Happy Scribe Logo


Proofread by 0 readers

covid-19 is the formal name of this disease, it was called the Novel Coronavirus and and Kov 19 novel coronavirus 20 19 denoted both the virus and the disease. And now they have sort of factored those out into sars-cov-2, which is the virus and covid-19 the disease.


Hello and welcome. I'm Shane Parrish, and you're listening to a special episode of an odd project, a podcast dedicated to mastering the best of what other people have already figured out. This episode is special because we're doing a deep dive on covid-19 really done an episode about a timely issue affecting a lot of people before. So consider this a pop up episode. We'll do this from time to time when we feel like we can add something to the conversation. This episode is also special because it has no sponsor other than us.


If you'd like to podcast with no sponsors and you want to support us, you can go to F-stop blogs podcast and click premium.


You'll save time, get transcripts and a host of other goodies and help support what we're doing today.


I'm talking with Baloji Srinivasan. I've said that wrong and I'm sorry. It's not easy to pronounce. I want to preface this conversation by saying the opinions expressed in here are those of the participants and not Farnam Street media. The participants are not medical professionals. I think doctors are some of the most incredible people in the world. And if you're having symptoms, you need to see a GP. Our guest is also not a pandemic expert.


He is, however, one smart dude, and he thinks in a very multidisciplinary way. This episode is full of what ifs, what could be hypotheticals, what's going on, and it borders on sci fi at some point, as you'll quickly find out this interesting interview. I don't know anything with the subject, but I'm about to dive in and learn. I'll tell you more about our guest in a moment. It's time to listen and learn.


Balaji, welcome to the show. How do I say that biology, biology, but yeah, that biology, as long as it's said with good intent, I don't care if people mispronounce it. So it's fine. So we're going to talk covid-19.


And I just want to preface this conversation by saying this isn't medical advice. If you're experiencing any symptoms whatsoever, you should go see. We're not doctors. I'm definitely not as smart as you are. I sort of represent, I think, the average person. I'm just curious about this subject. And you seem to know a lot more about it than most people. Can you walk me through a little bit of your background?


Sure. So I have a PhD non-empty, a Ph.D. electrical engineering with an MS in chemical engineering from Stanford. And some of the specific work that I did in grad school was around genetic circuits and microbes. That's looking at how viruses and microbes are wired internally. Their system diagramed, protein interaction networks, statistical and computational analysis of their genome sequences, that kind of thing. And when you think of that kind of work in genetic circuits or systems biology, that's at the intersection of electro engineering, computer science, genomics, stat's biochemistry, et cetera.


I also have some papers in clinical genomics, population, genetics and so on. I didn't touch the stats for informatics at Stanford for a few years to lecture in the stats department and then in twenty seven thousand eight hundred a genomics company called Council, which the why I was CTO and co-founder of that skill to do more than a million diagnostic tests and change the standard of care for so-called medical and genetic diseases. And we sold for three hundred seventy five million dollars separate from that of a partner in.


And it's a multibillion dollar venture capital firm where I did a bunch of investments and recruited people that helped build what became our crypto in our bio funds. So so that's some of the stuff that I've done in the kind of broadly bio space separately. I also had, of course, massively open online course for 250000 students that a bunch of angel investments in both cryptocurrency and companies, Bitcoin, Ethereum, etc., but also superhuman Soylent Land School. More recently, most people know me for crypto stuff because I've been active on Twitter for that topic.


And I also sold Coinbase and served the city of Coinbase until recently. So, you know, kind of jack of all trades, master of none. But actually my formal training is in genomics and I founded a large diagnostic testing company in diagnostics is an important dimension to this whole thing is you can see from recent FDA and CDC stuff that's been in the news so much, Minox benchmark's diagnostics statistics. You know, those are things I do have expertise in.


I'm not going to give you medical advice at all. But know I think, you know, many papers that are in today and places like any GM or The Lancet with an M.D. and a sheet like me as co-authors, one of the clinical stuff and one of the stats by informatic stuff. And they're usually equal copilots on the paper. So anyway, that's my background. I think I'm qualified to comment on aspects of this, especially those related diagnostics and stats.


But the topics inherently multidisciplinary, nobody knows pulmonology and infectious diseases and epidemiology and genomics and supply chains, all of these different things at once. So so I don't think it's useful to be very narrowly credentialed. So credentials can certainly help to an extent when when evaluating information. That's where I'm coming from.


Preaching to the choir here. Manuma, one of the most interesting Twitter accounts I've followed for years. I've always wanted to have you on the show. I didn't think this would be the reason we get on it. We'll have to do another one later and learn more about some of your thinking around different topics. But I think it's important for this conversation. If we can work to separate facts, opinions and assumptions and if we can try to preface things with this is true or this is what I think and here's why I think it or I'm assuming this to make this happen just so we're trying to get people the best information possible.


So I want to start with something super basic, right? Like I'm not well versed in this. So what's an epidemic?


So an epidemic is the spread of a new disease, dictionary definition, widespread occurrence of an infectious disease in the community at a particular time. But sometimes people use it to refer to like the obesity epidemic, which is a more kind of colloquial usage where at least we don't know if obesity is caused by an infectious agent. Most people think it's caused by eating. Maybe there's an infectious thing on top of that. So people sometimes just casually, but at least in this context, occurrence of infectious disease in a community.


And then it's your next question. A pandemic is when there's infectious disease that is occurring on multiple continents or the whole world and spreading. So it's in Japan in the sense of everywhere. And you know, what's interesting or just about that term for a second WTO, the World Health Organization has not actually declared this a pandemic yet. There's still, I think, kind of, I don't know, into mental model where they either think it can be contained or for political or sociological reasons, are deciding not to call it such simply from like a dictionary definition standpoint.


This is definitely spreading uncontrollably on multiple continents. So in that sense, it's a pandemic.


What caused the hesitation for labeling it a pandemic? I'm not an expert on the internal politics of WTO. I've seen people charge that it is due to pandemic bonds, that we get paid out or chose relationship with China or for various other kinds of things. I'm not I'm not privy to that. I wouldn't be able to give an informed opinion on why W chose not declared a pandemic. However, I think a number of folks in the space could be watching this, whether in diagnostics like my background or or epidemiology or genomics are kind of wondering why that official declaration hasn't been made, because it's independently observable.


You've been running the media on this for months. You seem to be well ahead of the curve labeling things and sort of like ascertaining what's going to happen. What is it we're even talking about?


What is like covid-19 to covid-19 is the formal name of this disease whose initial cold initially called the novel coronavirus and and Kov 19 novel coronavirus 20 19 denoted both the virus and the disease. And now they have sort of factored those out into sars-cov-2, which is the virus and covid-19 the disease. Now, I think the reason they chose that is they didn't want to. And actually I think this is a good, smart thing on their part. I didn't want to stigmatize a region by calling it like the Wootten flu or or the kind of flu like read.


It has a community called the China flu. That's not helpful for lots of reasons. Like that area is hard enough hit. You don't want to hit it like for the rest of time, kind to blame it on them or anything like that. So I agree with their decision to kind of give it the sort of abstract name. But covid-19 is the name of the disease and sars-cov-2 is the name of the virus.


And so what is it like? Where does it come from? I heard a rumor there's like a bat or something.


So there's something called molecular phylogenetic where you can actually build evolutionary trees of not just humans and dogs and cats and so on, but also viruses and bacteria and what have you. It's called molecular genetics because normal genetics, you're kind of just looking at macroscopic features of an animal like does it have four legs or two? Does it have four or not? And trying to classify them that way. Those are like features with molecular genetics. The features are more like, do you have an AI in disposition or C gravity or in this case it's an RNA virus, but the concepts are the same.


You're not using DNA, but RNA and the the molecular genetics of this coronavirus, you can you can put it in a tree with other related coronaviruses and you can see that it's similar to the coronaviruses that came out of Yunnan Province. So evidently there's some viruses that were in Batz or at least labeled as such on on GenBank, where they appear to be. Viruses have arisen in Batz from Yunnan province in China. So this particular strain of virus appears to be most similar to those.


And so that's why people have talked about, oh, did it come from bats? But we don't really know. I don't think I've seen anything that actually nails OK, this was patient zero and this was the exact moment people think it's something in the vicinity of this plant in the food market.


And one but I shouldn't say it doesn't matter. It does matter because if you know the origin of it, I can give you a clue for how you eat, for example, culture, the virus or or something. So sometimes a clue that's not that's obvious if you understand how it came about, like maybe the animal that spreads it is immune to it and maybe there's some molecular biological feature, that animal we can clone for treatment, that kind of stuff.


But the origins people speculate it's a so-called zoonotic thing where it made the leap from animals to humans and they speculate that had something to do with people eating animals of this at this seafood market, which is not just seafood, but has a bunch of other kinds of animals there. Is that because you're eating like raw animals or they're not cooked properly?


Well, you know, just eating the animal in general or certain parts or I believe that these wet markets in China are selling.


I have not personally visited the one and see the market. I don't think it's going to be a tourist attraction for the near term. My understanding is they sell both dead and live animals. And so we're used to buying chicken, you know, but there you can buy a chicken or a fish and kill it yourself and slaughter. Right. And so there's just folks who have written about how animals in that condition are very stressed out and their immune system may not be as strong and they're in close proximity to other things to jump back and forth.


I can't see how that thing arose, but that's likely. That's let's call it the most she's the most plausible thing. That is the default assumption right now. Right, that it came from that Kwanten supermarket, but with molecular biology. And I think eventually we'll will be able to resolve that and trace where the most recent roots of this virus are. But that's kind of an archaeological dig that may not happen for some time till things calm down.


How did we find Patient Zero? So, to my understanding, again, in literature, you see that in early December you start to see patients in Bellhorn and these patients were presenting with something that felt the community sort of like SARS 2.0 because it was something people remember from 17 years ago, from 2000 to 2003. So these patients were presenting and people were seeing more and more of them. You know, by the way, patient zero in the colloquial sense means the first human to contract the virus.


But in medicine, it's usually the index case or the first documented patient in the disease epidemic. Those aren't always the same person anyway. So the first documented patients were discovered in December, but if they were sick in December, then it probably means it got started sometime in November, got infected around the time because it shows that the virus incubates for one to 14 days. So if your first case is December 1st, the person probably got infected November, then things started to there.


I just want to walk through this head here so the patient shows up at a hospital, looks like SARS 2.0, exhibiting symptoms. Then they do a what do you do a biopsy? Like how how do you test for something like that? So the short version is that you take a swab, collect a sample from the patient's mouth or nose, you send it to a lab and you get back a positive or negative test result. The longer version is that it starts with sample collection.


You can take a swab, as you mentioned, from the nose or mouth, but also other ways of collecting samples like aspirates or washes, which which basically vacuum or wash material out of the patient's nose through into a sterile container. This is a respiratory or a virus that seems to target the respiratory system. So that's where people are typically collecting samples from the nose in the mouth. Then what do you do that sample? There are a few different kinds of tests.


An important one is the so-called art PCR, critical of the CDC published Real-Time Reverse Transcription PCR. As mentioned, you take these samples swabs, then you extract RNA, the that RNA into DNA versus reverse transcription, and then you detect the levels of DNA by amplifying it via PCR, PCR as the polymerase chain reaction, a standard technique, Tampopo can detect small bits of DNA even if they're present in a mix of other stuff. You can use it for things besides detection.


When you're doing this RTP, you also have some controls in there to check against various elements. Now, if all the controls come up right and did the right, the virus is present. You could see the curve representing the amount of DNA copies of the original RNA rise quickly above a certain threshold. And that shows you that the viral RNA that you're looking for is abundant. And, you know, the actual CDC test looks at this not just for one locust, but for three loci, meaning three different spots on the viruses.


One of the spots, I believe, is for all coronavirus. And to look loci are specific to the sars-cov-2 virus, the virus that causes cocain 19. Incidentally, the primaries used to detect these laws have had bugs in them, as been widely reported. Now, the primary thing you get out of this are TCR type test is basically zero one presence absence signal of where the virus is present. You also get some quantity of information on viral load, a measure roughly of how much virus was present in the sample.


So there are other kinds of tests, though they're worth mentioning, important ones, so-called sequencing assays that give you not just a zero one for the presence or absence of the virus, but the full RNA sequence of the viral strain, the ACS use in genes rather than ACT genes because it's RNA. An example protocol for this is the Illumina protocol that was published to sequence sars-cov-2. This gives a lot more information than the other two PCR protocol as you get full sequence data.


And that can be used for molecular genetics, for looking at viral evolution, that kind of thing. Now these tests, I should emphasize, are quite standard. People have been doing our PCR and sequencing on aluminum machines for basically a decade plus. At this point, the main difficulty here is that the sars-cov-2 virus is very contagious and dangerous. So you need to be even more careful with sample preparation than normal.


That makes sense. So how does it transmit?


OK, so first let's start with some definitions of the Arnaut, which basically an R with the subscript zero is the basic reproduction number. The Australian Department of Health defines its average number of cases directly generated by one case. In a population where all individuals are subject to infection, there's no immunity from past exposure to vaccination, nor any deliberate intervention and disease transmission. So if there are not three, the first person infected and passes on to three people who in turn each passed from three more, giving nine and nine people infected and so on.


So it's basically this exponential now going from that same situation, from health into the basic reproduction number or not is contrasted to the effective reproduction number where there is some immunity or vaccination or some intervention measures are in place.


So while some people say this clinically, at least by that definition, you can't reduce the basic reproduction number by human interventions like vaccination or quarantine or lockdown. That reduced number would be the effective reproduction number after human intervention. And so the Arnon is supposed to refer to what happens when a disease is tearing through a population. There's no natural immunity or vaccination against it, sort of like we have with covid-19 this we call it the basic reproduction. No human that basically no, it's not fully consonant.


It can be modified with things like environmental conditions. For example, people think he could theoretically make the virus break down outside the human body more quickly, that that's true for some other other kinds of microbes. So it's not a universal constant, even though we're not. Now, with that said, we kind of care about the overall reproduction number, the effect of production numbers. What can we do to reduce that? So we'll even before we have vaccinations, we can do things like cancelling large events.


People say viruses love crowds. They also look chokepoints like door handles, bathroom faucets, things that lots of people touch. And insofar as we're advocating campaigns of hand washing sanitizers, hand sanitizer, use face masks and so on, we are reducing the effective reproduction number are. Now, one important point, though, is that while we can tell that covid-19 appears to be highly contagious, the exact reproduction number that are basically effective is difficult to estimate. Precisely, and it has Whitemarsh.


And let me give you an example. With Facebook, they can go to their database and they can poll every single person who invited every other person. Right, so they can find this person went invited seventy three people and happened in these specific times, they can pull that out of their database from an event that happened in twenty seven or twenty two thousand nine in September. But to do that for a virus you don't have that. We aren't recording that.


You don't know exactly when it is spread. So you have to kind of infer it. And so it's necessarily kind of statistical. With that said, we have some very new tools for really interesting. So something Trevor Bedford has shown, this is the first time I think we're seeing like public viral bioinformatics on on social media. Trevor Bedfords, profit UW. He's at TRV on Twitter. And his lab is something called a project called Next Dream, where essentially being the kind of web interface where they can sequencing viruses of different kinds of things and looking at infectious diseases and tracking them over time.


And here's the awesome thing. In twenty twenty, it is possible to sequence a virus, get that into GenBank, which is like a like a database for all of these things. Analyze that, post the data online and then post about it on social media and have that all happen like in twenty four hours. The cool thing about what Trevor's analysis showed is he was able to take two viral sequences, one from a patient around January 19th and one from a patient on February twenty ninth.


He could show that it was very likely that the viral strain on February 29th descended or was closely related to a common ancestor of the one on January 19th. And the reason that's important is it showed that there was likely a cluster of spread happening in Washington that was not contained. The alternative scenario would have been the patient on February 2010, a strain that was related to someone to a Chinese dream, as opposed to strengthen, mutated and diverged.


And every time it sort of infects, somebody mutates. But my understanding, correct me here, is there's a core fundamental that doesn't change between them, which is what makes the vaccine possible.


Well, this is a sophisticated question I take. It's not true that every time it infects somebody, it mutates. But it is true that because viruses are like a replicating so frequently that this virus is actually quite mutable. It does accumulate mutations. But you're right that there have to be certain conserved aspects of biology or else the symptoms start to change. Right now, sometimes often viruses evolve to be less lethal and more contagious because like if you if it's like Ebola and somebody dies in a pool of blood, then they don't spread very far.


Right. The Ebola virus actually sort of kills people before they can spread it.


But for for this one, it's not obvious that it evolves to become less lethal because it's actually already super contagious and it's asymptomatic for a long period of time. So once you spread it, it doesn't care about you that much and might kill you or or you are pretty bad to that. That's bad because it's more like Spanish flu where it's super contagious and packs a punch as opposed to being super contagious like flu or extremely lethal like Ebola. It's your nearest Ebola, but it's scale is likely to be much higher than Ebola.


And so the total number of people hospitalized or killed by it, maybe, maybe much larger, are those other common variables associated with sort of viruses like hospitalization rate, death rate? What are the other variables in that sort of equation that are the common common ones people talk about?


Right. So there's the there's ANA, which does kind of influence your total number of infected. Right. The cumulative infected. There's hospitalisation rate. There's death rate. You know, one thing we don't know, by the way, is what the long term effects of this thing are. For example, polio didn't kill everybody, but it did put some people who paralyzed him for life, even though they survived and survived. But he was paralyzed. We don't yet know what the long term effects of this.


Sorry, I don't think it's something which is something that's so severe as to cause death and otherwise healthy people. Within a few weeks, the folks who have severe conditions may have long term impact that we're not aware of because it's so new that we don't know what that long term impact is. Maybe they have impaired breathing. I'm seeing the reports to say that this is a man. I haven't looked at all the literature, so I'm saying may, but it may have changed the reporting that may have neurological testing.


So maybe attack the nervous system, the immune system, just respiratory system. The severity is a continuum. It's not just dead or, you know, like bounce back. It's probably something you don't want to get. I mean, there's a lot of other variables that that people look at, so this is kind of a derivative of other things. But the doubling rate, how fast? Two cases double. And this one looks like there are seven point four days in the first intermediate time constants like the time for when you're infected to showing symptoms in time for when you're showing symptoms to when when you die, if you die.


Right. Certainly not everybody dies from this. So, like, those vary a lot. Politico's recent report is reporting like one to 14 days for infection, showing symptoms. And for those who do pass away like two weeks from symptoms to death with a large range, you know, those are some important parameters off the top of my head for thinking about this from a logical or mathematical modeling.


That strikes me as really interesting that you could possibly have a 14 day incubation period with no symptoms. But you're you're you're carrying it. And I assume that you're passing it on to other people during that time as well. Yeah.


So this is the really tricky part about this one. There's a lot of reports of asymptomatic spreading, and that's really dangerous because, you know, if you're asymptomatic, you don't even know you're really symptomatic. You don't even know. Right. Here's the one good thing about being asymptomatic is, you know, it means that there are still some people who can and can go out and do things right. But it may turn out that we tested very, very large number of people because there may be more spreaders than we think.


And it may turn out that we need to make testing extremely routine, which would basically mainstream genomics. Everybody would have, you know, like sequencing all the time, sort of like wearables, you know, your your your your heart rate and your steps and so on are being tracked continuously, sound sci fi. But it's actually not impossible to have, for example, a home sequencer that's plugged into your plumbing. So you tend to think this may not be the form factor.


There may be other form factors. But you the think or you put a sample into this somehow and it is automated sample prep and it looks at what's there. Right. Or maybe it's plugged into the sewage system or something. There's a ton of technical details to work out on something that's very high level kind of concept. Sample prep is important to different people is important. But in theory, from that organic material, you could do sample prep and then you could do what's called environmental sequencing to determine all the microbes and viruses that are there in that right.


And then on your phone and you'd see if you were getting sick or someone in your household was getting sick. And I kind of think something like that is going to be the types of things people will build, both to measure and contain this. And it's sort of a Hygiene 2.0 measure in the future.


So this will accelerate some sort of positive development around that. Is there the test kits that they're using now?


Are they real time like are they know where you be able to go to the airport before you board a plane?


Like, spit in something and. No, they're not. Even though it's called real time PCR. It's not it's not that fast. It's the real time aspect refers to something different. But still, that will be technical term. You're basically looking at the ratios of curves. You're not going to get it on the spot. That is a different term for that. It's like point of care diagnostics. You're thinking of like a pregnancy test or something. And I do think we want to get this stuff into a actual point of care is a little different from apprentice's ultimate where you can do it at home.


But ideally, you want to have something where the point of care is at home and you can kind of test yourself, you can plug into your smartphone. And we're not there yet with this yet. I think we're going to have to get there rapidly. And there's the potential tools that that that could be used for that. You know, there's also proxy's you don't necessarily have to go to the molecular level right away, but you can look at something like burring or even a thermometer and just measure your temperature.


Right. And and so then that's symptomatic. It's not going to detect the asymptomatic stuff, but it's maybe a cheaper and more widely deployable proxy because thermometers are well understood, variables are well understood. That's relatively easy to deploy.


And I think the WHO report said like eighty nine percent of people that were infected had a fever is somewhere around there.


Yeah, I don't remember the exact stat of my head, but you have to this is incredibly high like it was. Yeah, right. Like like basically it is, it's the most characteristic symptom. Right. And that's what kind of folks are asking drivers to record their temperature and so on. So orderings or things like that. I mean I'm not an investor or anything like that, but but I know that's pretty cool or it doesn't monitor your body temperature every time something like that I think will become very mainstream.


You can see the variation. Yeah, that's right. Actually, there's another study which which had to do a while back, which was wearables in general, the seemingly good diagnostic rate information where you can predict the onset of illness from them. There's been papers published on that. So so you might get farther with wearables than you might think you might need to go to a molecular diagnostics immediately.


Are we going to a world where, you know, you can. I have a company issued or a or Apple phone, and they basically say, don't go into the office today because we predict that you're likely contagious. Oh, absolutely not with it. Not with this particular virus, but with any virus in the future. Absolutely.


I think this is going to. So let me give the silver lining to this whole thing, because this is a dark cloud. There's a lot of people we are just at the beginning of this potentially. There's a lot of people in a lot of people infected. So I don't mean what I'm about to say in casual way at all. But the silver lining of this is this could be as much of an accelerant for biomedicine as the Internet was for software.


For example, if you think about this from a like a funnel standpoint, every aspect of the system, from diagnostics to treatment to the analytics on the population, you need to innovate like five or six different things. OK, we're going to want at home diagnostics, romance stuff to plug into your smartphone. But then on wearables that give continuous sensing, we're going to want pandemic surveillance stuff that's monitoring effluent and sewage systems both at home and elsewhere are going to want.


If I didn't mention telemedicine, we need that. We're going to want medical delivery robots so that people who are really contagious can kind of be in an isolation chamber and the delivery of robots can deliver them the drones deterring doctors or nurses. We're going to need much faster clinical trials for drugs and vaccines. We're going to need like these instant hospitals that people have built. We need to be able to add medical capacity and like a surge form to build this type of thing.


I mean, there's so, so, so much know, I think at the end of the day, what happens is, you know, an entire missile defense system, when there's an incoming missile, it has to react so fast. It's directed the missile. Right. A human being involved maybe too slow. So it has to track the missile, launch the counter missile and hit a bullet with a bullet to knock out the air and blow it up like the Iron Dome system in Israel, a pandemic defense system, the kind of thing the Chinese will build a cold with the third Great Wall.


And then somebody actually used that term in China as well. Like a few weeks later, there's a Great Wall of China, the Great Firewall. And I think we're going to see the Great Firewall of China, where they'll be sequencing and testing the heck out of everybody, scanning the country constantly to see, OK, is a virus rising and then quarantine that area, vaccinate that area and just just stop it from spreading. Because this is such an enormous scar on society.


No one is going to forget this. That I think is going to be a significant part of their defense budget going forward. And that's just that's just stuff I can see. Everything I mentioned there has been kind of happening like two months ago in China. You know, like like, for example, the telemedicine, the hospitals, the delivery robots. They built that in Geneva. Right. Like they've been at a wartime pace of biomedical innovation since January, which is part of what got me, like, really interested in this, because, I mean, the Chinese government and the Chinese people are serious people.


They wouldn't they wouldn't commit economic suicide. As my friends Twitter friend John Stokes would like to simply shut down their economy for nothing is not characteristic of them. So I took this seriously based on that like extremely indisputable signal of reality. And and the technological innovation that's coming up there has been amazing. And I think we're going to need to know that and learn from that in the West in order to deal with this.


What would be the second order sort of consequences of. A bio loll around a country, not specifically China, but just like we do a thought experiment here, and each country is now isolating itself from other countries and the spread of disease maybe gets reduced. But does like human resiliency go down like water, the sort of consequences you can foresee?


Well, so for one thing, before you travel to certain countries, they want to know that you've had this vaccination or that vaccination. Right. You've probably seen it before. Yeah. Now that's going to go to real time, meaning within a day or an hour before you get on the flight. Imagine the next generation, TSA, that was actually competent, you know, in countries around the world, which means you're your biological information before you can get on a flight or enter the country.


As in like you give a blood sample or saliva or something like that? That's right. I think that's where countries are going to go because screening is already happening for every traveler in India coming in like that. That just my. And so as countries react to this, they are going to require diagnostic testing, just like, you know, you have to make sure you're bringing no guns on a plane or bombs on a plane. Right. That could kill hundreds.


They want to make sure you're not bringing a deadly virus to a country that could kill thousands or millions. And so I think that kind of screening is going to happen. It's happening already in the sense of checkpoints within China, but that's going to be rolled out. And in a sense, it's just a generalization of, OK, you need X and Y and Z vaccine to enter the country like Singapore for a long time is required evidence. You've had yellow fever vaccine for your country country, right?


It says it's not like completely unprecedented, but the scale of it is going to be totally new. And so this is going to be a gigantic shot in the arm for every diagnostic company in the world, every medical imaging company. All of those are about to get like defense budget level spending, because you can think of this as a world war in terms of its scope. But it's a world war against virus.


Do you think covid-19 is the is the straw that sort of pushes this over or the piece of sand, I guess, that piles on and topples the sandpile over?


Well, I think it's definitely going to cause at least a recession, but likely something more severe than that. Some of those things where there's so many secondary shocks, if you remember, like the 2004 earthquake in the Indian Ocean that led to a tsunami that devastated many countries in the region. And that's kind of how I think about this. There is the badness of the virus itself. There's the second order kinds of things that riechmann, for example, supply chain shocks.


You have all the people in China who can't go to work, can't go to work. They can't produce the screws for your for to maintain your radiator or the even more dangerously like necessary components reagents for drugs. Because supply chains are really complicated. People don't understand how many different parts go into, you know, like a printer or even a laptop for sure. And or even like a container of milk.


You've got the label and the ink and the this is that the disease causing that, or is that the reaction of people like the flu infects, you know, a decent number of people every year and it may have a lower sort of death rate, but it's still very transmissible and it doesn't cause people to shut down factories or. True.


But the thing is that, you know, this analogy to the flu is is very bad. I think in no way attack deep dive on this. Yeah, no, no. I want to dip down on this because it's the most common sort of message out there in the media, like, oh, my God, why should why should we care about this more than the flu?


You know, the flu is going to infect 40 percent of people over the next five years or just making that number up. But it's going to affect a decent number of of people over the next five years. And, you know, we do have vaccines against it that are sort of like reasonably effective health care practitioners are trained to identify and treat and like.


So. So why is this?


Well, we have to we have to make distinctions. Right. Like a common cold is not is not HIV. Right. You know, there's there's there's there's infectious diseases. There's infectious diseases. And so analogizing it to something which is also an infectious disease is not helpful in this case. The one there is a similarity in the sense of it spreads like the flu strains and B of influenza of people are hurt. It seems to be airborne. Right.


Which is people are coughing and what have you. But compared to the flu in terms of severity is not the case. And there's there's a graph which you might be able to share with you.


But just how can I push back on that first thing? Like, how do I know that? Because I'm assuming and I'm totally naive on this, so I'm an idiot, but I'm assuming the number of infected people are astronomically higher than what's reported to me.


And I've got a I've got a great one.


Yeah, but if the death rate is accurate, wouldn't that reduce like if the number of deaths as a result of this is reasonably accurate, wouldn't the death rate actually go dramatically down when you assume the number of people that would possibly be infected?


Here's my short answer to this, which I actually tweeted a while back. But now I think it applies to Iran and Italy and so on. OK, Wukan was a normal city on December 1st of last year, and by January twenty third, all seven hospitals were filled. People are being sent home to die. The entire city was put under quarantine and it was the largest like. Quarantine in human history began. That was not a bad flu season.


So is that Chinese reaction or is that like a necessity? Like do you think that. Well, here's here's the thing. Like, you don't I mean, it's one of those things where, first of all, Iran seems to have had a similar experience where I know if you seen this guy on Twitter, Ali Assad, and is tweeting from Iran or DeLucca on, he's giving stats from Italy and people are showing the number of offenses section on Wikipedia from the Italian government, increasing exponentially in the first week the to the region of Italy that's been hit by this and have run out of beds.


OK, now, when you're seeing things from China, Italy and Iran, I don't think this is psychosomatic. I don't think it's a lot of people who are hypochondriacs going to the hospital like it's a hospital. Right. It's expensive. You know, it's even if you have universal health care or what have you. And I don't think that's it's not a trivial decision to go to the hospital and into. So I think that in very different societies now, we have to kind of take the warning and it just it sort of beggars imagination to believe that the Chinese government would have 750 million people under lockdown and quarantine because they're scared of us.


That's just not what this is. It's simply not the most parsimonious explanation. You know, you can you can sort of wait for the storm to hit somebody before they'll they'll listen to that. And some people will be like, I'm not saying you are right, but all the evidence we have indicates that the absolute number of hospitalizations and deaths and this hysteria is very high. Even if it is true, there's lots and lots of mild cases. The end is so large that it overwhelms the medical system and the number of people requiring intensive care.


Exactly the number of available beds, which further and how does it transmit? So it's airborne, obviously, but it's also surfaces and.


Well, yes, I mean, it seems highly contagious. So it being airborne, which really sucks because you shouldn't wash your hands. You should avoid handshake's. But that won't necessarily save you, because if it's airborne, right. It appears that fecal transmission is another vector. And that was something that happened with SARS, which is a related virus 17 years ago. But there's a famous infamous case of an apartment complex in Hong Kong called Amway Gardens. And about three hundred people there contracted SARS because they give some of them diarrhea.


And this is gross, but some of them flush and the sewage system was imperfect and some of the droplets were never coming back up. And so people were all infecting each other within the building. That's a more general thing, which is apartments are not biohazard contaminants. For example, if you think about your soundproofing in an apartment, it's not perfect. Right. You can sometimes hear the thumping Bass River. It hasn't been tested for military grade soundproofing or whatever the ultimate level is, anechoic chamber kind of soundproofing in the same way.


Yeah, it's probably true that you're going to be less communicable in your apartment, but there's still like you're still connected to the same air ducts and you're still the same. And people are pushing an elevator button. So it's not a hundred percent. You said you had some numbers before, so, yeah, I digress a little bit because we were talking about physical transmission. Yeah, I gave my guardian's example. Right. And so I just gave the example is fecal transmission.


Most people want you to say, well, you know, I'm not eating anybody's poop. Exactly. But unfortunately, in in in in my gardens, people were amazed that you realize that, number one. And number two, in many cities on the coast of the US, like San Francisco, most infamously you are definitely people are definitely inhaling fecal matter as they're walking outside. They're definitely getting it on their shoes and so on and tracking it and even in small quantities.


Right. And and that is something where if you have one infected person who's doing that, there's some that there's a guy, professor from UC Berkeley, who warned about this two years ago. Actually, an NBC report is like the streets of San Francisco are at a developing world level in terms of public health. So the warnings have all been there. But it's really like the city is kind of in immunocompromised city, San Francisco in particular. And it's going to be pretty dangerous if and when I think it may have already happened.


We'll see if a control community spread is happening. And like it seems to have been happening in Washington, Washington state, that is. Yeah, that won't be good.


Why are kids less or more resilient than adults like it seems to almost correlate to age, but younger is least affected to older being the most affected. Is that accurate?


I mean, you know, one thing with all of these these types of things is this is sort of these are emergency. Conditions right now, so it is good that we have demographic crosstabs. We can see that, you know, the old or the elderly seem to be more affected, much more affected, especially as you get 70, 70 year old. And we could see the young appear to be less affected because he seems to hit men harder than women.


But those are kind of macro scale demographic observations.


The molecular biology of that, I think, is probably still a hypothesis because it could be one of a bunch of different things biologically that are different between a kid and somebody older. It could be a bunch of things. Biologically, there are differences between women and men that are causing this. And I don't think we know the molecular biology of what are some of the second order impacts of what's happening now. So like around here, I'm assuming everywhere around the world at this point, Purell is like soul.


There are there are they're sick. You know, it's really interesting to me is like I was at the store the other day and I was trying to get some Purell like everybody else.


And the store when you go online, because it's like 300 bucks on Amazon right now.


So I think I was thinking about how retail everybody's like retail is moving quickly and, you know, they're catching up on this technology thing. And Purell was on sale everywhere because some guy like six months ago decided that they would have a sale this week on Purell. So not only is it out of stock and hugely in demand, it's on sale like two for one tons of stores. Anyway, that was just my retail observation. But what are some of the second order consequences of his Senate has to use?


Like everybody's just massively going to use hand sanitizer now?


Well, I mean, some people have said, oh, you're just going to get microbes that are resistant to hand sanitizer. That may be true. Probably true. I think, you know, this is something you both are in a somewhat different scale. So here's a few of them that I'm going through. All of them. Yeah, sure. I mean, to downplay that, I do think it's possible that likely, in fact, that you're going to get microbes that are eventually resistant to hand sanitizer if it's widely used for sure.


Still, you know, like let's use the hammer we have now and then come up more later for something like this. I think that's probably probably reasonable. So the kind of secular design thinking of here's a few of them and the one the supply chain shock. Right. Which is to say all this and trying to just shut down. Lots of things are made in China. All these small companies in the US that depend upon Chinese parts of these hardware companies, I mean, private companies are just going to die, many of them, because they just can't get the parts from from Chinese warehouses.


It's like it's funny to use the reverse analogy, if Amazon Web Services went down, a lot of sites would just go down. They just do not have a backup option. Right. And this is like that. The physical clouds kind of use a digital analogy to explain the physical world. So one supply chain, number two, travel shelf, airline travel conferences, events, rallies, sports games gone like canceled and lots and lots of countries. And maybe for an extended period of time, depending on how bad the airport gets.


Do you think they go away or do you think they switch to conferences, switch to virtual? Do you think like sporting all switches to TV now?


Yes, I think well, I think what happens is historians of the future may write something like twenty. Twenty was the year that the Internet actually began right where this comes for the primary in the mirror. So in the mid nineties for The New York Times first, but online, The New York Times dotcom website was just a tiny mirror and the physical paper was primary. Just a few articles that were online right then over time, the the website for more and more and more and more of the load.


And I don't think there's any particular announcement externally. Certainly today you would say NY Times dot com is a primary and the paper, the physical paper should print out at a particular time. That's the mirror. Right. So it gradually shifted from the physical is primary in the digital. It's a mirror to reverse where the physical is just a printout of the website. And in fact, there are certain things that are digital native that you couldn't print out, like the interactive graphics or or links to tweets and better tweets.


I mean, it's harder to print that, right? It doesn't make sense. The data science experts, once you start going digital native. And I think that this year is going to be it's kind of like, you know, during World War Two, many women enter the workforce and then even after World War two, a new normal had been reached. Right. This is what's called hysteresis in physics. It's like you apply a stimulus to a system, then even after you remove that stimulus, the systems in a different state than it was before.


There's some memory in the system. Right. So this is going to do is it's going to force everybody to look at remote work as not a auxiliary, but as a primary. This is one of the grand challenges, I think, for a tech as opposed to biotech to work on. And it's going to have its hands full building what I call the remote economy, not just remote work, like every job has to be a remote, capable, not just like legal work and programming and and graphic design and all that stuff.


We need telemedicine. We need autonomous delivery. We need people to be able to operate forklifts from a factory all the time.


So it sounds like you're saying this isn't going to cause it, but it's going to accelerate that these natural sort of trends or things we've been dabbling with because they'll be a huge health benefit or a perceived health benefit in the future to having things operate. But then what are the second order consequences of that social isolation? Don't we need to feel part of something and part of society and connected to our community? Like what are the psychological impacts of if everybody worked remotely?


So what I think happens is you get to competence that all the things that are canceled, people are going to want to bring back or bring back in some some other form. Just a couple of options for that. First work from home is going to get really good things like Home Office chat and so on. People are going to try. There's a hundred different things that you could argue or the reasons that people aren't as productive at home and people are going to push that in VR is an obvious angle.


There's a bunch of other obvious angles to try. Like maybe it's it's like a Facebook portal. I think the tricky part is anything that's related, maybe maybe tough because supply chains will be hit. Right. So there might be demand for the beer and they don't have the ability to supply them, which would suck for the VR companies. Let me figure out some workaround, though. So one aspect is that remote work is going to become really good.


The second thing is people are going to lean harder into social networks and EA Sports and all of these online compliments or proxies for offline socialization. A third aspect, I think, is that people are still going to want meatspace or physical space. I think you're going to start to see trusted communities where you're literally trusting the other person. You have very high trust that they are taking all kinds of hygienic measures. Right. So it's like a bio trust.


Yeah, it's a bio trust. Right. So that's a very high bar. Like, do you trust this person to take such good care of their own hygiene and health and to be so diligent that they're not going to put you and your loved ones in danger? Right. It's a pretty high bar. And so I think that when you start to get especially if people are locked at home for a while. I mean, all this homeschooling stuff, right?


Basically, you know, there's the educational system has been completely transformed with all of these children center and there's all these parents there. So I think you're going to have these community organizations form where a group of folks who know each other to not be infected or trust each other to maybe trade off. And then they start like educating the children amongst themselves or using online tools. And each each parent has the kids for a day or something like that. Right.


And it's almost like a forced way of, you know, I am an internationalist. I, I, I'm an Internet person. I love all that stuff. This is in a weird way, like like a forced globalization, radical forced globalization, because all the things that were benefits, like economies of scale, large groups of people, big crowds and so on, are now big demerits, at least in real life. And so forces very small scale in the physical world.


And I think you're going to get scale of the digital world as a complement to that until we get vaccines in all kinds of things and at least aspects of life. Conventional, right.


That makes a lot of sense. What are sort of some of the other second order consequences you see to the economy, both positive and negative to the whole? And one thing that I do want to point out that I like is that a positive is that governments are handling this very differently and that will enable us to see how different things play out over time and get feedback on how we could handle the next one better, I'm assuming.


Well, yeah. I mean, I think of this as like the decathlon for governments because you can't be ideological about it. You might need to for example, you might need to mix a border, shut down and quarantine with universal health care for the vaccine, with accelerated technological capitalism to push out the telemedicine and the delivery robots and so on. Right. Right. So doesn't fit neatly into ideological categories. It's just about competence. Know. And so who's winning that international decathlon right now?


Probably Singapore and Israel. Right. Know, I haven't looked at what Australia is doing, but I wouldn't be surprised if they're very much ahead of the game.


So what are what are they doing differently or are they doing differently? Well, I mean, Singapore's leader, Lee Qianlong, is Lee Kuan Yew son. If he's not just good at math, he is the he was he was the wrangler at Cambridge. That means senior wranglers is a guy who has the top the top mathematics on the graph, who has a degree. And in fact, he has posters like Sudoku Solvers on Facebook and in C or C++ and look like he's intellectually knows Coke.


So this is somebody who has like a technical background in Estonia as well. They've got a very strong computer science engineering influence in their government because they've kind of decided in the early nineties to become an Internet country around the time they got independence. Right. And so the critical point here is something much that the prime minister needs to code the response themselves. That's not what I'm saying at all. What I am saying is there's a critical mass of people with scientific and technical expertise in government who in politics, in the media and so on in these societies.


And those folks are capable of doing telephone calls like growth rate extrapolation rather than based extrapolation. You see towers post office.


Yeah, walk me walk the audience through the differences between growth rate extrapolation and base rate extrapolation.


Sure. So very roughly, and this is like maybe a little different than how Tyler himself is a growth rate. People model how the world can change. Basically, people think about how the world is basically. And so the three people are often mathematical, they're used to exponential growth for things like iPhones or Facebook or Bitcoin, they're looking for large deviations are often investing money or quantities or gold or technical or scientific. Basically, people are, you know, and they're they're often correct in the default that things aren't going to change.


I mean, an institution is almost by definition of base rate thing because it's kind of it's feed started. Right. Right. But basically, people are there about the conventional wisdom. They like mainstream things. They think, oh, that's weird. When you point them to something which is at 0.01 percent in a sense, anything that's below 50 percent of the population or a large enough thing that doesn't have a constituency they kind of ignore because it's just not important enough to matter.


The problem is for them, I think, is if something is growing exponentially, then it's one percent, it's 15 percent and it's seventy two percent and then it's one hundred percent. Like, pretty much everybody in media, for example, has a Twitter account and just goes like this really fast before they have a chance to react to it. And so I think the issue is that the most the people in politics are, you know, their lawyers or they are career politicians.


There's really a deficit of folks with technical engineering degrees. And the interesting thing is I was able to reframe this and think about this myself. I actually think that's on us in the sense of folks who have any form of technical or scientific background, I think need to as like a civic duty. And this is both for this crisis and then for future kinds of things. We have to do journalism, citizen journalism, like a year of it, for example, per person, something like that.


We have to actually get involved in politics and and things like this because otherwise there's going to be a deficit of folks with that training. Because one of the drivers of this is you can definitely make a lot more money outside of politics and journalism and whatnot by if you have a technical background by going to technology or to Wall Street. And that's definitely one of the factors. But this meant is folks within those demands are making decisions without an intuition for math and intuition for science.


And you can't teach them all of that in the instant that they need to understand and grok it to deal with the pandemic, whereas Singapore sort of like takes the top students and then tries to fast track every government, right?


That's right. That's right. There's more shared knowledge. Right. If if I say, oh, I can explain the Arnott and the Singaporean government, I don't have to explain to them. They understand that it's right. But if they hadn't heard it before, they've done a spreadsheets or what have you, that they can understand the implications of that. I'm not sure that our government in the US is selected on that basis.


What's your opinion on how should governments respond? What should governments be doing? What should they do? That's a great question. So would have been tweeting back on. One of the things I'm writing up is something I call the decentralized response because there's a centralized response and a decentralized response to China's response is centralized. The Chinese government actually has a critical mass of folks with engineering degrees. I think decision paying has a chemical engineering degree from Genella and Hu Jintao, his predecessor, I believe, has a hydro electric engineering degree.


Also from doing so, government has centralized response, which is CORNTON. Things do basically surveillance in the sense of both pandemics, surveillance and contact tracing. And the government was basically driving everything, you know, certainly like lots of private citizens were doing things. Lots of companies for us to do things was kind of the driver seat. I think in the US that's not working, the centralized response has been very poor, in particular, FDA and CDC and HHS have not been doing a wonderful job.


And that's because if you followed this whole thing, HHS declared a public health emergency, which you would think would push them through radically. It didn't. Instead, it meant that there was now a higher bar to get diagnostic tests out because FDA had so-called emergency use authorizations that had to grant to come collapse before they could test netnet a bureaucratic process inhibited the United States from testing for a critical 30 day period from January to February and give people a false sense of security that there weren't that many coronavirus cases in the US.


So the simple response so far has been a total, I should say, mostly a failure. The response, though, I think, has shown promise that there's these folks at UW who developed a diagnostic. There's that professor I mentioned who is estimated the number of cases. There's also the fact that the first patient who was written up and removed from medicine and January was treated under compassionate use, which I believe use the more recent rate to try Padthaway.


I need to double check this and verify this, but there's no mention of FDA approval. In Chiliad statement, they talked about local regulators approving it. And this is kind of a wrinkle, by the way, that just to explain this for a second, so not everything in health care is FDA approved. Clinical labs, for example, are regulated by another entity called CMS, the Centers for Medicare and Medicaid Services, under a program called CLEAR. But CMS and FDA are under HHS Health and Human Services, which is the parent agency.


And in CMS, the Clear Program for Laboratories works with local state agencies to regulate lab tests. The way it works is that the lab itself gets cleared by state agency like, say, the California Department of Health. Then any so-called homebrew test for LTT, which stands for laboratory to test the laboratory itself, develops or creates can be shipped based on the lab director's approval. Now, of course, to prepare copious reports on the sensitivity and specificity of the test to build the appropriate controls, all those things in that lab director has to have a doctoral degree and maintain mortification.


You know, this has to be a qualified person. But in general, overall, you have far more flexibility to ship an update of lab tests than a traditional medical device in a box. Now, the FDA hasn't like this because it's a path that's been outside them for more than a decade, contended they have something called enforcement discretion, meaning they've just chosen not to enforce the law against laboratory tests, but reserve the right to do so. And they've been seeking to bring lab tests under FDA's purview for so-called premarket approval, which means, among other things, huge time and cost of the test market.


And the FDA reviews every change you make to the test. If their software involved, which is for many modern assays, every committee would have to go through something called design control, which is like the sort of bureaucratic process around revisions to to a medical device. Now, you know, opinions differ on this, but many lab directors believe that adding premarket review on top of existing health care regulations really slows things down and adds a lot of paperwork doesn't have much value.


And so right now, those lab directors, usually they're regulated by this more decentralized system for Leptis approval cochlea. It's a system outside the FDA the FDA doesn't like. Now, once Health and Human Services declared a public health emergency over this virus, FDA actually gained a new power, the power to grant so-called emergency use authorizations. The idea is that it's an emergency. So an emergency is high risk and this year need lot more review for the tests.


And that means safety needs to slow things down.


And so it was the denial of these emergency use authorizations over the critical month of February that kind of forced all labs keep their hands tied. And basically all our planes were sitting on the ground while they got Pearl Harbor by the virus. And why did he do this? Because they saw an opportunity to win ground and bureaucratic turf war, to never let a crisis go to waste, to use the UHA to force lab directors to send in tests through them and FDA to believe that they're the only good actor in medicine and biomedicine, that anything outside them isn't the higher standard.


At a cultural level, they basically don't care about the cost or time they impose on others for this ostensible high standard. And nor do they think that significant premarket review and bureaucracy may actually reduce quality overall due to less ability to innovate and faster costs, which could lead to less competition, innovation and substantial delays. And that's an argument that's been going back and forth for many years. But at least now the American public is seeing that in this case, the cost of delays and contestable may be measured in needless hospitalizations and deaths, because under this emergency use authorization system, latticed approval was changed from the more decentralized system to a centralized.


And we're positioned themselves as a bottleneck for the UHA and no labs could get approval other than CDC itself. OK, and he was told yesterday and that basically meant for a country of 300 million people that test until today. That's essentially what happened with this case in Washington is interesting, which is a different example, decentralized, not diagnostics, but decentralized drug prescription. So typically, to get access to a drug, you have to go through an FDA process.


But there are other ways to do it. One way you so-called off label prescription where a drug that's approved for purpose, a doctor can prescribe it for a purpose, be it hasn't been approved by the FDA for that purpose. However, the drug company historically has not been able to tell the doctor, hey, you can prescribe it for privacy because that would be considered so-called off label marketing. And so that's a weird thing, right? A drug company cannot tell a doctor of a true fact because there'd be considered like promotion for a purpose if the FDA hadn't approved.


That was basically most very recently in twenty eighteen, I believe, rejected on free speech grounds in U.S. first kronur. And so like FDA power to kind of centralize drugs kind of dropped a little bit there. Another aspect which is quite relevant to us here is these right to try. Laws are passed recently and I think in about one states mean that when someone's at death's door, you no longer need to pursue compassionate use. Going through the FDA, you can get, I believe, a local state.


And the laws vary in different states, but a local state regulator to sign off on it and let it go through. And I believe that's what happened here with the prescription. So that's another example of the disclosure of diagnosis of drugs, is that clears that up, for example? I think it's clear.


I mean, I just don't understand a lot about that sort of world. I mean, it made sense as an explanation to me. I just don't know the sort of like the context.


Yeah. Going back to sort of like how should governments respond? Like, what is a decentralized response? What is it? What is it?


And then what on the individual level can we do?


So I think the single most important thing that the government should do is emergency expanded right. To try laws. And without do is, it would unblock, so we've already seen what happens with the diagnostics bottleneck, right, making it something where it had to go through FDA approval and also the CDC's definition, as if he was blocking tests that the physical level and the definition, which said only people who had returned from China are eligible for covid-19 testing that twin kind of centralization.


But like pretty bad decisions that were central point of failure that's led to the epidemic or really in the US as part of a pandemic that we have now. So. So when you can't yell at institutions like institutions that made that bad call, you can't bottleneck more stuff through them. Right. The alternative is do or one alternative, which I think we should do the most important thing we should do emergency spending right. To try, which gives everybody the ability to clear anything related to the coronavirus.


And so it's just approved the state level rather than being pushed up to the federal. There's several reasons to do this. First is to remove some bureaucracy, correct? That's right. And basically decentralize it. The reason I think this is a good idea is several fold. First, if this is pandemic, you know, when you have 50 states that are dealing with 50 epidemics, you just simply cannot bottleneck that through Washington, D.C. for approval. And everything just doesn't work.


Number one number two is this is already how lab tests are regulated and they've been pretty much fun. Number three, yes, it is possible that there are some things that are approved that don't work or even that are unsafe, which is less likely but possible. It's more likely that it's just kind of a placebo and doesn't work. But in this case, the cost of delay is so high that it's OK to have some false positives. It's OK because folks will die.


Otherwise, that's to say you'd rather have had an imperfect test. I mean, the exact numbers do matter, but rather have had an imperfect test during February and then have had essentially almost no testing because at least we would have realized the potential scale of it with an imperfect test. There's things you can do to mitigate. You can do if you know the false positive rate, you can sometimes do retesting. If the errors are random to test cancellations throughout the day, you can use those standards when necessary and do random sampling.


An imperfect test is better than nothing for something like this. Not always. Math matters, but but but frequently that makes it so that emergency can try to try and get drugs, diagnostics that are approved. States, I think, would unblock a lot of stuff. And then information sharing would happen because if Washington State or Massachusetts or California or Texas or whatever figured out something, then that could be propagated to other states and you wouldn't have to bother.


OK, what else should schools shut down?


Should government shut down? How do you like where? Where do you draw the line on this?


What's your opinion? Great question. So there's some math to draw the line on it. So essentially one way of doing it, it's not the only way of phrasing it, but one way of doing it is for given that, you know, a particular infection rate in the community, let's say it's you know, here's a good example. Let's say there's one hundred people in San Francisco who have the coronavirus and Terrence's populations, about eight hundred eighty thousand.


And you have a office that has a thousand people. What is the probability that one person in that office has the coronavirus? I have no idea.


But you can walk me through it. Yeah. So you can you can calculate that out. And it's about about 10 percent. So just to recap the numbers here, like one hundred people in San Francisco to have the virus population, about eight thousand one thousand person gathering reason, it's about 10 percent is even though the probability is very low. You have only one hundred people who have the virus out of eight hundred thousand enough. The number of trials you're doing and equals a thousand is very, very high, pretty high.


And so you start to take a low probability event. We try it a bunch of times and then eventually you potentially get a sprinter at the office. Right. And the reason you might say, oh, well, isn't it paranoid to say one person was the coronavirus, that the office will spread it to a bunch of people? Right. That's an excellent question. Yeah. And the issue is there's choke points, so almost everybody touches the elevator about it.


Right? Bathrooms, elevators, common areas like cafeteria doors, the fridge for food.


Right. If you had a security camera trained on it. Right, you would see those things that four hundred people touch. And that's probably how if people go in, these folks who've spread conferences and so on, if you go back and run the security cameras and you look at exactly how it happened, I would not be surprised if it was something like that. Someone washes their face in the in the restroom. They turn the hand back up, the hand back to the faucet handle back.


Right. And then 30 more people touch that. Right. That's a great way to get super spreading. So the answer to your question is we can use math to quantify this. You know, what's your risk? So just to go a little bit further, let's say there's now not one hundred people enough, but a thousand people in S.F. who have the virus out of eight hundred eighty thousand people. And there's, again, a thousand people in your office.


Now, it's probably announce more like, you know, more than 60 percent. Right. So, you know, I posted infographics with this kind of calculation, OK, so you can do it yourself if you want this, because here is that the second major thing that people should be doing is they should be stopping large events south by southwest Houston rodeo, large conferences with a large event.


Is it like more than people? Like is it an international event? Like is it an event, a local city event? Like what? Define a larger issue?


Good question. The answer is it depends upon the percentage infected in the people coming to the event, which we don't know. You can you don't know directly, but you can estimate right now if you do, the analysis of Trevor Bedford had a bunch of people and he was able to estimate a cluster of medium size like five hundred seventy. Right. So five or seven people in Washington have a case related to that cluster. There's other ways of estimating it.


Based on 10 deaths in Washington state, there's probably thousands of undiagnosed cases out there. I can get to the math of that. So this is what I call a Geiger counter, right? You can't just by feel tell whether a place is really radioactive or not using a Geiger counter. Right. And when you're dealing with an invisible threat, whether it's radiation or deadly viruses, you can't just kind of intuit your way to it and say, oh, that feels big to me or that feels small to me.


You have to do it on the basis of, OK, what's the estimated infection rate and then how likely it might get. Is there to be somebody at this that gets it? Plus also, this is very important. We have empirical evidence as this is not apparent. Did you see what I have been tweeting on the the parades in Philadelphia? Yeah. Yeah. So to have at least two examples, one from basically one hundred and one years ago, under two years, I was the Philadelphia St.


Louis bingo.


Exactly the contrast between how they responded to the Spanish flu or Philadelphia to sort of like went ahead with everything in St. Louis, sort of like shut everything down. And it affected Philadelphia way more than St. Louis, correct?


That's exactly right. That's right. And the same thing actually happened and move on for this virus two months ago. Right. Right. One had a gathering of forty thousand families, which is a a really big thing. And that's a great way to just take something that was a viral fire and just make it go completely vertical.


Hey, thank you so much. This was a fascinating conversation. Awesome. Cool. You can find show notes on this episode, as well as every other episode at F-stop blog slash podcast, if you find this episode valuable, shared on social media and leave a review to support the podcast, go to F-stop logged membership and join our learning community.


You'll get hand edited transcripts of all the podcasts and so much more. Thank you for listening.