Auren Hoffman (00:00.394)
Hello, fellow data nerds. My guest today is Neil Buddy Shaw. Buddy is the CEO of the Clinton Health Access Initiative, and he was the Managing Director of GiveWell and the co-founder and CEO of ID Insight. And Buddy, welcome to World of DaaS

Buddy Shah (00:01.852)
Ready?

Buddy Shah (00:19.113)
Thanks, Auren, great to be here.

Auren Hoffman (00:20.85)
I'm really excited. Now, a lot of your work kind of has focused on like Asia and Africa, where I think it's like a harder place to gather data. What are some innovative solutions you've come up with to like get that, get the data that you need?

Buddy Shah (00:37.725)
Yeah, it's a good question. And to be honest, data scarcity is a huge challenge for decision makers in a lot of low income countries in Africa and Asia. Because whether you're a government official, an NGO leader, or a philanthropist trying to figure out how do we do as much good as possible, which problems to tackle, which regions or communities to tackle, if you're operating in an environment where there's just not much data, it's very hard to make the best decision possible.

So given that reality, there has been, I think, a huge amount of innovation in trying to figure out how do we get the right type of data to actually guide decisions to figure out how to do as much good as possible. So just a couple that I've seen personally and worked on in some way personally, one is actually on the use of fairly basic machine learning. So in India,

especially in northern India, there's a huge problem of girl children in rural areas dropping out of school early. And when a girl kid drops out of school, she's at much higher risk for negative life outcomes like very early marriage, poor labor market outcomes, and worse health outcomes. And there's an NGO that figured out a program of basically if you go village to village, you can find the girls that have dropped out.

use village pressure to essentially convince their parents to re-enroll them in school and then provide kind of supplementary education so that they get back on track. The big problem was that we found that at least in the areas they worked, the percentage of out-of-school girls was heavily concentrated in a subset of villages. So like 10% of villages had 50% of all out-of-school girls. The issue was

you couldn't actually look at a government roster or any administrative data and see where are those 10% of villages that have the majority of out-of-school girls because the government data was actually quite poor and inaccurate and it didn't correlate directly with things like the economic indicators for villages or caste composition or anything like that. And so we actually used machine learning based on

Buddy Shah (02:54.921)
the areas where the NGO had worked deeply, and then essentially trained it on all the administrative data we could get by combining different government data sets in order to predict across North India, which are the villages that have the highest density of girls who have dropped out of school. And just using that targeting, basically then the NGO could only go to those selected villages that had the highest density of out of school girls, and they're able to, with the same budget,

basically re-enroll twice as many girl kids and get them back up to learning. And so that's one area where ideally you just have the roster village by village where all the out of school girls and then an NGO or the government could target very effectively that high density area. We didn't have that data, but having a pretty basic machine learning algorithm helped you predict quite well actually where those highly concentrated areas of out of school girls were. So that's like one.

type of innovations, like where there's a lack of data.

Auren Hoffman (03:52.102)
I assume there's like a lot of people who are doing like low cost surveys or they're taking cell tower data or they're, you know, whether browsing data or something like that or kind of like high level data from different kind of data plan usage and have any of those been successful.

Buddy Shah (04:11.425)
So I think I'm more optimistic currently about the low cost survey tools. And then I think over time as cell phone usage and penetration gets spread, especially to the lowest income communities, I think we're going to see more innovations there. But just on the low cost survey issue, it's really hard, again, to figure out what's the economic status, health indicators, education indicators at a village, district.

level within a state in Kenya or Nigeria or India. And some of the innovations there have basically been, and one that we worked on at ID Insight called Data Delta, is actually having a essentially Uber-like network of locally based people that have, you know, are farmers or do other jobs, but then the government or an NGO can ping them. They get some basic survey training and they say, hey, go to these randomly selected schools.

or hospitals and collect some data. And that data could be like, is there teacher absenteeism, which is a big problem in certain parts of the world, or what's the student attendance here, or what are the waiting times like at this health facility? And because normally when you have to collect that data, there's this big exercise, some firm or government hires a big survey collection firm, they call them to the capital city, they train them in.

the questionnaires and they send them out to the villages and you get like a one-time snapshot every year, every few years. And what this enables is actually just like much more high-frequency, high-resolution and cheaper data collection. Now those things have not scaled. I think there's some really good examples in India and Kenya of that type of decentralized low-cost survey data collection showing promise, but we're still a long way towards it scaling globally. So I think that's one that I'm...

Auren Hoffman (05:37.623)
Yeah.

Buddy Shah (06:06.881)
quite interested in. And the key issue there is just going to be, who's the payer that's actually going to put up the amount of capital required to build out this large scale decentralized network of low cost part-time surveyors because there aren't necessarily a commercial place. So it's going to have to be heavily government and philanthropically.

Auren Hoffman (06:27.05)
Why, like how much penetration in these places are of cell phones, of smart phones? Like, I assume that's going up like dramatically.

Buddy Shah (06:39.961)
It's going up dramatically. So I don't have the latest numbers. One of the challenges has been the amount of data that you're able to capture, whether you have an old brick phone versus a smartphone, is meaningfully different. And so while cell phone penetration, even in most parts of the world, are very high, I think that difference of having smartphones catch up is going to be huge. And then always, I think the question is,

Auren Hoffman (06:55.116)
Yep.

Buddy Shah (07:07.005)
the places that have the least penetration are the ones that you most wanna target because that's where economic health, education outcomes are worst. So overall I do see that gap is closing rapidly. And as it does, the main question is just gonna be who's incentivized to drive innovation in low cost data collection, just given the nature.

Auren Hoffman (07:26.55)
What would be like the data set you wish you had, but it's just been really hard to get? Like if you could like wave a wand, what would be that data set?

Buddy Shah (07:35.321)
Yeah, I mean, so one is just like monthly or even, you know, quarterly data on disease burden down to the even district level, if not the village level. It sounds weird to say this, but in health where there's so much measurement across the world, even then, to actually have an accurate view of what's the burden of malaria or HIV or TB.

or some new infectious disease, you're often relying on data that's a few years old or model data that is going to have a lot of inaccuracies. And that actually matters a huge amount. And I'll just give you two discrete examples to illustrate how much it matters having high quality data. So one is comparing between countries. So if you're trying to figure out how are we going to save the most lives with an investment we want to make.

even the same investment like a cheap anti-malarial drug or bed net to prevent kids from getting malaria and dying in the malaria season in two low-income high burden countries, say Guinea and Chad in Africa, actually having the high quality data on burden of disease and cost of delivery means that you can reach and save eight times as many lives in Guinea as you can in Chad. And so, it's just a...

the technical details matter a huge amount. And then not just between countries, but within countries, similarly just from malaria in Uganda, if you look district by district, it actually had high quality data on malaria burden, as well as what types of programs are being run. We found through Givewell's research, we found that you could actually save 10 times as many lives per dollar spent if you target certain districts.

over others if you had that kind of granular data in an ongoing way. And so it sounds super basic, but actually just having high resolution, at least relatively regular data on basic things like disease burden, I think has big practical implications for what kinds of programs you design, how you'd roll them out and where you would do them.

Auren Hoffman (09:47.818)
What are some of the more surprising statistics in public health that like someone like me wouldn't know about?

Buddy Shah (09:55.965)
Yeah, I mean, there's a surprising statistics in public health that, well, some of the stark ones are just how big the gap is in certain health indicators between rich countries and poor countries. So, you know, a woman, say in Guinea-Bissau or Chad, is 300 times more likely to die in child birth than a woman in Japan or Sweden, right? And that's just...

it is so stark to think about a 300x increase in likelihood of dying during childbirth, especially when the ways to prevent that are super cheap and easy. And I think on those kinds of metrics, there's just a lot of very staggering differences. I think on a more even less.

Auren Hoffman (10:40.086)
But even in the US, I mean, you have that right. I think a woman in Bergen County, New Jersey, I think the expected life expectancy is like in the mid 90s or something. And then there are certain counties in America where like a man has life expectancy in like the 60s. So we're talking about a 25 or more year difference even within the US.

Buddy Shah (11:00.558)
exactly for all of us to do.

Buddy Shah (11:05.477)
Exactly. Yeah. No. And so I think any of those kinds of differences across countries or within country are quite striking. I think another more optimistically is, you know, in certain countries in sub-Saharan Africa that were particularly hit by the HIV crisis, life expectancy, you know, a couple of decades ago was in the early 40s. And just within a generation, we've seen that jump to mid-60s. And so...

I think that's another one where it's like, a parent's generation, people are dying in their early 40s. And the next generation made so much progress that they're living to mid to late 60s. And then, I think there's a bunch of like the correlations that are spurious, that the kinds of things you learn in school where you look at correlations and people who take baths over showers are more likely to have cardiovascular disease. And it's just because it's.

Auren Hoffman (11:33.774)
It's amazing.

Auren Hoffman (11:57.387)
Hehehehe

Buddy Shah (12:00.545)
correlated with being old or other lifestyle things.

Auren Hoffman (12:06.242)
One of the things, the more common criticisms you hear about, like the effective altruism movement is that it's just like super clinical and, you know, partially we give locally because we're closest to it. Like, how do you think about that?

Buddy Shah (12:20.465)
Yeah, so I mean, I think there's two parts of that. One is just, you know, what should we do? And then the second is, how do you actually make the argument in an effective way? So, you know, on the, what should we do? Like the clinical, very cold, rational aspect. I actually think, you know, I personally think makes sense. And it's grounded on kind of like two beliefs.

that I think more people agree with than at face value. The first is just that we should care about all humans, if not equally, much more equally than we think, right? So accepting your family, you know, there's our obligations or feelings towards saving someone's life or helping someone out, whether they live 50 miles from us or, you know, 2,000 miles from us, it's...

I do think if you really ask people and they were able to see that individual in real life and not just have it be a conceptual thing, I think they'd feel as much of an emotional pull to help someone if they came across a kid that's severely malnourished and at risk of dying from an infectious disease in a rural part of another country than someone that's has suffering but much more modestly in their own country.

Auren Hoffman (13:49.29)
You don't think there's something human about like really trying to help your own tribe. I mean, obviously humans want to help other humans, but really want to help your own tribe first.

Buddy Shah (13:58.833)
Yes, I mean, I think that is just like an evolutionary biology fact. So not disagreeing with that. But I think that people are, if they're exposed to it enough, are at least convincing and would probably agree that, yeah, theoretically that makes sense and that I probably should care if not exactly as much about someone living in Nigeria as I do someone in the US,

Buddy Shah (14:28.713)
that's the first part is I think making that argument in ways that don't feel alienating or judgmental of caring more about your neighbor than other because we're always gonna have those instincts. And then I think like, how does the effective altruist movement actually try to make that argument? And I think that's where the biggest mistake has been is that effective altruist folks in general tend to think that you can convince people simply by

rationality and empirics. And that's just not what the data shows about how people get convinced by stuff. People get convinced by stuff because people around them or their friends or people they trust believe something and they engage in conversation with about it. People get convinced. I think. Yeah.

Auren Hoffman (15:11.222)
emotional, you know, you see someone in need and you emotionally want to help them.

Buddy Shah (15:15.417)
Exactly. And by personal interaction. And so I think what effects of altruism has been, has done least well, even though I really agree with the general principle of how do we do as much good as possible and let's care about all humans relatively equally, is try, is being creative about how do you make people feel what it's like to live in extreme poverty under a dollar to a day or, you know,

to have a one in 10 chance of dying in childbirth or things like that. And I just don't think that there's been as much investment in creating that emotional connection to why both there's so much need, but also like how good it feels to be able to, you know, actually counterfactually save someone's life for a couple of thousand dollars, which is an insane thing if you're able to, you know, do that and actually feel the emotional result of it.

Auren Hoffman (16:14.002)
When I started learning about the EA movement a decade ago, it was a lot about what you're talking about. Okay, we could save someone's life or we could make someone's life marketably better for a very low cost and here's how we do it. Recently, it has often been about talking about future humans from 100,000 years from now life. And that's where I, at least I personally kind of got lost.

Buddy Shah (16:18.793)
Uh-huh.

Auren Hoffman (16:43.122)
in it. It's like these people aren't even born yet and they're not even going to be born anytime in my lifetime. How do you think about those types of arguments?

Buddy Shah (16:52.381)
Yeah, so I mean, I also just my own personal career and even like emotional and rational orientation is around there's a lot of people living today. There are a lot of tools that we know how to deploy to improve their lives in really meaningful ways. And so that's how I focus my energy. I will say I buy it theoretically, you know, the idea that the vast majority of humans are gonna live in the future.

um, make sense. The second point that like, you know, we should care about that. How much we should care about them versus today or the next generation. I think, you know, I'm still forming my views on, but I buy that even the ones 300, 500, 10,000 years out have some moral worth and we should be thinking about them. And I think that's also fairly intuitive to people. If you just see the reaction around climate change, uh, I think it's something that people can get behind.

And then the third bit, which I think is the most challenging, is just, OK, even if you do care about people, if you buy that, the majority of humans are going to live thousands of years out into the future, and they're worth our consideration. Is there anything we can do about it? The future is so contingent. Is there anything we could plausibly do about it that could increase their well-being? And that's where I was initially very skeptical, but now do see.

some potential ways where we can at least start to put into place processes and ideas that could protect those future generations. So overall, I think, you know, A, there's something there, B, we should approach it with a lot of intellectual humility. And we should start focusing probably on the things that we actually have a higher degree of confidence could positively impact the future rather than getting too distracted with things that are much, much lower likelihood.

Auren Hoffman (18:48.47)
The one way people have really got their lives better is by the areas that they're living in growing economically, having better governance, etc. If you think of the average person in South Korea or Singapore in 1950 versus today, 70 years later, we're just talking about just a massive difference in every single outcome you can imagine. And living standard, health outcome, etc.

And, you know, it's very hard to bring governance into, you know, a place, you know, some of these places, you mentioned like Chad versus Sweden or something. And you probably have a much different kind of governance factor in one versus the other. To me, that does seem like the way you would really move people along, but is a much, much harder problem. Like, how do you weigh that?

Buddy Shah (19:41.629)
Yeah, I mean, first of all, completely agree. The most, I think, the most consequential improvements in human well-being have been driven by the national governments, right? So what China's done over the last several decades, what South Korea and Singapore have done, what even the Indian government's done over the last few decades, are just by far the most consequential things for human well-being. I think the question I ask is, what's actually tractable and doable for...

me and for organizations that I lead. And so exactly, you know, like how much am I actually going to be able to improve governance? I think it's an open question, but I'll say two things. One is while I think those domestic political processes have to work themselves out on their own terms, but while they do, there's a huge amount of good that can be done.

Auren Hoffman (20:14.578)
Yep. Yeah, you can't like go eat a coup at these places here, so that's it.

Buddy Shah (20:39.441)
in those areas. So for instance, you know, distributing anti-malarials in Chad is going to save a lot of lives immediately. And also I think there's reason to believe if you look at Korea or Singapore in the forties and fifties was roughly the same GDP per capita as Ghana and some other countries in West Africa. And then you see huge divergence, at least one of the theories for why is obviously governance. But the other is that both Singapore and South Korea started with much higher levels of both education and health outcomes.

And then that provided enough human capital such that when you flip the switch on governance, they're able to take off. And so I think even just solving in a kind of bandaid way that the health and education pieces can prepare governments to then really take off once the governance challenges are solved. That's one. And then I think the second is how you approach kind of philanthropic work or work as an outsider matter. So, you know, Clinton Health Access Initiative, CHI, we...

only work through the government or private sector. So we very rarely, if ever, will do our own parallel healthcare delivery. It's always with an eye towards strengthening government system and helping that government system improve their delivery. And so I think there are ways to strengthen that internal infrastructure, even as countries figure out their own governance over time.

Auren Hoffman (22:02.514)
And sometimes, you know, that means you have to work with people that aren't necessarily good people. And that may be doing bad things, they might be stealing from the state, they might be doing other types of stuff like, but, you know, there might be no other way to help folks without dealing with some of these bad people. Like how do you square that?

Buddy Shah (22:23.165)
Yeah, I mean, I think two ways. One is we always work at the invitation of the Minister of Health or head of state. And there's just some screening process there of the 36 countries in which we work. There's got to be both the demand as well as the belief that we can do good work. And then the second is to just have best practices in terms of the conditions under which we do different things. Like obviously,

we want to be as effective as we can within a particular political reality and not contribute to any kind of governance or a rule of law that we wouldn't agree with, but that's not always our job to change. And so we're just trying to be as good technocratically and technically as we can within the realities that exist and while not giving them any extra energy. So, I think there's extreme cases where we...

wouldn't work in a place because we don't think we can be effective and lawful. But overall, we found that working through the bureaucracy, there's just, you know, there's a huge amount of good you can do, even if some of that stuff is happening in the country.

Auren Hoffman (23:35.618)
A lot of ambitious young people kind of see the nonprofit world as slow moving. You mentioned bureaucratic sometimes, maybe not as impactful as other arenas. Like, what do you say to those people?

Buddy Shah (23:49.277)
You know, I really think that you have to be more nuanced than that. I think you could make that argument for entire ecosystems, the public sector as a whole, private sector or nonprofit sector. And it's much more about have you found a particular industry or organization that where they are essential to solving a particular type of problem? And then do they do that as effectively as possible? So.

I don't buy the generalization. I think that's certainly true. And because of incentive misalignment, might be more true in the nonprofit sector than others. But I also think if you're being nuanced about it, A, there are just certain types of problems that exist because of market failures or government failures. And there's just an essential role for non-private actors and non-government actors to solve a problem. And

If you think that you're going to solve those problems either through the capital markets or through government, then you probably have a misdiagnosis of what's causing the problem. And then second is, if you found the problems where it really is because of a market failure or government failure, and nonprofits are especially well positioned to solve it, then you need to be really sophisticated in thinking through what are the right nonprofits that are actually going to have that sense of urgency.

to tackle that problem. And then the last is, I think just, you know, owning that different types of organizations and ecosystems require different skill sets, right? So there's parts of the private sector that are faster moving, say like less regulated parts of tech versus parts of the private sector, like healthcare insurance and delivery in the US, where you just have to be willing to own, like one is a more bureaucratic process, another is gonna be a faster moving process and the returns are gonna be different and map that.

Auren Hoffman (25:35.202)
Yeah.

Buddy Shah (25:44.021)
to your own personal preferences and ability to operate in those environments. And so I think in the nonprofit sector, there are unique challenges around incentive alignment and things like that, that you have to be willing to just own and have self-awareness enough that you're gonna be able to function well in that type of environment. But I think that goes for any type of role, even within the private sector or government.

Auren Hoffman (26:09.442)
The decline in HIV AIDS lethality has been one of the biggest success stories in modern medicine in the last 40 years. It's basically gone from like a decimals almost to basically being a very treatable preventable thing. If we had to write a book on it, I'm sure there are some really good books. How did this happen?

Buddy Shah (26:34.813)
Yeah, I mean, and this is something that CHI, the Clinton Health Access Initiative, played a huge part in that story. So basically, exactly as you said, HIV used to be a death sentence. And then pharma develops these breakthrough antiretroviral drugs that almost overnight change it from a death sentence to a chronic condition where you could live almost a full agent, and now certainly you can. The big problem was that...

those antiretrovirals were $10,000 per patient per year. And 90 plus percent of people dying were in markets in Africa, the Caribbean, Southeast Asia that just couldn't afford it. And pharma understandably wanted to recoup their R&D and protect their markets in the US and Western Europe. And I think the first big breakthrough was, Chai and others doing this thing we called market shaping.

which is basically aggregating a bunch of demand. So, you know, the World Bank, Gates Foundation, government of South Africa, government of India, and pooling capital, and then going to pharma and saying, hey, we can guarantee you X hundred million dollars of volume per year, if you drop your price to much closer to your marginal cost of production, and then we'll put into place intellectual property agreements, tracking to make sure those drugs don't make it back to your profit-making markets.

And through that kind of innovative financing, Pharma was able to both protect its, you know, profit-making markets and also produce at huge volumes and low marginal cost in Africa, Southeast Asia, Latin America, and the Caribbean, and, you know, pretty rapidly dropped the price of life-saving drugs from $10,000 to under $1,000, now 20 years later to under $50 per patient per year, which has been huge. So that's, I think, one big part.

I think the second was just some bold leadership. I have to say George W. Bush started PEPFAR, which was billions of dollars in US government funding for HIV. He didn't have to do it. He didn't really get any political points for doing it. But that mechanism in and of itself,

Auren Hoffman (28:37.506)
George W. Bush.

Buddy Shah (28:56.705)
helped to create a sustainable market for huge volumes of diagnostics, therapeutics for HIV, as well as the in-country work. And so I think that was another huge part of the story and probably just an undersold legacy that George W. Bush has had. One of the, I think, most impactful things that humans have done for one another in the last several decades. And then of course, it's just the leadership in country. You see a huge shift.

from the early days of the HIV crisis, where certain heads of state thought it was, you know, a Western fabrication or an attempt to sterilize communities and then actual community activists, people living with HIV, agitating on behalf, changing social norms, changing government opinions, and then governments across Asia and Africa showing leadership in really just once we had the tools in place of getting them to the people that needed it. So yeah, I agree, huge, huge victory. And I think there's...

a couple of big discrete moments that account for that.

Auren Hoffman (29:56.65)
Yeah, it's so much of a victory that you don't even hear about it anymore, which is great. Right. I mean, I'm sure it's obviously still a problem and everything, but just the fact that, you know, if you've met, if you remember in like the eighties and nineties, like that stuff that everyone was talking about it, like it was such a big deal. And now it's like, um, it's like, uh, it's like, you know, it's like measles or polio or something, it's still important, but it's, it's less, it's less impactful.

Buddy Shah (30:18.729)
Yeah.

Buddy Shah (30:21.905)
Yeah, and two things I'd say about that. One is the part that still is, is actually pediatric AIDS, which is extremely sad, but whereas around 80% of all adults with HIV get diagnosed and are on treatment, it's more like half for kids, and there's still a hundred thousand kids who die every year. And that's one thing that we're really focused on. We actually think we can eliminate pediatric AIDS deaths, and we need that same kind of both like...

Auren Hoffman (30:36.987)
Ah.

Buddy Shah (30:49.153)
public pressure and government support to get the tools that we've already developed out to the kids who need them. And so, you know, there's still unfinished business there that we're really excited to accelerate but agree it's been a huge success. And then the other part is just how do we learn from that lesson in HIV and do similar things and the other big remaining killers.

Auren Hoffman (31:11.826)
You mentioned some machine learning initiatives earlier, given the advances we have with AI, what are some new applications where you could see impacting public health in the future?

Buddy Shah (31:25.153)
I think we're early days, Oren. We're thinking about it a lot. A couple of pretty discrete examples. And obviously, I think there's a big difference between fairly basic machine learning things that I described, like the girls education stuff, and true, more generative AI applications. So the first is, huge problem in many parts of the world is just the ratio of qualified health care workers to the population, right? Very few physicians to the population.

very excited about is you could essentially develop a sophisticated AI decision-making algorithm for a community health worker. Someone who's a high school graduate close to the community, they're already delivering care and driving.

Auren Hoffman (32:03.821)
Yeah.

Auren Hoffman (32:07.434)
Yeah, they're, they're a nurse or health worker. And now I've got this AI assistant to help me triage things and everything like that. Yeah.

Buddy Shah (32:13.905)
Yeah, both triage and treat. And I think it's very plausible to have, you know, over, I don't know, five years, whatever the time will be, 10 years, community health workers that can functionally operate at the level of at least a primary care doctor, which is hugely exciting and obviously there's regulatory. Exactly. Yeah. And I think like just the technological.

Auren Hoffman (32:30.742)
Yeah, and obviously that's not just true in the developing market, that could be true in the US as well.

Buddy Shah (32:40.917)
pathway is so plausible and it's just a matter of can we clear the cultural and regulatory burdens and put the money behind developing those. So that's one.

Auren Hoffman (32:48.65)
Getting some of that data to do that is hard and there's privacy things around it in the U.S. There's HIPAA stuff. In some ways, maybe it's even easier to do it outside of the U.S. or something, you know, because there's maybe less regulatory burdens. Like, how do you think about that?

Buddy Shah (33:06.609)
I think that in certain content, there's a couple of considerations. A, there are often fewer regulatory burdens, but you also have to be very careful about potential backfiring, you know, either for often companies or outsiders coming in and using national data, which we're very careful about. So you'd want to do it with government buy-in. I think you'd want to find a forward looking government that really wants to lean in and seize the benefit for their population to develop the ideal kind of...

Auren Hoffman (33:17.74)
Yeah.

Buddy Shah (33:34.665)
lean innovation oriented regulatory regime around it so that you could collect data, but the government feels comfortable about it, communities do, and use that as a training ground to develop some of these tools in particular for those populations. So that's one. I think the second is just while there might be fewer regulatory burdens, it is just way more data scarce, right? So if you think about the amount of health data we have in the population in the US, it's just way more than...

Auren Hoffman (33:57.783)
Yeah.

Buddy Shah (34:01.525)
how often people are getting diagnosed on a regular basis in these countries. And so overall, I think if we're gonna do it, we should do it like very well. Think about what are the data sets you need to develop, go and get the political buy-in and actually in a tailored way, collect that data to...

Auren Hoffman (34:16.63)
Yeah, I'm a big believer we should try to do it here in the US and then just allow anyone in the world to do it. And it's definitely possible to do it in a way to protect everyone's privacy. But there are reasonable concerns that people have that we just have to, we have to LA and work on it.

Buddy Shah (34:35.077)
Yeah. And then the other, in addition to the decision-making algorithms, are just AI for diagnostics. We already have some pretty good machine learning stuff that should get way better on things like a community health worker being able to diagnose cervical cancer with automated visual exam where you essentially have an ML algorithm that takes pictures and helps with diagnosis. You can do that for TB, skin cancer. There's a lot. And so...

Auren Hoffman (34:50.924)
Yeah.

Auren Hoffman (34:59.406)
skin cancer, yep.

Buddy Shah (35:04.133)
I think the main thing, I have some trust that in high-income markets will figure out the financial incentives and regulatory paths. So there's going to be a lot of innovation here. I think the disease burdens are so different in other parts of the world that we need to make the markets so that cutting edge tech companies and pharma actually invest in the applications there. So this machine learning diagnosis for cervical cancer.

or cheap x-ray detection for TB, it's just gonna be, they're gonna be lower volume markets in rich countries because there are fewer people that have it. And so we need to create the financial instruments that incentivizes cutting edge companies to actually try to develop those technologies for the diseases that disproportionately affect low income countries.

Auren Hoffman (35:39.095)
Yeah.

Auren Hoffman (35:52.382)
One of the issues in just the health world is that there's just a general decline in trust of experts globally. Certainly that's happening in the U.S., but it's happening all over the world. What do you attribute that decline in trust to?

Buddy Shah (36:13.545)
I honestly don't have a good theory of it. I think I could, I mean, there are guesses I could make. The one that comes to mind, which I don't know how much explanatory power it has, is that it is much easier to try to become an expert oneself, right? Just because the access to information in differing points of view.

is so much higher now than it's been in the past. And so I think there's some almost like fooling oneself into thinking that you can arrive at the right conclusion through first principles and self-learning, which I think in a lot of ways is actually a positive thing in the world, like decredentialing and being critical of experts actually very positive, but I do think it has certain negative implications where...

Auren Hoffman (36:58.689)
Yeah.

Buddy Shah (37:09.065)
you know, you haven't had that training, you're trying to parse through very complicated academic literature to figure out like, does X cause Y? And just like the whole correlation causation challenge, but magnified when you're looking at dozens of studies and trying to piece together a whole picture, ground that in some understanding of like the biomedical mechanisms. I think it can lead people to think that they've read as much and come to a different conclusion on.

things like vaccine effectiveness or links with autism or any number of things, whether in health or others, where I think we should have robust debate, but I'm just skeptical that all of us should try to be experts in everything and come to our own independent conclusions. I do think we need to have trust in a robust scientific community and have those debates play out openly and rigorously.

within that world. So it's not to say just like default to yes based on what experts say, but I think some calibration's required.

Auren Hoffman (38:16.594)
One time I have seen some things where some things that recently have happened that maybe they always happened where like, OK, someone is a.

but really good at treating cancer or doctor really good at treating cancer. And so, well, then they get, you know, they kind of get looped into like actually forming health policy. Um, and because they're an expert in cancer, there's halo around them. And then, but it turns out actually they're not really that great at health policy related stuff like it. You know, they're not any better than, you know, um, another person would be at that. And so you almost have the halo going the other way.

sometimes and then and then people are like, well, why is this person like they're saying, oh, we just got to shut down this whole thing or I can't go to work or whatever. Like you can see how that would happen, right?

Buddy Shah (39:11.197)
Yeah, no, for sure. And I think that's where...

Buddy Shah (39:16.717)
you know, A, just like having certain rules of the game around scientific method also prevents, yeah, experts in one thing, flexing into it, acting like experts in something else and, and potentially to negative effect. But yeah, I mean, I do think it's hard.

Auren Hoffman (39:33.174)
Are you worried about this kind of like, again, the conspiracy theories go, but like a little bit of the capture, like a little bit more of this capture. I have a friend who's a well-known cancer researcher and he did this thing on, you know, in this thing in the subset of people with smoking and to his surprise, like at least in this particular study, it didn't really affect their outcome.

Um, and he got a lot of pressure from his peers not to publish this because they said, look, we'll do more harm than good. You know, other people will, you know, go on it, but he's like, well, that was the finding. So, and, you know, and he didn't know what he should do. Um, you know, should he just, you know, because you can imagine, you can, you can make both cases, I guess, right? Like, how do you think we're playing out in this kind of like, it's somewhat politicized world.

Buddy Shah (40:26.073)
Yeah, so there I have a very clear view that it is just essential to protect the scientific process and individual researchers coming to conclusions based on the data and letting them publish that with a much lower bar for like, how is that going to get interpreted by others? Because I think once you start having that type of not quite policing, but social pressure to not

release results that you think stand on their own two feet in terms of the analysis, because you're worried about how it's gonna be interpreted by others. I think it starts just to erode our ability to actually produce new scientific content and is just really dangerous. So overall, I think individual researchers should just ask the questions that they think are important, answer them as best they can, subject it to peer review.

And we just need to be very knowledge oriented. And once we start worrying about political ramifications of how that knowledge could be used, and putting that pressure on scientists, I think we're in trouble. There's other mechanisms to mitigate those effects. But I don't think it should be on the shoulders of individual scientists to self-censor results that they think are actual robust scientific findings. And I would be worried about that if it

increases in frequency.

Auren Hoffman (41:56.126)
Okay, a few personal questions. I'm really interested in people's names. And obviously professionally you go by Neil Buddy Shaw, but individually you go by Buddy. How do you think that's affected your life?

Buddy Shah (42:13.118)
Yeah. It's interesting because Neil is, it's not like I go by Buddy because that's a long Indian name, right? Neil is actually a very common Indian American name in part because it's both Indian and... Exactly.

Auren Hoffman (42:18.804)
Right.

Yes. Yeah, I know so many Indian deals. Yeah, yeah, yeah. They sometimes spell you spell it N E I L and it's often N E A L.

Buddy Shah (42:32.689)
Yeah, and sometimes even N-E-E-L. But, you know, interestingly, I wrote my college admissions essay on this because my, you know, basically it was pure happenstance. My older brother was a couple years old and my parents like, oh, you have a new friend, his name is Neil. And he said, he's not my Neil, he's my buddy. And so everyone in my family just started calling me buddy. I went to school and wouldn't return.

Auren Hoffman (42:35.1)
Yeah, that's true. Yeah.

Auren Hoffman (42:42.95)
Ooh, okay.

Auren Hoffman (42:53.911)
I'm out.

Buddy Shah (42:58.301)
respond to Neil when the teachers called me that. So they called my parents, they're like, oh, I think your son's too young for preschool, like bring him back next year. And they said, just try calling him buddy. And it stuck. But I think, you know, whether it's just my disposition or because of the name, I have kind of own like, essentially trying to be friends with and be able to get along with many different types and groups of people. And so being a buddy, and so that's kind of like stuck as at least part of my orientation.

Auren Hoffman (43:26.414)
because I've known you for a while and you are a very friendly person. And maybe it is because of that, right?

Buddy Shah (43:31.186)
Yeah, so who knows how much is like the influence of the neighbor. That's just how I was. But yeah, it's stuck and I made a conscious choice in going to college and then after that to just use it as my professional name as well.

Auren Hoffman (43:43.85)
Okay, so there's all this stigma about like the Forbes under 30 under 30 list. You're on it. Like, how do you think about that?

Buddy Shah (43:52.713)
Yeah, well, I'll say how I think of it, it's like very different from when it first happened. I think I was the first year that they had the list was the year I was on it. And now, so now I feel like very sheepish when I still see it pop up in my bio. So I'm like, you know, telecom seemed to delete that. I think at the time though, just like to give the generous read for it, especially in the social sector, where there are fewer...

discrete markers of success. I think in the, my friends at, like I was a social entrepreneur in a nonprofit sector, my friends who were entrepreneurs in the private sector, they're just much clearer markers of success in terms of who you raise money for, then how highly valued your company is, exit, and beyond those ones, just how successful your company is on some standard metrics. The social enterprise sector has been trying to figure that out. And so I think in those early days, as it was becoming a more robust sector,

Auren Hoffman (44:31.916)
Yeah.

Buddy Shah (44:48.189)
it actually did create some incentive for people who had a lot of outside option value to take the risk and start a nonprofit that could solve a big problem. I think it's a pretty bad way to incentivize people to do that and I think there are much better ways to do it. But at least in those early days, it probably served some positive for just like getting talent over the hump to...

go into certain spaces, but yeah, now I try to distance myself from it.

Auren Hoffman (45:22.474)
Yeah, I definitely want to be in the Forbes 100 over 100. That's my goal.

Buddy Shah (45:26.873)
Exactly. I mean, that's the other part is just like in so many of these spaces, in certain cases like you have accomplished something meaningful under 30, but the majority of it, including myself, was like, I hadn't accomplished anything meaningful at that point. I had started an organization that had some promise and that was about it.

Auren Hoffman (45:43.151)
Alright, last question we ask all of our guests. What conventional wisdom or advice do you think is generally bad advice?

Buddy Shah (45:50.513)
Um, I, you know, I know a lot of people say this, um, but it's like, follow your passion. Um, and I think part of the reason it's bad advice is like, I think the more precise way of saying that is like, understand what you're really good at and follow that. And I think sometimes there's just like a big difference between what people think that they care about or are passionate about. And

deep self understanding both of where they're exceptional and also when they're most in flow state. And so I would just focus much more on that self understanding of where you think you're truly phenomenal and leaning as much into that.

Auren Hoffman (46:35.862)
It's hard when you're young, maybe when you're 50, it might be, I'll have a better idea of like what your superpower is. But when you're 20, it's very hard to know those types of things, right?

Buddy Shah (46:44.789)
Yeah. And that's why I think there is like, there's a time aspect to it where the early parts of your career should be probably about learning as much as possible about different strengths, weaknesses, times where you feel energized, times where you feel less so. But I do think with active introspection, you know, like if you do enough in our thoughtful enough and expose yourself to enough different environments, I think, you know, I don't know the age, but

I do think you can start to develop and hone that understanding of where you're most uniquely positioned to contribute in the world and where you tend to be most excited and try to find that Venn diagram, which sometimes could be where you're most passionate, but might not be.

Auren Hoffman (47:29.506)
This is awesome. Thank you, Neil, buddy Shaw for joining us. The world of dads. This has been a ton of fun.

Buddy Shah (47:34.933)
Thanks, Auren. It was fun being here.

Auren Hoffman (47:37.134)
All right, that's great. Really great. Thank you very much.