Posts
Comments
I am going to 24 hour grocery stores in the middle of the night once every 2 weeks to reup to 6 weeks worth of food. I imagine this sort of thing is harder for people in denser areas.
Those are direct effects, but many other people won't receive care or will have care postponed by general capacity overwhelm.
I'm having trouble imagining a concrete scenario where it would be possible to use this information to gain an advantage.
This isn't a random sample afaik. That said, 2x worse than a very bad flu year falls well within the current projections (flu: 8% of population, up to 60k dead, COVID maybe 15-20% infected 100-200k dead). Could also do 3x that though pretty easily.
I'd like to see a model for knock-on deaths from hospital overwhelm.
I don't know how this would work. Trading on info always seems great until it comes time to actually do it. eg you get actual inside knowledge that suggests a trade, but wait, how do you know whether this inside knowledge is sufficiently leaked that it is priced in?
meta but why argue about this if there is high confidence that serological data will determine things one way or another over the next week or two? Is there decision relevant leverage?
Thanks for sharing this. I wish more people would share their reasoning about how to model rather than just their finished models.
I call it orthogonal sum games.
I rag on the California government a lot, but big positive update on their early response. Some will want to credit the big tech campuses for closing fast, but this is more a function of how cheap it was for them to do so relative to other businesses.
Tangent but self-locating uncertainty is an interesting angle on human ethics as well.
the zoom link to ask your question in the call seems like it will increase the amount of time per question by a lot (and thus decrease total questions) unless you have someone buffering the question line well.
https://en.wikipedia.org/wiki/Dog_behavior I'd guess textbooks would be the best place to find research.
That prediction is qualitative. Widely (unknown), and unknown time parameter, unknown how much tiny is, unclear which companies would win.
Assumption: we shouldn't expect to be able to make strong quantitative predictions unless we also expect to be able to get rich playing the markets.
That leaves qualitative predictions about things that might change. What sort of predictions of this sort were made during other crises? How did they pan out? Who has a surprisingly good track record? Can we solicit from them? What surprising actions are they taking at present? What can be inferred from that?
I think that status anxiety is more unpleasant than is intuitive.
Spoilers for people who don't want dogs ruined for them.
The author is correct that dogs are an artificial status prosthetic for humans. People misinterpret dog body language. Many of the things that are colloquially interpreted as happiness are actually status anxiety. The way we interact with dogs, especially with regards to food, makes them believe they at the bottom of the pack hierarchy and therefore potentially at risk of losing their place entirely. This likely is exacerbated by the fact that they don't contribute to the pack in any way that they are wired for. It is possible to relieve some of these anxieties but obviously most wouldn't care to.
I feel like there's wet spaghetti code at both ends of this? On one side we have economic impact projections that vary about an order of magnitude and the other we have expected death projections that vary over 1-2 orders of magnitude.
easy way is (per litre of water)
1 tsp nusalt or no-salt (potassium chloride)
1/2 tsp salt
1/8 tsp epsom salt (food grade) (it's okay to eyeball small amounts of this, high therapeutic index)
6 tsp sugar
For the first, don't think in terms of the US and its suicidal litigiousness. Think Iran, think rural hospitals, think what people will do if someone is dying at home and no hospital will take them.
He's still basically claiming over reaction but at least this is something major.
Two things that are paralyzing enormous numbers of potential helpers:
fear of not having permission, liability, etc
fear of duplicating effort from not knowing who is working on what
in a fast moving crisis, sufficient confidence about either is always lagging the frontline.
First you have to solve this problem for yourself in order to get enough confidence to act. Something neglected might be to focus on solving it for others rather than just working on object level medical stuff (bottlenecks etc.)
Good notice. I don't and I'd put my money on no.
Regarding days until hospital overwhelm. Literally every single model I've seen so far uses zero hospital elasticity which I've seen healthcare workers say is unrealistic. granted, with an exponential, a doubling of hospital beds only buys you a week. But that's the uncontrolled scenario.
Regarding cytokine storms. In the absence of hospital treatment, what are best practices with OTC available options? NAC? How much etc. Who would best be able to answer this?
Prospective study of probiotic supplementation results in immune stimulation and improvement of upper respiratory infection rate
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5995450/#!po=68.7500
fewer incidences of upper respiratory illness (19% vs 46%), fewer incidences of flu with symptoms of upper respiratory illness i.e cough (4.5% vs 16.4%), had less severe symptoms and were sick for fewer days than the placebo group.
Of specific interest:
Estimation of the asymptomatic ratio of novel coronavirus infections (COVID-19)
https://www.ijidonline.com/article/S1201-9712(20)30139-9/pdf
the asymptomatic ratio is thus estimated at30.8% (95% confidence interval (CI): 7.7%, 53.8%)
If this is true then we should see another spike in Hubei as they lift restrictions right?
Oh, yeah don't know how the differences would play out down stream in the model or in real life. Just thought it was worth pointing out since 3-5x vs 20x+ is very large and thus it's worth investigating the model to see what's causing that difference.
Joscha's article seems to overcorrect in the other direction. Messing with ad hoc models I get a mitigated peak 3-5x hospital capacity, not 20x+.
You're building a causal model to do that. A babbler only knows that whatever they said to you failed to make you smile and nod, so they will say what they said to you less in the future. Or less to people like you while seeking those who respond better.
One of the consistent patterns with such people is that they have to spread the message really wide. They love awareness raising type things. If they have enough audience they can find at least some agreeable people who smile and nod.
ad hoc masks with hepa filter paper
Made the hacker news homepage, hospital playbook. I imagine useful for those trying to flesh out good protocols for what to do if a housemate or family member takes ill.
The requirements on multiplying ventilator use through sharing is
1. Equal tube lengths.
2. Equal lung capacity.
3. Equal lung resistance.
4. Same patient weight (approx)
the question is can any of these requirements be broken though clever use of 3d printed valves or other JIT solutions?
related:
http://www.overcomingbias.com/2017/03/better-babblers.html
https://www.lesswrong.com/posts/xqAnKW46FqzPLnGmH/causal-reality-vs-social-reality
The short answer is that you're most likely not going to drag a social reality person into causal reality.
Subway use in New York is only down 20%. And mask use is rare. How is NY not completely screwed?
>Maybe the U.S. and European markets had reflected a much safer world than anyone at the time of prior disasters had expected?
that's part of what super low rates/yield imply right? That people expect a very stable world. Higher multiples are flimsier in the face of new evidence.
And to start compiling a list of common problems with the parameter estimates being used. Eg I am seeing some models extrapolate naively from the fact that most cases are coming from the least controlled places with the widest uncertainty bars.
I don't have a strong stance either way but I think there are some interesting points for the other side. Let's say this is on the order of a ww2 number of deaths. Look at the impact of ww2 on the economy. Except with the present situation there's no need to halt international trading, and none of the productive assets get bombed. This still leaves a lot of room on the side of it being worse than many of these sunny projections, but what I want to point at is that it is also really common to drastically underestimate how bad things can look locally and still have trends mostly do alright due to the enormous differences of scale between what it takes for things to look bad and what it takes for things to be bad everywhere.
I can't remember the term it's antibody specific amplification or something, but some promising vaccines will actually make you more susceptible.
I think we will likely get some acceleration out of heroic volunteers, if not in the US then elsewhere.
I would strongly encourage people to try brainstorming some questions. Even if you don't come up with anything directly useful you might jog someone else's creativity. Remember to go for quantity over quality on your first pass.
Epidemiology questions that, while we probably can't do much about, would be useful to try to ad hoc model given how bad official info has been so far:
Are estimates of doubling time off from bad modeling of rapid test ramping making it seem faster than it is?
What is actual hospital elasticity? Is there an existing gathering of data on this from previous disasters?
How long do human trials need to be before they are rolled out to the majority of the population? Just to the extremely vulnerable? What is the gears level model here?
What granularity of travel restriction makes the most sense? In general, how can cities and counties act knowing that federal response may (will continue to be) be too slow?
Which physical objects have longer supply chains and thus can be expected to be less robust to disruption?
What mental health problems can we expect to spike hard in the next 1-6 months given people feeling shut in and helpless?
What are the most predictable second order disasters?
Does moral hazard show up anywhere here?
What's most likely to be ignored during this? Civil liberties? Already seen discussion of that. What's even more ignored?
I've seen people from a Stanford lab asking on facebook about being put in touch with someone from an MIT lab. How can lab cross talk increase?
If UV 210nm turns out to be effective, how can you build your own flashlight/lightsaber (from the virus' perspective) out of off the shelf parts?
Which continuing failures of the FDA are highly predictable? What can be done to mitigate that expectation at the hospital and lab level?
How can models take into account reference classes. e.g. Many models are averaging naively which means essentially all the data points are from the least controlled regions with the widest error bars.
" On an exponential curve, exactly correct reactions are not a thing. "
This is fantastic
Mask supply via alibaba/aliexpress seems to have ramped. I predict no shortage within a few weeks.
If you develop a dry cough, do take cough drops and sip liquids to prevent yourself from coughing.
Reasoning: I've seen one doctor claim this is helpful because one of the causes of severity is how deep into the respiratory system the virus is able to travel. Apparently lower respiratory infections are more severe because the immune system has a harder time fighting it. Coughing tends to drive the virus deeper into the respiratory system.
This would be inside a sealed container that you then open outside.
People operate mostly via mimesis with a little bit of reason sprinkled on top. So it has been mostly about modifying the environment so that they will be more exposed to high status information sources saying the useful things rather than trying to pelt them with facts myself. e.g. if you show them a dashboard, people will often go back and check it themselves. If you invite them to a social media group posting updates they will find their own info and convince themselves.
Introduce changes in routines slowly over the span of a few days. Arrange cleaning supplies so that they are available but not in your face initially. Then a couple days later add to them or make them more prominent. This is much like animal behavior training. Elephant and rider. There's a famous quote that says something to the effect: 'the job of intellectuals isn't to change things now. It's to be ready with plans so that when the crisis comes and people are scrambling about for options you already have something neat and tidy to drop in their laps.' If you are gentle with the facts up front, when they do get convinced by the turning of the social tides they are much more likely to come to you as you've seeded the idea that you've already been tracking and following things. At that point if you've already bought the supplies and printed the checklists you make a nice pain free glidepath for them.
If you want to investigate the intuitions themselves e.g. what is generating differing intuitions between researchers, I'd pay attention to which metaphors are being used in the reference class tennis when reading the existing debates.
I like the idea of prepping as a service. You pay us to stay up to date on risk thresholds + research the ideal kit of necessary supplies for likely types of disasters. When the risk threshold triggers we deliver a kit to your house along with instructions. The risk threshold is low enough that this happens 'too early' by most people's standards, so you wind up crying wolf. But that's okay, it just means you accrue your own extra stock of some supplies.
It's possible that the easier solution is just put everything in a box with an ozone generator. Faster too. Evidence of efficacy needs reviewing.
Casual/co-op games that aren't super high mental overhead combined with voice chat can be more connective/sustainable than the slightly more formal video calls. I can only stay on a video call for about an hour to an hour and a half, but I can jump in to and out of group games regularly.