Solani Daytoni (3:01)

Yeah, I mean my focus is global health and in that we have a similar problem. It's sort of different layers of problems. One is that a lot of data is just not actually collected at all. And what we have to do instead is extrapolate from what we have. So let's say for data on diagnoses of mental health conditions or chronic pain or things like that, we usually only have data from a few rich countries sometimes. And if you want to try to estimate how many people in, let's say, India or Nigeria have these conditions, we don't really have surveys or we don't have the data that's collected in those places for many conditions like that. And so what happens instead is that you use the data that's already collected in the rich countries, put it into a statistical model, use other information about demographics and how that differs between countries, and then just try to extrapolate, essentially. And the less data you have, the harder it is to make those estimates. And also these relationships between some demographics. And maybe rich people are more likely to have these conditions in one country. That might not be the case somewhere else. And I think we're sort of flying blind on that. The other problem is that even when we do have the data, it might just be a different type of data. We might have collected surveys in one place, but we have medical records in a different place, and these are completely different. Like, the types of people who might go to a hospital in India are not very representative of the whole population of India, for example. And so that really biases our understanding of a lot of diseases. But I totally agree that this is a problem across many fields, not just health.

Podcast Host (4:49)

And the. The thing that fascinates me, I'm thinking of the book the Weirdest People in the World. It goes across discipline, right? Like most of the psychology testing we have, 85, I think,%. Is done on Western people without any, you know, 85% of the world is ignored. And I remember thinking when I was preparing to talk to you, like, in. In matters of health, it's not funny, right? Like, in certain areas, you can go, oh, ha, ha, yeah. But, like, another thing that I've noticed is that a lot of the data that we do have that's been collected on actual humans, not mice.

Solani Daytoni (5:33)

Right, Right.

Podcast Host (5:34)

Is on men and not on women. And I always think of Semmelweis, the very famous example of making. Forcing the men to wash their hands before attending a mother about to give birth. And like, how. How do you. How do you solve for all this?

Solani Daytoni (5:50)

I mean, it's difficult. The. The way that often Researchers try to do this is they try to over sample certain populations. So let's say with a general survey, you, you reach out to, let's say a thousand people. If they don't pick up the phone, you'll call them again, you'll call them again. And what they do is for populations that typically don't answer these surveys, they'll try, they'll do bigger sample sizes, they'll try harder to contact them. And then you sort of, yeah, you sort of put in more effort so that you collect the same amount of data that you want to have. But that's quite, that's quite hard. And like it's, it's difficult to sometimes find the funding to do that or even know how to model these different population. Like how do we know which groups are being unrepresented in the data if they weren't in the data to begin with?

Podcast Host (6:41)

Right, exactly. And, and that also leads to a problem that we see in a variety of disciplines. The self selected sample. Right, right. It's like I go crazy when I see the. Every millionaire does this. Every successful entrepreneur. No, no, that's entirely wrong because look at the sample. And is there a solution?

Solani Daytoni (7:09)

I think that often the solution is to have either philanthropic or government run surveys where the goal is to create a data set that can be used by a lot of people from different fields. And the goal is to create something that's nationally representative. I think when it's private interest, it's, there's often this, you're only looking for a certain thing. You don't, you don't have to fulfill every purpose with that data. Whereas when it's being run by government, there might be 100,000 different uses for that data. And collecting that at a representative level is really important for that purpose.

Podcast Host (7:48)

And yet also the idea that you've got a variety of gatekeepers, and I know that that's a favorite topic of yours as well, and I want to spend some time on that. We might as well right now, I wonder if the gatekeepers, the peer review, all of that isn't actually disadvantaging people who might be doing what I call real science. And then there seems to be this pathway that is, let's make it sound really cool so the media will pick it up so that it'll get included in this congressional report. And when you keep digging down and down and down, you find that basically it helped mice.

Solani Daytoni (8:43)

Yeah, yeah. There's also this idea in science called the Matthew effect, that people who succeed once succeed again. And they're sort of chosen to do follow up projects. And in a sense you can see why that would be appealing because the best predictor of what someone's going to do might be what they've done so far. But it really disadvantages newcomers into a field or people with different perspectives. And I think that that makes it quite hard to diversify the field to other types of research.

Podcast Host (9:13)

And is one of the answers that, you know, like for example, we have a fellowship and grantee and we're increasingly getting really fascinating medical especially, and it's coming from groups who I don't think would be able to get into a peer reviewed journal. And I'm increasingly funding that type of thing because these are brilliant, brilliant people who are coming up with novel non consensus ways to look at things. And sure, maybe it's a moonshot, but like penicillin. Oh damn, I should have cleaned that up. Oh, wait, what?

Solani Daytoni (10:01)

Yeah, have you heard of this idea of science advances one funeral at a time?

Podcast Host (10:07)

Yes.

Solani Daytoni (10:07)

Max Planck, Yeah, it's sort of similar to that idea. There's this economist, Pierre Azulay, who tried to study whether that, that aphorism was actually true. And he did find that publications and citations increased from a lab once the lead scientists had died. And it's sort of a sad and depressing finding. But at the same time, I guess I can see how individual scientists who have a particular topic of interest might be so entrenched in that way of thinking or that type of approach that it takes them leaving the field for something new to happen. And that's especially the case in academia where there's this very narrowing funnel from doing your PhD to becoming a professor. A very small fraction of people make it to the end and that affects the research agenda in a field.

Podcast Host (11:06)

I was always fascinated by examples where people that I really admired, primarily physicists who, when you looked into it, like many of the episodes that happened with them were really like the movie Mean Girls. And, and I'm thinking specifically of Robert Oppenheimer when he was told that David Bohm, who was a brilliant physicist, was suspected of being a sympathizer for communists and was told by our government, suppress him. And so they, there's written records of him saying to his colleagues, if we cannot disprove David's thesis of hidden variables, we must ignore it. Does that, does that go on a lot? Is that those kind of rivalries and crazy, really petty types of behavior?

Solani Daytoni (12:09)

I mean, I would guess that that happens in many fields, if not all of them, because humans. Yeah, unfortunately, academics are humans just like everyone else. And I think it's often the way that you make change is change the incentive, sometimes change the structures. Occasionally I'll see really talented researchers leaving the field because they find it really hard to deal with that kind of reward structure of, you know, publications beyond everything else. Yeah, I, I, there are so many ways that you could improve that, but it's, it's a difficult problem.

Podcast Host (12:45)

Well, now I'm going to put you on the spot. Sure, let's improve it. Tell me how we could change the incentives, how we could change the process so that we got much better outcomes from all of the money that we're spending on the research and development of drugs and other therapies.

Solani Daytoni (13:05)

That's a very big question.

Podcast Host (13:06)

I know.

Solani Daytoni (13:07)

I think I'll probably start by talking about kind of the way that research is done in labs. So one of the reasons that I'm not a scientist anymore is because as a scientist you're kind of expected to do everything at once. You're expected to write papers, do the, figure out a research question, find the participants, find the animals, whatever to do the research on, sometimes write this programming code to do the analysis, present your findings, and then go through the whole conference and networking procedure. And all of that is just often one person or it's expected of people to be these all star scientists who can do everything. And I think it really slows people down because it's hard to keep up with the advances in each part of these, each of these different skill sets. And it's hard to do it all at once. If you're a young scientist trying to learn how to code, that might take a few years to get good at. There's a situation I often see where scientists are just learning the basics of how to code. They'll make mistakes in really tedious parts of the process and not know because they're not computer scientists. Like they're, they're supposed to be thinking about the research and the question. And so I think spreading that out between people, having people do different tasks and work together on science as a team instead of, you know, individuals, I think that can make a big difference. It's this idea of division of labor and science. You know, if you can have one person who's the software engineer, one person who's reads the literature, one person who writes and presents the findings, that kind of organizational structure could make things a lot faster. I used to work at an organization called Our World in Data and our structure was very similar to that. So I Did research and I did writing. But then I'd work with my colleagues who were pure data scientists. Their focus was trying to extract the data from these Messy sites or PDFs or dashboards and get it into a usable state for me to use and for everyone to be able to see on the website. And that's something that would take me years to figure out how to do. They would probably feel the same way about the writing and research process. But being able to work together on something like that means you can do multiple projects per year instead of just one big one that drags out. That's the big one that I would see. And then in terms of drug development, I think there's another problem, which is that the bottleneck is often clinical trials and how to test drugs and understand whether they work, how effective they are, how safe they are, and so on. Currently, that process takes an average of about a decade. I find that incredibly depressing. There are medical breakthroughs in the pipeline right now that work that won't get to patients for another decade from now. And obviously being able to test whether they work is really important. Otherwise we wouldn't know which drugs to prescribe. So that process really matters. But it could be done so much more efficiently than it is now. And that's something I've been thinking about a lot and just how to improve different parts of that process. Because currently it seems like there isn't really a group focused on doing that across the board. There are people who are patient advocates for certain diseases. There are pharmaceutical companies who just want to increase their margins. There are biotech companies who want to make it easier to get into the field, but they're not. There are very few people who are interested in the whole pipeline. And how do we make it easier across the board for people to volunteer into clinical trials? Or how do we make it easier to design clinical trials so that they're, you know, instead of testing one drug versus placebo, we can test five drugs in the same trial. That's very hard to find people to work on because it's. It's not. It's. It's something where you need to coordinate on a problem that's distributed between people, if that makes sense.

Podcast Host (17:23)

Yeah. And it's a logistical challenge.

Solani Daytoni (17:26)

Yeah, yeah.

Podcast Host (17:27)

Trying to get the various. But what do you think about things like innovation, like Claude code, that, uh, somebody joked that even an idiot like me could probably write code in cloud code? Uh, do you think those kind of innovations will help?

Solani Daytoni (17:42)

I think they will. I think they'll definitely help in some areas. What's hard though, is that human biology is really complicated and we just don't have the data that could be used to, like, the things that cloud code is often good at is where we have a lot of data. We have like writing collected on this topic or there's a bunch of code available online. We just don't have that for the human body. In the same way, the data is very fragmented between different hospitals, different diseases, the different types of measurement, they're all in different places. There are very few places where you can get information across the body on how a drug might interact with some organ. Even the data sets that do exist are generally on healthy volunteers. There's a big data set called UK Biobank, for example, and it tends to be highly educated, healthy people, like we were talking about before. And you wouldn't be able to see how a drug affects a particular system in their body. And we aren't collecting that data in data sets like that. Even if we were, we actually don't have the tools to collect them at the right level. So one thing that I learned recently that was incredibly fascinating was just how fast things happen in biological systems. So proteins, which are used across your body for all kinds of things, they're made by turning a gene into RNA and then protein, that sequence of protein folds into the right shape for it to do its functions. And the speed at which the folding happens is on the level of microseconds on average, and that's incredibly fast. And we don't have any way to capture that. And that kind of speed is also how fast enzymatic reactions happen, how fast collisions or interactions happen in cells. We don't have any way to capture things at that level and we can only really approximate them. And so there's so much knowledge that we actually don't even have. We don't know what is happening at that granular level to be able to predict what is going to happen in a system. And you can make some rough approximations, but that still means there's a lot of uncertainty.

Podcast Host (20:04)

Yeah. When you mentioned the Matthew effect, I remember reading a scholar who made the reference to our immune systems as basically following the Matthew effect. And I found that really fascinating. I guess I could see that basically his point of view was that our cells all have the suicide switch, which gets flipped if they're not being useful, and that when you watch this way, for example, when we have a novel virus like the coronavirus, you see the immunes, if you were going to animate it you would see a couple of cells, like taking a punch and then they're not working. But then when one got through all, the way it moved was towards that, which suggests power laws to me. And it does seem like our bodies are complex adaptive systems and that we might be able to glean some useful information by looking at how complex adaptive systems function in general.

Solani Daytoni (21:11)

Right.

Podcast Host (21:11)

But I'm fascinated by the bottleneck. How would you go about, given all the tools that we have, and I asked you to design a new system and said, we're going to underwrite it, we're going to make it as ubiquitous as we can, what does that system look like?

Solani Daytoni (21:32)

It's interesting because I'm often thinking about incremental changes that we can do within the system that we have. And redesigning the whole thing is much harder because it depends a lot on what balance of risks and benefits people are willing to make. And different countries also do this differently, and it's hard to think of how to do it from scratch. I think that the thing that I would keep in mind is it really does matter to know how effective drugs are, and if we want to do anything with them, we do have to have some way of testing that. And that can't happen after the drug is already available because it becomes really hard to test the difference between having it or not. Before it's available, it's easier to randomize it to certain people and not to others. And that allows you to tell a difference without, you know, while excluding for other confounders. Right. So that somehow needs to be in the process, and probably in some way you need to have some scaling up where you start out small with some easy experiments that are cheap. I don't think that doing them in animals is a great idea because they don't necessarily translate very well to humans. But also like the ethics of doing animal testing, I think, are bigger than people imagine. If there's some way to do very small initial trials, just see what happens with patients who are more willing to. Or have severe conditions. They're willing to try experimental drugs. If there's a way to do that at a small scale, learn from what works, scale that up into larger trials, and run those trials much more efficiently. I think that's. That's the sort of process that I would like to see. And the one thing that I would mention is the, the way that we're running clinical trials right now, it's like everyone is just doing their own version of it, Right. Like every pharmaceutical company who is running a trial is doing their own trial for one drug, comparing it to placebo. And there are people who have designed better ways of doing this. So like I mentioned, instead of testing one drug at a time, you could do a larger trial where you're testing five drugs versus placebo at the same time. Or let's say it's just a whole population and within a certain condition that you have, we have different drugs available. We don't know which one works better. You just get randomized to the ones that you might be prescribed anyway. And that would help us understand which drugs work better than others if this was happening at a larger scale. So each individual company just gets to adapt a module of this trial instead of developing their own one each time. I think that would make the model process a lot more efficient.

Podcast Host (24:28)

And, and what do you think about. I know that there are benefits and detriments to AI. My cousin is married to a guy who's a medical doctor, but also a PhD in computational science. Ended up working for McKinsey and working primarily for pharmaceutical companies. And what they would do is go into the drugs that didn't work and use AI back when it was called machine learning. And what they discovered was really interesting. For example, he was telling me about one drug that showed no efficacy at all. It was for a particular condition. I don't remember what it was, but overwhelmingly female. And after running it through all the machine learning and everything, he came back to the pharmaceutical company and said, actually this drug will be incredibly efficacious in postmenopausal women who are slightly overweight.

Solani Daytoni (25:32)

Okay.

Podcast Host (25:33)

And when he told me that, I'm like, are you sure? That seems like really cherry picked to me. And I worry about. And I never followed up with him to see whether they released it and tried it in trials. But what about the ability of using artificial intelligence intelligently?

Solani Daytoni (25:58)

I mean, I think there's a lot of. It is just like you said. I think there is some grain of truth in that. A lot of times when we don't know how to treat a condition yet, the way that people go about it is trial and error. They just see what works. Just try a bunch of hundreds or maybe thousands of different compounds. Compounds. See if they have any effects. Sometimes in the lab, in cells or in small scale trials. And I think that is a legitimate way to, you know, just spread your risks if you don't know which hypothesis is correct. And we've managed to find a lot of really important drugs through that process. So azt, the first HIV drug, that's Right. Came from this trial and error process. It was originally meant to be a cancer drug, didn't work as that, and then was discovered as being an important HIV drug. There are others where people thought that it might work for certain things, and then it turned out to work for something else as well. So there's this new schizophrenia drug that was approved last year, I think, called Carxt. I think it has a different brand name, but that was initially tested for Alzheimer's disease. And they found that it seemed to specifically reduce the hallucinations that some patients would have. And it was sort of forgotten for a while because it also had these nasty side effects of diarrhea and vomiting. And eventually scientists figured out a way to combine that pill with a different pill that prevented this to ameliorate. Yeah, yeah. And now it's. Now it's a very important new type of schizophrenia drug with a different mechanism than most of them do. And that's something that wouldn't have been found through this process of trial and error. And I can totally see how AI would be really helpful in going across this whole library of drug compounds that we have and different diseases that we have and trying to find potential matches. There actually there were a few Covid drugs as well that I think were found through this method of just. Let's just screen thousands of different drugs and see what works in the lab. What's difficult, though, is this cherry picking, as you mentioned, it's hard when you run so many tests. Some of them are just going to work by chance in the lab and they won't work in humans. You still do need this process of let's actually follow up, make sure that it's not just a false positive, run a trial on humans to see if it actually works. And so I think that's why biology and clinical trials are still going to be the bottleneck. Even if we have AI to massively speed up this pattern recognition of trying to find potential drugs, that last step of being able to predict will this actually work in humans is still really hard.

Podcast Host (28:53)

Yeah. And that seems to me to be the kind of ultimate bottleneck here. You know, you mentioned that there are drugs right now that you're pretty certain, you have a high degree of. You, you've established a high probability that they work and they're not going to be around for like 10 years. Give me a couple of examples.

Solani Daytoni (29:14)

So I would. I don't know specific ones. I just think probabilistically, there are definitely ones in the pipeline right now. And the way that I would say that is because there are a lot of drugs that have only recently been approved that were in the pipeline for years or decades. And just knowing that that process hasn't changed very much, that's still likely to be true. So, for example, the malaria vaccine approved three years ago, I think it was developed in the 90s.

Podcast Host (29:47)

Now, I did not know that.

Solani Daytoni (29:49)

It really shocked me when I learned that. I mean, like, half a million kids die from malaria every year.

Podcast Host (29:56)

I'm a big fan of Africa and.

Solani Daytoni (29:58)

Right. It's. It's a huge problem. And yet the problem was funding and clinical trials. So the. The vaccine was developed in the 90s by, I think, researchers who were initially at the Walter Reed Army Medical Research Institute. And because it's a disease that mostly affects the poor, there's no commercial incentive to develop that vaccine. Like, if a drug company produces it, they're not going to earn a profit. It's essentially an act of goodwill that they develop it at all. And so you needed philanthropic or government funding to test this drug at all and to scale it up. And that came in the early 2000s for initial field trials. They did those initial field trials, I think, late 90s, early 2000s, but it was still, like, it wasn't an incredibly effective vaccine. It has an efficacy of about 40%. That's how much it reduces malaria. Still pretty important for one of the biggest diseases worldwide.

Podcast Host (30:57)

Totally.

Solani Daytoni (30:58)

But there was this question of, could we get something better? I don't know. And is it worth taking this through larger trials? The researchers who worked on this malaria vaccine struggled at every step of the process to get funding to continue testing. That took another. I think it was only in 2015 that they finished doing tests for this vaccine. And part of that process even involved the researchers themselves trying to find funding to build clinics in Africa because there weren't enough clinics to actually run the trials at all. Like, if you have a drug or a vaccine for a disease that doesn't affect us here, you have to be able to develop the tests or the hospital, sometimes distribution. Yeah. Hire the nurses and the doctors who will run these trials on the ground. That process took such a long time. I found that incredibly depressing. But it's also the sign that if we fix this, there are actually so many opportunities, so many medical innovations that we could. Fine if. As long as we fix this pipeline. The pipeline. The bottleneck is not necessarily scientific difficulty. It's not that things are impossible to develop. I think sometimes it's economic incentives or it's just the process of how we're testing drugs that is stopping us from developing treatments and cures for some diseases.

Podcast Host (32:28)

And why do you think there seems to be this incredible reluctance to allow humans who have all of their faculties, they're not like Alzheimer's and they're not mentally impaired in any way. Why are we not letting them put themselves in trials of their own free will?

Solani Daytoni (32:50)

It's both a question of the risk aversion, and I think for a company, you wouldn't necessarily want things to go wrong. And for that to become known, you do want to have certain. You want to pass certain thresholds before you feel safe, like, safe enough to test this in a human population. And so there's. Part of it is that reputational risk. I think the other part is just often the difficulty is not necessarily whether people want to participate, but that it's just difficult for them to. Like, it's difficult to find time to go into a clinic, you know, every two weeks for three hours or something. And that's often why a lot of people who participate in research are university students or people who are unemployed or they have time. And it's not just people with those conditions, but it's people who also have the time and don't have alternatives that I think we sort of need to make it easier and more appealing to be part of these trials, like having health coverage or making it simpler paying people to participate or making it part of the general process of getting treatment. If there are three different drugs that your doctor could prescribe and they don't know which one, and all of them seem roughly the same, that should just be a clinical trial instead. And if we randomize those drugs, then we'll actually learn information that could help people in the future.

Podcast Host (34:27)

We got a proposal for our fellowship program from somebody who is trying to design a system where people can participate in clinical trials from their homes.

Solani Daytoni (34:37)

I'm a big fan of that idea.

Podcast Host (34:38)

When I saw the. You know, because we get thousands of applications, and I saw that. That pull quote, and I'm like, why aren't we doing that now? And is it a sense of the inherent nature of. A lot of things are life and death? And obviously reputational damage comes in all of those things, those negative things. But the precautionary principle taken to an extreme destroys society, it destroys innovation, it destroys all of those things. And there's gotta be a balance where you're. You're not being reckless and you're not saying, well, we'll just give it to everybody. We'll put fluoride in the water and just see what happens?

Solani Daytoni (35:29)

Yeah.

Podcast Host (35:33)

Is part of that the coming back to the media, you know, the focus of attention?

Solani Daytoni (35:40)

Right.

Podcast Host (35:40)

It seems to me that the media have the steering wheel, so to speak. And I don't think I'm being too out of left field here. I don't know that they're steering all that well.

Solani Daytoni (35:53)

Well, especially when it comes to science journalism, I think that it's just bad

Podcast Host (35:59)

and like, how do we make it better? There's so many smart people who can also write well and communicate well.

Solani Daytoni (36:08)

Yeah. I mean, this is the reason that I started writing, essentially.

Podcast Host (36:13)

I was going to say, other than you.

Solani Daytoni (36:16)

It just seemed to me like there was this huge gap between what scientists knew or, you know, what we understand versus what people were aware of from reading the news, especially in science. This idea of all of these kind of clickbait headlines of this drug worked in mice or this drug worked in a rabbit and therefore cancer is cured. And it just leaves me really depressed because there are genuinely big breakthroughs happening that people haven't heard about. And it's these other flashy headlines that they read instead. Or, you know, chocolate is going to give you cancer, and the next day you read chocolate protects you from heart disease, and you have no idea what to really think. I don't know what the reason for it is. I think partly I wonder if it's expertise thing that a lot of science journalists don't have a background in science. They have a background in journalism and they don't necessarily know how to read the literature. They don't know how to evaluate what's good and bad research. You could also blame at least part of it on the audience. Like people are clicking on the headlines that seem the flashiest. But I think it's also this thing that a lot of media in science journalism doesn't really treat the reader as an adult. It doesn't, you know, the way that I write, I sort of don't expect people to have any background knowledge in biology or health, but I do expect them to be interested as long as I keep it at their level and bring them up to speed on this issue. I sort of have this view that anyone can love science as long as they understand it. And if I can bring you from the basics to something really in depth, then that's my goal has been achieved. And yeah, I just think that there are so many interesting things out there that people are just not aware of because they haven't been explained it to in a. In a easy, you know in an accessible way that is engaging and interesting. And that's, that's part of the reason that I started writing and started the magazine works in progress.

Podcast Host (38:21)

And, and it seems to me that, you know, I guess I used to have a friend who would say when people would ask him about business, everything is sales and marketing.

Solani Daytoni (38:36)

Yeah.

Podcast Host (38:36)

And in a way there is some truth to that in this field as well. Right. Like, I totally get if you are going to maximize, if you're trying to maximize the objective function for clicks.

Solani Daytoni (38:52)

Right.

Podcast Host (38:53)

You're going to be led to a particular type of story. And if you're trying to maximize it for understanding, you're going to be led to a completely different type of story. I wonder, is there an example of something that is truly an incredibly innovative drug or process or procedure that history will look at and say that was kind of like hand washing and antibiotics that's happening right now, that nobody right now is just meh.

Solani Daytoni (39:28)

Yeah, there's actually a lot, there are a lot that are just on the cusp of either becoming available or like some versions of them have become available. I'll tell you about two. They're both types of drugs that are not pills, but they last really long in the body. So one intramuscular shot or something like that lasts for months or maybe a year with a single drug that's very potent, very effective, very safe. And that completely changes the way that people get treatment, get prescriptions. It makes it so much easier to take treatments for a chronic disease, for example. So one of them is sort of in the sphere of HIV drugs. So PrEP, which is the way that people prevent infections by taking oral pills every day, usually that is increasingly being replaced by a new type of treatment that is long acting antiviral. So it's an injection into the stomach usually I think that lasts about six months, protects people to nearly 100% efficacy. It's more effective than the daily pills because people forget to take the daily pills. But also that particular drug just seems to be incredibly effective. And that has been a breakthrough on types like drug chemistry, like formulating drugs. The way that it works is that it forms a little depot in your stomach and then slowly diffuses out of that over a period of months. And even very tiny concentrations are enough to prevent the HIV virus from infecting your cells. And that is a huge revolution that I think people are sleeping on and aren't really aware of. So it's not even just that one drug. We've developed lots of different new versions of doing this for other conditions as well. So contraceptives, for example, are another example where it lasts very long. But we've now figured out ways to do this for multiple different drugs. You, you either have the drug in a little oil droplet or in some kind of sugar or something, and eventually over a long period of time, it dissolves into the body. There's another type which is called Sirna, I don't know if you've heard of that. It's sort of like the cousin of mRNA. And what that does is it's a small molecule of RNA that can go into your cells and silence a particular gene from producing its protein. So with Alzheimer's disease, for example, one protein, ApoE4, seems to create a much higher risk of developing Alzheimer's disease. There are liver conditions and there are things like high cholesterol, where often it's just this one protein. It might not be that the whole disease is controlled by that one protein, but that one protein might be a key part of a pathway. It might be the bottleneck in some way to that disease process. And if you're able to silence or slow down the production of that protein, that could make a huge difference to the development of that disease. And so these new drugs, essentially, they get into your cells, they find that specific gene and they silence it. And it's like, shut up, you're not going to produce this protein anymore. And the way that they're formulated is that when they enter a cell, they get immediately trapped in a little bubble like thing and occasionally one of them seeps out, one of them leaks out and it's able to silence that gene. And because that happens so infrequently, the effect lasts a long time, like months, sometimes years. And scientists have developed a bunch of new drugs, mostly for liver conditions so far that use this method and they're incredibly effective and incredibly long lasting. So there are new cholesterol drugs like that that reduce cholesterol levels by like 60, 70% with this one injection that lasts for four or five, six months.

Podcast Host (43:33)

And are those commercially available today?

Solani Daytoni (43:36)

I think there are two that are available already. There are a lot more that are in the pipeline that will be in the next few years.

Podcast Host (43:43)

I was reading about one that just specifically limits LP and it has a huge, I mean it's like 90 plus percent reduction in LP.

Solani Daytoni (43:55)

Yeah, exactly. So that's another one that's still in the pipeline right now. I'm guessing it will be approved this year or next, but yeah, Reduces lipoprotein A by 95%. Or more really extraordinary. Right. Uh, in that case, it's the, the whole, that specific type of LDL cholesterol is almost entirely determined by that one gene. And so if you can silence that one gene, you can make a huge impact on the development of that, of that type of cholesterol.

Podcast Host (44:28)

And I, I'm very excited about all of this stuff because like the original statins, I had high cholesterol and way back when I said, no, I'm going to do this slow release niacin and oat bran muffins and got it way down. But back then, statins had pretty bad side effects. And, and how much of that hangover do people have? Right. Like hearing this and reading that thing that I read about that I'm like, I, I can't believe that we've made this much progress. How much of the old problems. Let's just stay on statins for a minute. Original statins, not, not the greatest thing in the world. What we have today, like put it in the water.

Solani Daytoni (45:20)

Yeah, I, I think it's, it's often this earlier drugs, often, especially when they're taken orally, they tend to have digestive side effects like nausea and vomiting, diarrhea. What's different about these drugs is that they're injected into the muscle or just under the, underneath the skin. And that means they don't have those digestive side effects. You still do have to be careful that they don't also have other effects that you don't know about. So that gene that is harmful for one disease might be beneficial for some other part of your body. And trying to specifically target the treatment to only reach your liver, for example. That's been the difficulty that scientists are having right now in developing new drugs for other conditions. But it's sort of crazy to think you could find a gene that's responsible for a disease and eventually we'll be able to target these incredibly precisely and have such large effects.

Podcast Host (46:20)

And I kind of think that 100 years hence we might be looked at like we look at like the barbers who bled people. Am I overly optimistic about what we might be able to achieve in the next longer period of time, like 100 years?

Solani Daytoni (46:44)

I think scientifically, I often think about that because I write and read a lot about the history of science and medicine, and it's shocking to me just how much people didn't have 50 years ago, 100 years ago.

Podcast Host (46:57)

It is wild.

Solani Daytoni (46:58)

It's, it's wild to think, you know, people didn't even know what the structure of DNA was until the 1950s. They had no antibiotics until the 1920s. They didn't have statins until the 1980s. Like the one that really blows my mind is CPR was invented in 1960.

Podcast Host (47:17)

Yeah, it was, it was like in the 60s.

Solani Daytoni (47:19)

Right. And it's, it's just crazy to think, think what was happening before that. Like, people might have some idea of, I don't know, change the person's position if they had a heart attack, but generally they wouldn't have any idea what to do. That to me is. And that's not even a medical, that's not a drug, it's like a procedure. And that's true for so many other things that we really underrate how little we understood in the past. And that doesn't mean that, you know, people were stupid. They were trying really hard, but it was, it was, it was hard to figure out these things on the frontier without the tools sometimes or to make progress. It took just so long. I think that's why I'm sort of optimistic in the sense that I think we're going to make a lot of progress on diseases that people currently think are untreatable. But the difficulty is often still going to be the financing or like, how do we get this to people who need it and how do we get it through clinical trials? How do we make it commercially viable for companies to develop this drug with these like long duration drugs? That I think is still an open question because they change the whole price, drug pricing, insurance thing massively. If you're, if you have one drug that you only need to take once, it lasts for a year or two years, the pricing of that is going to be very different from a pill. It's like much higher upfront costs. Should you do a subscription model or should you do something else? And I think we haven't really worked that out at all. And this isn't also like vaccines because not everyone needs to take these drugs. It's not something that the government can just pay for everyone or subsidize for. So I think that that kind of question is going to be the bottleneck.

Podcast Host (49:08)

Yeah. And the cynic in me says that if they could find the right pricing mechanism, man, there might be all of a sudden. Yep, all you have to do is get the shot once again every year you're good to go. That'll be $10,000.

Solani Daytoni (49:28)

I mean there's, there's also, there are these new gene editing technologies, for example, that essentially cure like sickle cell disease or blindness or deafness that are Caused by individual. Their pricing is just enormous. Some of them are like $3 million for a one off treatment. And while you might be able to okay that if you spread out the pills over a whole lifetime, maybe it would cost that much or you spread out the dialysis or something. But how is someone going to afford all of that to be paid at once at the start when they're young? It's just a hard thing to solve, I think. And it's that pricing that people need to work out.

Podcast Host (50:10)

And it also is part of the third rail. Right. Because if you really wanted to get the drugs into the system, you would accept that only rich folk were going to be able to use them and experiment on them. But I think that the, you know, the, the whole class of orphan drugs that have no market. Right, right. Well, they have a market for the people who have that particular disease. And it's, it is interesting to me because that's my history is in asset management. And it's, it's a market failure and there's just no other way to look at it. It is a market failure where you have something that will cure a not insignificant group of people and you don't put it on the market because that group of people can't pay for it.

Solani Daytoni (51:12)

Right. I have a few ideas of how this might be solved and it's still very difficult, but there are certain ways that we've tried already. So with the orphan drugs, for example, there's this idea of doing a priority voucher. Essentially a drug company that develops a drug for a neglected disease or a rare disease will get a voucher. And that voucher allows them to, for any other drug that they have, they can move that to the front of the queue on being reviewed by the fda, for example. That's smart. And so it becomes then a huge commercial incentive to develop this drug and then use the voucher to speed up the process for another much more commercially viable drug. And that helps to develop those drugs in the first place. I think what's hard about it though is that it's really zero sum. You're really just shifting around the prioritization within this queue that already exists. That's quite hard. And it also means that once you've developed the drug, you might not have any incentive to actually manufacture it afterwards. So there's another idea called an advanced market commitments. And what happens there is instead of, well, instead of funding the development of the drug or individual drugs or paying at the end, what you do is you have, let's say governments or philanthropies, agree to pool some amount of money. And they say, if you develop a drug that meets these criteria, then we will pay you out of this fund and we'll pay you based on the volume that you manufacture. And that actually gets to people. And what that means is, firstly, you're not betting on what will succeed. You don't know which drugs will make it to the end, and you're letting the drug developers take their own risks on that side, but you're signaling that if you produce this, there will be a market for it and we will pay you this amount. And often that is enough to develop, to get drugs to the finish line. Because companies know that, oh, there is actually this reward at the end, that if we produce this, it's not going to go away, they've made commitments. And secondly, it means that you're incentivizing not just drugs to reach that threshold, but drugs that patients actually want, because the only way that they get paid out of it is based on how many doses are actually administered to people. So this type of approach has been used for some vaccines, like the pneumococcal vaccine for type of pneumonia that bacterial pneumonia. And rich countries had already, like, there were already vaccines for the strains that affected rich countries. But Africa has other strains of this bacteria, and there weren't any pneumococcal vaccines that targeted those strains. In the 2000 and tens, I think seven countries and the Gates foundation came together and they put together this pool of funding and this advanced market commitment. And they said, if any pharmaceutical company that can develop a vaccine that meets these standards of effectiveness and safety, if you develop that, we will pay you this amount per dose that is administered to kids. And so that was effectively a subsidy for creating this vaccine. It was sort of a signal if you develop this, there will be a market for you and we'll pay you to scale it up. And what that meant was that millions of kids got these vaccines that wouldn't have been developed otherwise and also that were scaled up much faster than other vaccines have been because the average vaccine that is used in Africa or South Asia, there's. There's really very little incentive to manufacture that at scale. And the prices that governments are paying are usually at this, like, not for profit level. Right? Companies are not expected to get a profit. They don't have any incentive if you don't give it to them. And I think it's. It's important to make that, to sort of fill that gap with how do we make this actually profitable. How do we make it worth it for companies to develop this and in a way that actually reaches people that, like, last barrier of making innovations that people can actually use, I think is underrated.

Podcast Host (55:42)

Yeah. And I would think that in countries like America, you could also add to that mix tax incentives that would be favorable for the company to pursue an orphan drug or one with a limited market market that accountants could find all sorts of ways to abuse.

Solani Daytoni (56:02)

Yeah, yeah. And I mean, the other problem is that a lot of these diseases they have, let's say, you know, something that affects one in a million people or one in 100,000 or 5 million, and that's each of these diseases has such a small market for it, and it's hard to find researchers who are invested in developing a drug for that particular disease. But collectively, there are so many diseases like this, I think it's estimated like 5 to 7% of the population has a rare disease. That's a lot of people.

Podcast Host (56:36)

That's huge.

Solani Daytoni (56:37)

And if there's some way to make some kind of mechanism that would work for a lot of those diseases at once that might help solve the problem. Like, let's say some company developed a gene editing platform or this Sirna concept, and they have to develop the overall system or the mechanism for the drug, but then they can swap out the specifics of which gene it targets or which organ it targets. And if they can do that based on the specific rare disease that someone has, that would be much more viable, I think. And trying to find ways to approve drugs that do approve platforms or mechanisms that do that. That, I think is the way to unlock treatments for rare diseases.

Podcast Host (57:28)

And delivery systems, as you mentioned earlier, are also incredibly important. It seems to me that those types of things which are outside of the drug, but the drug needs them to work. Right. I think that there's got to be some way that that type of process could also be made to be more pursuable.

Solani Daytoni (57:55)

Yeah, yeah. That reminds me of. I don't know if you know, but there's been like an enormous amount of progress in treating childhood leukemia for a similar reason. So in the. Before the 1970s, the survival rate for. For someone with childhood leukemia was about like 15% or so would survive more than five years. Now that figure is around 80 to 90%. And that is a huge change in the last, like, 50 years. And the reason is not. I mean, some of the reason is new medical innovation, but a lot more of it is about the actual treatment regimen, like how that works. So which order to prescribe the drugs and what doses to give. And what was really hard about that as well, in the same way, was that individual patients with childhood leukemia are fairly rare and it's hard to run trials for them. Like, one particular hospital might not see enough patients to run a trial. And it's very hard to find all the right participants to run a trial at all, like across the country. And so what researchers did, and I think this was led by doctors and researchers at Boston's Children Hospital. And what they did was they created this clinical trial network across the US and later Canada and Europe, where they tried to find patients with leukemia children essentially across the country and enroll them into the same trials. And that meant that instead of seeing like one or two patients with this, you would have hundreds of patients with these conditions that can be participants in this trial for otherwise very fatal disease. And just testing out which regimens are going to work better for which patients is like being able to decide that these are patients who have a high risk of relapse or who have a high risk of side effects from that drug. And tailoring the treatment based on that would have been incredibly hard without doing a much larger trial across different sites. And from what I understand, that has been the big driver in like improving these survival rates. And it's just developing these treatment regimens based on much larger studies. And I think it's incredible.

Podcast Host (60:21)

Yeah. And again, in my adjacent quantitative work in finance, why is it so difficult to get people to understand the difference between a large sample and a small sample?

Solani Daytoni (60:35)

Right.

Podcast Host (60:37)

Because it baffles me.

Solani Daytoni (60:38)

Honestly, it's so important. You can't understand, you can't tell the difference between noise and what is really in effect without a large enough sample. And that might also depend on how effective the drugs are. Something that is extremely effective, you can tell the difference between two small groups, but something that is only moderately effective, you need a much larger group. Or if it's a rare condition or something that only happens rarely, you will need a large sample size to be able to see the difference between these two groups. And the fact that we're doing all of these different trials fragmentedly, it means that all of that becomes a lot harder.

Podcast Host (61:20)

So if, if I hate czars, but I'm going to make you one anyway. So. But only for a day.

Solani Daytoni (61:27)

Okay.

Podcast Host (61:29)

Like, what are the three probably highest leverage things that we could do? And let's make them incremental. They don't require a huge investment or a switching up of the infrastructure. But what are three High leverage things that we could do fairly easily that would improve outcomes at least noticeably.

Solani Daytoni (62:01)

I've thought of one. Hopefully I'll think of the other two while I'm describing the first one. The first, I think, is make it really easy for people to participate in clinical trials. And the way to do that, I think, could be just people who are patients seeing doctors for any condition. If there's a way for them to just show their interest in being contacted by a clinical trial researcher later on, maybe that's just a checklist in their usual form or something like that. If that could be done across the country, clinical trial researchers can easily follow up and they can see in some sort of secure, private way, here are some patients that have these diseases that are the ones that we're investigating, and these ones are interested in participating in clinical trials. Even just something simple like that would make it so much easier for researchers to just find participants for their trials, because I think that there's actually so much interest and it's just hard to match patients to trial researchers. They don't know where they're happening, and it's currently very hard to do that. Another one that I would suggest is either federally funded or just some sort of coalition of people in different doing, developing different drugs, run clinical trials together. And that doesn't have to happen in the same way for each of the like it. You don't have to start testing each of the drugs at the same time. You can actually do this sort of relatively flexibly. But just deciding on one protocol and saying, regardless of which drug reaches this threshold, we will test them all in the same way. We're going to look at the same outcomes, maybe whether they have a progression in their tumor or something like that. We're just going to run them all in the same trial, start the trial in the same way, recruit from the same hospitals, but instead of just doing this for one drug, we do this for five or seven or something like. Like that. And this has been done in the past, but it's just been hard to set the stage for it, hard to coordinate it. So during the pandemic, there was this big trial which did this called the recovery trial, where they tested, I think, more than a dozen drugs in the same trial in two years. And that made it. That meant that you could see the effects of. You don't know which ones are going to succeed at the outset, but while you're running this trial, you can compare the different drugs against each other and you can compare them to just one control group you don't have to recruit five different control groups just to test five different drugs. And it just makes the whole process a lot more efficient and simple. So that's a second one that I would suggest. What's the third one? A third one maybe a little bit harder, but we already do have the data to do it is actually having a platform that people deposit data into once they've run their trial and that other researchers can then reuse.

Podcast Host (65:11)

I think that's. I think that we're not doing that. Yeah, just blows my mind.

Solani Daytoni (65:17)

We are doing that in small fields and we're not doing that across the board. And it's very hard to access that data right right now. But currently, if you wanted to try to understand, has this been done before, what were the results in a different trial? Or you want to see across multiple, like 10 or 15 trials, what were the characteristics of those trials that made them more efficient or meant that the recruitment process was faster? We don't have a way to do that right now. And just having some kind of environment, it. It has to be sort of secure and private so you're not able to see the patient's individual data. But if there's some way to have a platform where other researchers can just learn from trials that have already been done, that would save them a lot of time.

Podcast Host (66:07)

I think all of those are really good actionable ideas. The third one, though, it just blows my mind that we are not doing that. Like, we have an AI division at o' Shaughnessy Ventures, and one of the things that we first thought that we would do when we, when we're at scale would be, wouldn't it be cool to have AI just generate null hypotheses? Because no one likes to learn via negativia. No one wants to write a research grant saying, I suspect that I'm going to reach a null set here and prove that this doesn't work. And yet there is so much information that you can glean via negativia. And so our idea was we're just going to have the AI just generate hypotheses after hypotheses, sent them to a central database that everyone can access and

Solani Daytoni (67:09)

like, duh, yeah, that's very cool. It's wild to me as well. These negative results are actually extremely helpful in helping you, one, reprioritize, do things that don't redundantly run the same types of trials that other people have been running and failing at. But also they will often help you understand what exactly is going wrong. If five Trials succeed and one fails, it's helpful to know why that failed. And if we don't publish those negative results or we don't have a way to reanalyze them, that becomes much harder. A lot of the. I feel like a lot of advances in the history of science have been people trying and failing dozens or hundreds of times, and those failures are really important. And knowing this didn't work in rabbits, but it worked in dogs. Why is that? And just following up on, on those successes and failures. And I mean, this is another problem that I think that academia has, which is that there is this bias towards positive findings. And that means that people can't learn from what didn't work and they can't understand why those things didn't work, and they can't reprioritize their research in the same way as they would if that information was available.

Podcast Host (68:28)

Yeah, and, and you know, in quantitative research, a lot of stuff doesn't work. And so. But we kept a research graveyard because we learned a tremendous amount from what didn't work. And it seems to me that that is applicable across discipline and, and our reluctance to learn that way. I guess it's just part of human os.

Solani Daytoni (68:55)

Yeah.

Podcast Host (68:56)

You know, everyone wants the positive result. You know, they. Well, I'm the one who came up with this breakthrough, but the only way, in my opinion, that you, you get there in a variety of fields is you make a lot of mistakes. And the mistakes is where the learning is. And it just baffles me why we are so opposed to that particular type of thing.

Solani Daytoni (69:23)

Yeah, I mean, I've, I've often also seen and heard this from scientists where, you know, if their hypotheses failed, they sort of see it as like a personal failure. Oh, I didn't, I didn't figure out what it was going to turn out to be. And, you know, it's very strange and it, it's sort of sad in a way. And I think that it's not necessarily their fault. It's this system where we reward things that succeed and don't reward things that fail for, you know, actually telling us something about what doesn't work. If there's a way to give scientists credit based on just the methods that they're using and not the results that they're getting and have some way to easily store these null or failed results, I think people really undervalue that.

Podcast Host (70:11)

And it seems to me to be like one of the easiest arbitrages available to really increase not only our knowledge, but our processes and Our tests and everything that we do. Right. And so I just find it kind of inconceivable that we haven't done that yet.

Solani Daytoni (70:32)

Yeah, I mean, it's also. This is coming back to the problem with journalism. I mean, I can totally understand from, like, an editor's perspective, you don't want to publish stuff that says, and guess what? This drug didn't work either. And, like, I totally understand that, but there has to be. Be some place to put that. Put those results that people can learn from, because.

Podcast Host (70:56)

But yeah, of course, you're right about that. You know, the cat sat on a mat is not a story. The cat sat on the dog's mat is a story. So we're drawn to stories, we're drawn to conflict. But the. It just seems so overwhelmingly valuable and that we're not doing it. And then ultimately the story you pitch the editor is, hey, it was only because we learned how we failed here, here, here, and here that we got here. Then you got a story.

Solani Daytoni (71:33)

Yeah, right. Yeah.

Podcast Host (71:35)

And so it seems a little bit a failure of imagination on the people doing the research.

Solani Daytoni (71:42)

Yeah.

Podcast Host (71:43)

Now, of course, as you mentioned earlier, you can't be all things, but that's why I very much like your idea. I'm a huge believer in cognitive diversity because, you know, there's no way to get a guy to. To come up with a list of things that would never occur to him. Right. You're going to fail.

Solani Daytoni (72:04)

Yeah.

Podcast Host (72:05)

But putting together very, very different skill sets, different ways of approaching a problem, of thinking about that problem, I think that that could be incredibly useful and beneficial. So I love the idea of, like, the team approach, but it also seems to me that does the entire scientific process, because the other thing that journalists do that just drive me crazy, and I understand that it's driven by people with a particular point of view that they're trying to. To advance. But anytime I hear somebody say the science is settled, it drives me absolutely insane.

Solani Daytoni (72:46)

Right.

Podcast Host (72:47)

It is. That is the exact opposite of the scientific method.

Solani Daytoni (72:50)

Yeah.

Podcast Host (72:51)

It. It's never settled. And like Feynman the physicist is like, no matter how beautiful your theory is, if the tests say it's wrong, it's wrong. And, and, and so. But the true scientific method is like, to me, if you. If you were going to try to find a music match for it, it would be punk rock. It would be like, no, I'm not going to take your word for it. We're going to test, and we're going to test, and we're going to test and you know, find out what transpires. Yeah, but that seems like, again, the media, in my opinion, is mostly responsible for scientism, not the scientific method, but science trademark.

Solani Daytoni (73:39)

Yeah, I mean, I guess I sort of. See, it makes me think of two things. One is sometimes it takes a really long time to reach those consensuses at all. And the idea that it's just all fixed after that is still not true. Being able to find those exceptions or something doesn't work in this particular area. And not treating that as a false positive or something, but actually trying to follow up on that, that can actually help you understand things at a different level of the phenomenon. Then there's this idea that often there's this situation where people thought that the science was settled or we knew how things worked. And then someone comes around with a different theory that just puts it all together in a way that makes so much more sense. There are two things that I'm thinking about. One is the theory of continental drift. So until the 1920s or so, like, the idea was continents had always been separated. Right. And there were so many mysteries around that that people hadn't really figured out. And why are these fossils seen in Australia and also in India, and they look exactly the same, even though they're so far apart, or why do the edges of the continents look so similar? But it's hard to figure out exactly what was going on in the past and put those lines of evidence together. There was a scientist, Alfred Wegener, who put together.

Podcast Host (75:21)

I've heard of him.

Solani Daytoni (75:22)

Yeah, yeah. This theory of plate continental drift and his. He put together some like five or six different lines of evidence that all of the continents were initially one and that they had drifted apart. And that explains the similarities between the fossils and the different places and the continental shapes. I think there were also some geological similarities that he had found and things like that, but people initially dismissed him. So I think in. In the US there was this conference where they came together and they tried to form a consensus and they said, like, present all of these points of view, and ultimately they just rejected it. Meanwhile, and I think in Europe and other parts of the world, there were some sort of true believers of his hypothesis. And they continued doing research on it and they found more and more evidence. But in the end, what convinced people was actually the US Navy doing research during World War. I think after World War II, maybe during the Cold War, where they were trying to find ways to develop ways for submarines or the Navy to escape being recognized by foreign ships. And in order to do that they had to find a way to navigate submarines without letting. Without letting their location be known. And they developed the tools to do that. And in doing so, they discovered some patterns on the sea floor that didn't really make sense. And eventually those patterns led people to rediscover this theory of plate tectonics, of continental drift. And it all sort of came together again after that. And then the consensus was formed that the continents had initially been one and they had drifted apart. And all of that happened within a few years or maybe a decade. And it just seemed so crazy to me to learn about this idea, that someone had proposed that idea decades ago, and then it was just rediscovered. And once you had the right pieces of evidence, and once people had put it together once again, all of a sudden everyone believes it again. And how quickly that consensus can change, that really surprised me. The other one that I was thinking about was in immunology. So until. Until the 20th century, until the mid 20th century, people really didn't know how vaccines worked. They were. They were making effective vaccines. They didn't really know why they worked. I think in the 19th century, there was this idea that something causes or vaccines. Vaccine basically depletes your body of the specific nutrients that a disease needs to thrive. And so if you develop a vaccine similar in some way to the actual microbe, it's going to deplete your body, and then the real microbe won't be able to thrive on it and multiply. It was really only in the 1950s and 60s that people figured out what was the process for immune cells to multiply into the billions and recognize certain pathogens and then multiply. As you said before, where it's this power law. As long as one or two of them recognize that they can multiply and create this memory that lasts a long time. In the 1940s and 50s, people had developed a new theory of how all of this worked, called clonal selection theory, that if you had one matching, you had an immune cell with a matching antibody to an antigen from the pathogen, that immune cell would be stimulated in some way, it would multiply, and then you would have the memory for this in the future. And in 1960, after that theory was developed, there were immunologists who essentially said, the field of immunology is basically solved. We have figured out all that there is to know, and all that's remaining is just a few little details to be worked out. What's amazing to me about that is this happened just a few years before people figured out that there were different types of immune cells in the body. There are B cells and T cells, and they do totally different things. And all of the higher level of this understanding that we have now has actually changed our ability to make new vaccines and drugs and things like that substantially. And yet back then, people thought it was all worked out, even though they had a theory that did explain the evidence. They just hadn't figured out the level that would help us make new breakthroughs.

Podcast Host (80:08)

Yeah, and unfortunately, that is seen across discipline as well, where people have the idea and everyone dismisses them and says, that is the way it works. And then a couple hundred years later, again, back to Semmelweis, right, With hand washing. And that was because of a social convention. It was seen as unmanly for men to wash their hands. I mean, how crazy is that?

Solani Daytoni (80:40)

That was crazy.

Podcast Host (80:40)

Well, this has been absolutely delightful. Where can people find your work who are listening or watching us now?

Solani Daytoni (80:47)

I think a few places. One is Works in Progress magazine, where I'm an editor. We publish ideas that are new and underrated to improve the world. And that is a print magazine, and it's also a website. I also write a substack newsletter called Scientific Discovery, where I write about breakthroughs and just how sometimes how little we understand, sometimes how much we understand and what the remaining problems are. And then I also run a podcast on medical innovation called Hard Drugs. Hopefully a very memorable name. And we talk about breakthroughs in biology and medicine and unsolved diseases and how we can make more progress on them.

Podcast Host (81:32)

And I can personally attest that that is a very interesting podcast. I. I watched and listened to several earlier today, and I was actually like, wow, this is really cool.

Solani Daytoni (81:44)

Oh, I'm glad you enjoyed them.

Podcast Host (81:47)

All right, our final question here at Infinite Loops is a little unusual. We're going to make you emperor of the world. You can't kill anyone. You can't put anyone in a re education camp, and your tenure lasts only for the time it takes you to give us two things that you're going to speak into a magical microphone and you're going to incept the entire population of Earth with the two things that you say. By that, I mean, whenever their morning is, they're going to wake up and they're going to say, you know, I've just had two incredible thoughts, and unlike all of the other times, I'm actually going to act on both of these thoughts right today and hopefully forever. Okay, what two things are you incepting in the world's population?

Solani Daytoni (82:37)

This is a lot of pressure what do people usually answer to this question? I'm very curious.

Podcast Host (82:44)

Not telling.

Solani Daytoni (82:45)

All right. Well, I do think one of them is consider completely changing your career into biology or medicine. I think there are just so many unsolved diseases and we have had a huge amount of breakthroughs in like the technologies and the instruments that we can use to develop new treatments. But you actually need the people to run these trials and do these tests and do the research on these rare diseases or common diseases that we haven't thought about in a different way. I would say it's just a very exciting time for biology. There's the breakthroughs in genome sequencing and AI and protein design and Sirna drugs and all of these different tools that make it possible to do research like never before. That's one. What's the other one? Well, I guess the other one is since last year, I've been pledging to donate 10% of my, my income to effective charities. And I think it's something that more people should consider. I actually find it really rewarding. I've sort of wanted to donate to things and just not really put aside time to think about where exactly that should go. And I think there are often just really effective ways that you can improve the lives of people in extreme poverty, your people with untreatable diseases, by donating some of your income to effective charities that work on those things, or scientists or privately owned initiatives, things like that. But just having that sense of deciding what is important to you and what is effective and setting aside some of your income to that, it's something that people should consider. I don't think everyone should do it, but the reason that I decided to do it was after I heard about some of my friends doing it. And I just thought that sounds like a great idea. And one of them actually asked me, why don't you do it as well? And I really struggled to think of an answer. So maybe that will resonate with other people too.

Podcast Host (85:01)

I think both of those are great. And especially the setting aside 10%. If you put some of your own skin in the game, it makes it more interesting, but it also makes the world a better place. And we could certainly use for the world to be a little bit better.

Solani Daytoni (85:21)

Right. There are so many of these market failures where sometimes you just need people working on the problem and they hadn't thought about it and they don't have the resources and those aren't going to go. They aren't going to be solved on their own.

Podcast Host (85:34)

I agree. Thank you so much for coming on.

Solani Daytoni (85:36)

Thank you. This was really fun.