone still thinks they were right. There is nothing stopping Emily Oster or others from unilaterally ‘forgiving’ whatever they would like. It seems good, even, if done genuinely. If done as a way of saying ‘I was right,’ even if she was indeed right, then not so much.

Here’s quite the thread.

Petite Charette has some very strong opinions about what it would have meant to ‘be driven by science and data’ and what it would not have meant, not all of which I agree with. It does illustrate that this was not a ‘we did not know the facts’ set of issues, it was a ‘we chose based on what we knew and people did not like how that choice was made’ situation.

People are still rather raw about it.

The best way to earn forgiveness is always, of course, to not only admit error but also then to Fix It. CDC director Walensky is attempting to do so, with explicit goals including ‘tell the public what we don’t know’ and ‘inform the public faster of what we do know’ and ‘in a pandemic you don’t have time to wait’ which is a nice change since this time around you could have fooled us. Internal pushback is proving a problem – the CDC’s culture says that you absolutely do have time to wait no matter what, which is not how I would recommend doing Disease Control yet here we are.

Meanwhile, all of this is about information being given relatively quickly. None of this is about recommendations or permissions or actions happening quickly. At some point, things stop being prospects for reform and the better option becomes replacement. This was a communications issue but was not only a communications issue. If the CDC thinks it doesn’t have other huge problems to fix, well, CDC Delenda Est.


China doubles down on its lockdowns in the wake of the party congress, impacting hundreds of millions at a time. Those who went to Chinese Disneyland for Halloween got locked inside, which is actually a pretty good scare in that case.

The cost of continuously having to check-in with your status, of having to worry about potential lockdowns happening anywhere at any time for indefinite periods, has to be taking a cumulative toll. There is no obvious next off ramp for a long time. There is also a lot of ruin in a nation and the impact will be gradual. It could add up to quite a bit before things are forced to change.

Other Medical and Research News

Pfizer reports Phase III success in its maternal RSV vaccine trial, reducing severe illness in newborns by 81.8%. People are saying this might be our last really bad RSV season. Which would be great, and a large improvement over this past week for the whole family.

From Saloni, I missed it last week, a Wired piece on how to speed up science based on lessons from the pandemic. Everything here seems great. We should do all of it. The problem is that there are other things, that matter more, that Saloni is not even willing to write in Wired. Until we acknowledge the regulatory obstacles in our way, we will continue to move slowly.

Saloni also asks, this time in The Guardian, should we do challenge trials? Yes.

Study on what would happen with FDA deregulation of medical devices. Innovation +400%, Firm Entry +1000%, Prices -40%, even some safety improvements.


New Comment
3 comments, sorted by Click to highlight new comments since: Today at 1:51 PM

Incidentally, how about this other lab leak of the 1918 flu virus? Perhaps our safety protocols are not so great?

I think far more relevant is that SARS 1 leaked from a lab four times.

There are four documented leaks and it leaked four times is not the same thing. There's no reason to assume a 100% reporting rate. It likely leaked more than four times. 

Booster Boosting

No noteworthy new attempts.

*According to publicly available information.

New to LessWrong?