How do you put a price tag on a human life? It sounds like a cold question, but for grant makers, it is the necessary calculus of doing good.
"Every time you choose whether to take a more dangerous job at a higher wage... or choose a house that's closer to environmental toxins but is a little cheaper, you are implicitly putting a price on how much you value a year of your life."
In this episode, Ben sits down with Deena Mousa, a grant maker and thinker at Open Philanthropy and Coefficient Giving. Deena takes us inside the difficult decision-making frameworks used to allocate finite resources—from the "Coefficient Dollar" to the complexities of measuring pain.
"In high-income countries, the question is 'Will AI replace radiologists?' But in a lot of low-resource settings, the question is not whether AI will replace the radiologist, but whether you can access a radiologist at all."
We also dive into the role of AI, why government procurement might be the world's most underrated problem, and, on a lighter note, Deena’s specific writing habit regarding white noise.
We cover:
The Calculus of Altruism: How philanthropists use "revealed preference" to value a year of life.
The Pain Paradox: Why standard health models struggle to account for chronic pain and suffering.
AI for Good: Why the risks of "AI washing" in nonprofits are real, but the potential for capacity building is massive.
Systemic Bottlenecks: Why fixing boring government procurement processes might be more impactful than flashy new policies.
Life Advice: Why you should probably ignore the advice that resonates with you most.
"Often, the people listening to a piece of general advice are exactly the group of people that should be doing the opposite. They are already leaning too far in that direction."
Summary contents, transcript and podcast links below. Listen on Apple, Spotify or wherever you listen to pods. Video above or on Youtube.
Contents
00:00 Introduction
01:17 Valuing Life and Health
05:46 Challenges in Measuring Pain and Health Outcomes
13:32 Creative Process and Research Methodology
18:38 Journey and Early Experiences
22:23 Debate on International Aid and USAID
29:20 Impact of AI in Global Health and Development
36:25 Overrated or Underrated
44:59 Exciting Projects and AI for Good
46:14 Balancing Cause Areas and Funding Decisions
58:31 Advice for Aspiring Philanthropists and Innovators
Transcript (This is AI assisted, so mistakes are possible)
Ben: Hey everyone. I'm super excited to be speaking to Deena Mousa. Deena is a grant maker and thinker on global health and development, as well as thinking on innovation and progress. Deena, welcome.
Deena: Thanks so much, Ben. It's great to be here.
Ben: Your work at Open Philanthropy and Coefficient Giving and really in Global Health overall requires making decisions about where to direct finite resources.
So at the heart of that, you have to grapple with the question of how do you value life and health? What is the value of life? How do you think about that and even begin to approach that question?
Deena: Yeah, it's a really great question and it's a hard one to answer in large part because it is so personal and there are so many different ways of approaching it.
There are. Fortunately, some studies and research that we can ground some understanding in. So there are two broad literature on this. One is on stated preference, which is where you ask a group of people or survey participants in a lot of different ways how much they value in monetary terms a year of health.
and this can be by asking them, for example, how much they would pay for a particular drug or treatment, how much they would pay to avert a particular outcome, how much they value particular aspects of their health. And these kind of converge on. one form of the answer, which is, what do people say when you ask them how much they value health?
Then there's another literature that is focused on revealed preferences, which is what we actually do and how much we actually pay when we have to make the choice. So every time that you choose whether to walk home or get in a car, every time you choose whether to take a more dangerous job at a higher wage premium or choose a house that's closer to some environmental toxins but is a little bit cheaper, you are implicitly putting a price on how much you value a year of your life.
And so there are quite a few papers that take a lot of these decisions and try to back out of them, how much people, implicitly care about their life in monetary terms based on the actions they take. And this sort of leads to another cluster of answers. and by looking at a combination of these two research studies, you can get to a triangulation of what is approximately right.
This is of course a very hard question because when people make decisions, they're not necessarily explicitly factoring in all of these effects, and they might not necessarily be accurately diagnosing them. And similarly, in the stated preference, it's quite hard to know exactly what decision you would make when it comes down to it.
So there is quite a bit of uncertainty around that, but that gets you at least to some estimate of how people feel and react in real life.
Ben: And so in the US when you are thinking about this, either thinking about grant making or in your research, do you have a range in your head that you're thinking about?
I'm aware that it'll probably be different in us, say to Italy, say Nigeria. but do you have a range on what, let's call it a statistical life, year of life or a life year to differentiate it from some other thinking? What do you say?
Deena: Yeah, absolutely. So one, upfront clarification is that different institutions have different numbers for this and in large part because they're doing different things.
So the value of a statistical life, quote unquote, that is used for example, in the United States when deciding on, health insurance or, healthcare issues is different from the one that you might use in a different country. But that in turn is a pretty different question from the one that we are answering, which is how to allocate our finite resources across interventions that mostly target health and ones that mostly target income.
So we, when we calculate, the value of a particular grant. We have a metric we call Open Philanthropy dollars or coefficient giving dollars now. And that is used to equalize health and income gains. so that allows us to basically compare across health and income increases to figure out how much you should value one versus the other. Though of course the metric of coefficient giving dollars isn't necessarily comparable outside of the institution. It's, not like a, objective dollar metric, if that makes sense.
Ben: Sure, yeah. You can't compare it to what the US government is saying about anything to do with dollars. But even within that, do you think your order of magnitude is about right when you're using this a hundred thousand and, spreading it, they say, that's within plus or minus 20%, That's the sense I have.
Or do you think there's a risk that you are completely wrong on that? And maybe we can think about that because you had a really, interesting essay on thinking about pain and I think pain's a really interesting thing to think about because on the one hand you could think for minor pits of pain we use say aspirins and things and they're really cheap and we don't seem to value it very much.
And another aspect actually, if you ask doctors and you think, oh, this person's in pain, we've got cheap medications, or they're out, it often doesn't. Get into their calculus that much because we think, pain or maybe acute pain is this is not life-threatening at the moment. You'll get round it.
On the other hand, if you speak to some people, it could be, really valuable to try and avoid or chronic pain is, different and it comes in enough and it's also more nebulous because there's some part of it which is psychologically based, some part which is maybe biologically based and they interact.
So I was interested in, whether you've got your reflections on how you've been thinking about pain and, whether this is the type of metric which could be used at all with this, or whether we should really be thinking about something else.
Deena: Yeah, absolutely. It's, a really good question because I think pain is an area where DALYs get a lot harder.
In large part because it is quite possible, if not likely, that a DALY is not a DALY in all circumstances. You might care significantly more about alleviating a marginal quote unquote DALY's worth of pain. The more acute pain that you are in. For example, and there was one study, that I referenced in Japan that implied that this might be the case.
That the more serious the injury or illness, proposed to individuals, the more that they were willing to pay to avert the same amount of additional pain. And so we don't necessarily think of pain or negative health outcomes as linear. but we might have all sorts of different curves for how much we value amounts of health at different points in our life and at different, levels of discomfort.
so I think this is, an area where yeah, certainly the model and here is six we are using might be a little bit blunt. With that said, it is, like you said, it's such a hard question to answer because a lot of people might have quite divergent beliefs and feelings about what that, what the shape of the curve is for them individually and It would be quite prohibitively hard, I think, to apply that sort of differentiated DLI metric, especially depending on the disease. And looking at how it affects different people in to different orders of magnitude and so on.
Ben: So does that mean we just have to be really cautious with these type of models and thinking when we come across something like pain or is it a kind of this is the best we have, so even if we're a little bit unsure, it's gonna give us some sort of answer?
Deena: Yeah, I definitely think it's, a good reason to be cautious and a good reason why. The quantitative answer is not necessarily the be all, end all, it's a good heuristic. It's a good guide to break down a problem. But for example, with with chronic pain or with, areas where the DALY figure that is given appears to be a little bit questionable or potentially a little bit too high or low one thing we often do is compare it to the DALY burden we've used for other diseases and just think conceptually about whether this is reasonable.
So for example, we might say, the DALY burden associated with cluster headaches is x. Now let's compare it to the DALY burden that we've used previously for o other quite, painful diseases, for example, is it significantly higher or lower? and this can help us triangulate a little bit, what are we really saying when we're using this multiplier for this disease?
Ben: Sure. And when you were looking into pain and, doing your research on that, was there anything which surprised you around that? Or do you think is actually, this is really misunderstood, we should really clarify this more.
Deena: I was very surprised by the degree to which we don't really fully understand pain.
We think of pain as a direct response to the magnitude of an injury we receive. you stub your toe, you feel some amount of corresponding pain, but that's not necessarily always the case. The amount of pain we feel can be very disproportionate, either higher or lower to the actual damage that our body has felt.
And so it is like a much more complicated and psychosocial, fa thing and phenomenon that we experience as well, which I think I underrated to some degree. And for example, also our understanding of like chronic pain and how chronic pain comes about, is quite limited. Our understanding of, phantom limb pain and why it only impacts some people and not others is, similarly a bits body.
So I think I, I underestimated the degree to which there are a lot of open questions around the biological components of pain and how pain literally works.
Ben: Yeah. Particularly in chronic pain. I think the scientists call it heterogeneous. We really don't know some of the causal aspects. So you might present with these same type of symptoms, but the underlying can be completely different.
And we might not even have that big of a clue. And it's interesting maybe making a segue into something where we might think we have clearer numbers around, maybe thinking about the UK or any system which has got a constrained budget. How would you deal with funding for extra money, for instance, for a diabetic versus perhaps a preterm baby or maybe say a rare genetic disease at birth or something like that.
So the overall disability life years, the dalis come out at maybe 20, 30,000 pounds, maybe a little bit more in dollars for the diabetic. So you go, oh that's really good in terms of, life, but the preterm baby is coming into hundreds of thousands of dollars. Same with genetics.
So we could be talking half a million up, up to a million, million or more. And that under kind of pure thinking, just cost benefit or say this. First or the utilitarian kind of goes, it's the diabetic all the time. But then when you ask people, and here in the UK we try and ask people, although we argue about whether this is it, people end up saying no, we think the baby should be worth a chance.
Or, particularly with the genetic and then there's arguments around that. How do you think we should think about that? And do you actually fall on one side of the fence around, how we should make that decision?
Deena: Yeah, it's a really, it's a really hard question and an important one. I definitely won't presume to be able to give a particular answer or say, necessarily what the government should do in a circumstance like this.
It's an impossible one. But I will say that, the, allocation methods that I'm talking about are. Something that's like quite useful for us in our context as a philanthropic funder who are, have a particular pool of money. We are allocating it across cause areas we are explicitly going into it being cause agnostic and impact maximizing.
So we know that is our approach and aim going in. Whereas a government is in a very different position to an individual donor or a philanthropic funder. And they ultimately are tied to a social contract with their citizen. They're democratically elected, they're intended to serve the voice of their people.
And to a large degree, I think the sort of bare utilitarian calculus needs to also incorporate what people believe, how they feel what their opinions and perspectives are on how funding should be allocated and what the priorities of the government should be.
Ben: Excellent.
Perhaps you could talk about your creative process or your research process. Say you're picking up a new topic like AI in health or you're doing some of this research within pain. How do you go about thinking about that? And also you've had a journalist background, so you've been interested in stories and the narrative around that.
So I'd be interested if you've got any particular creative process or anything you'd like to share around that.
Deena: Yeah, absolutely. I picked up freelance journalism, in large part as a forcing mechanism. I thought it would be a really interesting way to. Create a structure for myself to get myself to explore the questions that came up day to day that seemed confusing or potentially interesting to me.
so it, I think has been quite good in that it has given me another lens such that when I read something that is, seems a little bit contradictory or seems a little bit unexplained or intention with something else that I know I have a mechanism by which I can say, okay, I, actually wanna get to the bottom of this.
and I'll go and interview several people and I'll, if it starts to look like an actual story, put together a pitch to an editor who will then, be able to act as like a little bit of a scaffold for me as I go off and explore the question and distill what I what I believe in writing.
And that has been quite helpful in large part because I, have this strong belief that, writing is thinking and, in order to. Be able to understand something very clearly. You just have to be able to write about it very clearly in a way that someone else will understand. and that process I think has allowed me to dive, shallowly into a lot of different areas that I find somewhat interesting.
And, so I'm, really glad that I've that I've been doing that over the past couple of years. I think it's been a really enriching process for me, in thinking about the world. Yeah, I've, oftentimes the topics that I'm looking at intersect somewhat with the topics that, I'm thinking about in a professional setting.
but it's a quite different lens on them than doing, research or making grants. And so I think it's quite complimentary in my experience.
Ben: And do you like making notes as you go? Do you read and write? Do you write two three hour chunks in the morning? Are you an evening person? There seems to be 1,001 ways of doing it, but I'm interested in your actual process.
Do you write by hand or you, straight on computer, or is it just all completely random?
Deena: I tend to type and I usually do all of my interviews first and get a good picture of what I think the story is in my head, and then tend to write the draft all at once. I usually like to leave it for about 24 hours and then come back to it fresh and do the editing then, because I often think that you're a bit too close to it when you've just written it for the first time.
So yeah, that tends to be my process and I like to reserve Saturday, Sunday mornings to, write for a few hours each week, which I think has been helpful for, forcing some cadence and consistency. But certainly everyone has their own sort of writing habits and patterns that I think they're so personalized and individual that, I'm not sure anyone's, writing habits would work necessarily perfectly for anyone else.
Ben: I think there's been a book looking at this. And if you look at people who, are thought to be really great, you can find, they work at every single different hour of the day with every single different habits. Although there's maybe at the meta level some quirky things about them, which seem to have a, pattern, although the actual individual quirk is hard to know.
Do you ever listen to music or things or is it like music or, no. And is there anything about your process you think is perhaps a little bit quirky? Like you don't write with a pet crocodile beside you or anything, or? Really, a lot of people seem to be, at least in humanities creators, a lot of them seem to be high on drugs or alcohol or whatever when they write.
But I'm assuming you are not, but, yeah. Anything quirky or do you think you're just normal?
Deena: Unfortunately, nothing's so out there. I, I can't listen to music with lyrics while I write. otherwise I think the words will distract me quite a bit. But, I do often put on white noise particularly like the sound of rain, I think is helpful for, just being a little bit more focused and concentrated.
No, I can't think of anything particularly weird. I'm a bit, I'm a bit more of a night owl than most people. And so I think often my in spite of plans usually to finish something early in the morning, usually my, most productive writing comes quite late at night. so that maybe is, is slightly quirky though.
I think that's a bit common with writers.
Ben: I think not having lyrics is maybe somewhat common, but actually wanting brain music or white noise is perhaps a little bit more uncommon. So that's that's an interesting note. Maybe that's a good segue into your journey as to where, how you've got here.
Because if I read it correctly, you were a teenage medical inventor of some sort, or at least a kind of social entrepreneur back where you grew up and then you did a little bit of a consultancy gig and then have come into that. I dunno, what did you learn growing up being kind of social entrepreneur like, and, any thoughts of that in terms of how you went from there to here?
Deena: Yeah, I grew up, so I was born in the us and I grew up between upstate New York and Cairo. I spent my summers in Cairo. So usually about three months of each year, I would be there. And I grew up with, yeah, a family of much older siblings who I think, really shaped a lot of how I think about things.
And, I think just being constantly exposed to, adult conversations and like around the dinner table. I think to some degree, shaped my upbringing a little bit. I did work with my sister on a solar energy social enterprise in, in Egypt that distributed solar water panels that, heated water for households and also for organizations, around the country.
So I think that was a particularly valuable experience. given the fact that, like you mentioned, I worked at McKinsey as a consultant and now at Coefficient giving as a grant maker, getting some exposure to implementation and to operations was incredibly valuable because I think it gave me a taste of how.
Unintuitive difficult. It is, in many ways actually getting something done on the ground is quite hard and quite hard in ways that are almost impossible to predict at the outset. so in each particular circumstance, something will come up, but you don't know what it is until it happens.
And, yeah, often things take a lot longer than you project they will at the beginning. And I, so I think, that, that's been a really important perspective for me to have as I've, consulted people from mile High and also given grants to operators to be able to understand a little bit, of what they're going through.
Ben: I guess that will tilt you into thinking about impact if you've done social entrepreneur things as a teenager. Is there anything you've changed your mind about then over the years in thinking about it? you hint at one in the sense of actually implementing things always seems harder than the paper because unexpected things always come up, but, any grand ideas where you think, I used to think this, but now I think.
Really it's that.
Deena: Yeah, it's a great question. I think that is certainly one of them that you mentioned, that, operations is much harder than it looks. and that strategy often, di dilutes the nuances of the details of actually getting something done. I think in consulting, for example, I went in half wondering whether this this common refrain that consultants are just risk management, might be true.
it seemed shocking just because of the sums that, fortune 500 organizations pay for consultants. But I actually think I, I came out of it more positive about the value of consulting, if that makes sense. I think one thing that surprised me is how. Integrated, a lot of our consulting teams were with the organizations where we worked and how almost over time essential we became.
And I think that the scope and scale of that across, industries, across organizations, big and small, across sectors, I think was quite surprising to me. So I guess I came out a little bit less cynical about consulting than I went in, which I I think might be the reverse of what, what happens to a lot of other people.
Ben: it's good because you actually did go into consulting. 'cause otherwise some people would've said, don't do consulting at all. Go straight to making actual impacts. So I'm glad you got something out of it at the high level, I'm interested in a couple of systems piece. So there's been a lot of debate on, for instance, US Aid or even just international aid in general.
And on the one hand you could say we lose usaid. There's definitely some projects which had high value impact. Those are not happening and there's that is negative versus having it right in that world. That counterfactuals definitely negative. I guess some people would though say that there are some projects which maybe not higher return, and we really wanna build on local infrastructure and local projects for people, because there's a whole set of examples when, foreigners go in, try and give money, with good aims and have essentially messed it up.
It's gone to the wrong people. They haven't understood these things, the unexpected has happened. And so that money has not only been wasted, has been wasted, but has been actively poor. And there's examples like that. I don't think there's a consensus answer with within that aggregation. But I'd be interesting from in your reflection this mean that actually it's more important the work that you guys do in terms of philanthropy, because there's a little bit less international aid philanthropy from the government level, or is just this whole idea of aid has perhaps been a little bit more down rated, so it's making your life harder.
I'd be just interested in any reflections on whether USAID was any good pros and cons and how it's affected your work.
Deena: It's a very timely question. I think there is a very reasonable argument that usaid would have benefited from reform. there, there is a group of people that is or a sort of line of thinking that is, the bureaucratic nature of foreign aid in the US and in other large countries has overwhelmed, the institution and it needed a significant overhaul.
I think I sympathize with some aspects of that argument, and I think it is re reasonable to say that, even a pretty heavy handed, overhaul of USAID could have been good. There is a version of that could have been, very net positive. I think a version that, some have talked about since of, like massive slashes to foreign aid, that are indiscriminate or, just go across the board pretty heavy handedly are.
certainly not the version of, the overhaul that I think, would have been most effective. And I think that is something to note, which is like, there is a, there is a, good and a less good version of reforming usaid, if that makes sense. And I think in a, just the magnitude of funding that comes through foreign aid, in global health and development is staggering.
And losing a large proportion of that really would, significantly alter the field. regardless of the sort of spread of projects. And maybe you can say that the range of cost effectiveness that, happens within institutions, it is just so much funding that it, inevitably leaves a massive gap and a massive hole.
I do think that makes our work as philanthropic funders, more important and more sorely needed. And often opens up, more opportunities for us to do cost effective work. Though it is also of course much more difficult when a lot of USAID programs were supporting foundational infrastructure that then enabled us to build, for example, on top of them.
when that infrastructure goes, in some cases there isn't an actor who is well positioned to fill the gap.
Ben: So net, net, although reform of usaid, a reasonable person could have argued for that. What we have now in practice is probably net negative because of the absolute levels of, funding cuts and this issue of some infrastructure with no credible stakeholder or actor being able to do that.
So those, not only those first order projects have gone, but a whole infrastructure layer has been, withdrawn, made, life difficult. Was that a fair summary of thinking around that?
Deena: I think that sounds right. yeah, there is a version of USAID reform that, that could have been certainly net positive, but I think, a combination of magnitude of the cuts and the way in which the cuts were decided, if that makes sense.
In terms of how to prioritize projects and how many people opined on that and how long that process took, I think, resulted in, in less than optimal implementation. And therefore, I definitely do think it was net negative.
Ben: Another systems piece. Perhaps this is wider in, in thinking around cost benefit type of analysis or at least simple utilitarian thought.
I still think with suffering in the ecosystem around some of the effects that F-D-F-X-F-D-X and SBF had on thinking about that partly because of funding and partly because they were so explicitly we now think naively utilitarian around their thinking. Do you think there's still an echo of that and do you think I, mean I guess this is on some of this, but I think it's wider philanthropy in general.
And we are probably a little bit younger and a little bit more apart from it maybe. Has it made your thinking around that systems piece change or been influenced at any way?
Deena: it's a good question. much of the fallout of that happened before I joined coefficient giving, and I was very much looking at it from the outside in.
I think coming into my role, I am less of a sort of hard line utilitarian than maybe some say SBF or pure rationalist or pure effective altruist. And so I think I was already coming from, a little bit of an outside in view, and a bit of a sort of more, more sort of mixed perspective, if that makes sense.
I definitely do see it as having shaken up a lot of the effective altruist or rationalist community and a lot of the thinking around it, I think, a lot of people reflect it after that. Which I think is quite healthy sometimes. For sure.
Ben: Yeah, I'm definitely an outsider looking in.
I think one of the things I reflected on was, what would they call it? Normy stuff. So I now really understand why you just have a regular independent board if you can do it well, because this is the things which gives you distance over some of the things. And when I see things which don't have that, which I mean there's no guarantee, right?
A lot of those things go wrong. But I now understand why those boring, normal measures are put in and maybe there's too much bureaucracy with some of those. But you can see now the counter examples where they failed because you just went on high trust with a couple of individuals with no oversight accountability.
So it was a, win for normal boring covenants. which is which is weird, which is an interesting meta layer. Meta layer thing. Maybe thinking about a meta layer thing. You've done some work on technology and particularly ai, which within say health or climate or a lot of global development is potentially an enabling layer.
Yes, AI in itself might settle certain projects, but maybe it's more importantly is going to enable a lot of other things. So I'm interested in what you think is actually happening within ai. I think global development and health, you wrote a really interesting essay on sort of radiologists, how it hasn't, for instance, there are still a lot of radiologists.
In fact, we don't have enough radiologists, even though. AI is doing that very well, and that's in a narrow thing, but it looks like it could be supporting quite a lot of a infrastructure layer and then into global development in general. has this altered your kind of thinking, and does this mean we should be in investing more in a metal layer, or at least in transforming that metal layer into something practical?
Or is it really the kind of, yes, this is a new technology, but our old models and our old ways of thinking? Just incorporate that and that will just be the same, but with different measurements. I don't know Your thoughts on ai.
Deena: Yeah, it's a big question right now. maybe I'll start with some things that worry me and then some things that I'm excited about in the area.
So a few concerns I have upfront. One is this idea of using benchmarks to determine how much AI can do in the field or in practical terms. for example, looking at AI performance on a benchmark, or a test related to a particular job. And then to say this means that AI can replace.
87% of people who do this job, oftentimes the map is not the territory with benchmarks. And so there's quite a big difference between the questions on a benchmark and the true process of actually doing a job or doing a role. And usually it is quite hard to come to a like detailed and nuanced understanding of what it really means to do a job and how people spend their time in that role and what are the long tail outcomes that are, come up very infrequently, but are actually quite important to get right.
So I think that is one worry I have with ai. another, I have with AI in development in particular is incentivized AI washing. people are very excited about. Potentially implementing AI in a lot of new contexts and settings, which I think is, great. And I'll, also get to that in a moment.
But I worry about setting upfront too much this idea of we would really like to fund AI related projects. So we really like to invest in AI related projects. And then encouraging, nonprofits, organizations that are doing great work by default to feel pressured to incorporate AI where it doesn't necessarily make sense or it isn't necessarily adding much impact so that they can then say, yes, we are AI enabled, we, qualify for this particular program or this particular incubator.
I think there is always a danger with technology of, implicitly shoehorning it into processes where it is not quite the right fit, because it is the next really exciting thing. and I think, in part, in particular, we wanna be careful as funders in terms of how we are shaping the field in that way, and whether we are driving incentives in a negative direction in that sense.
But with all that said, I think this is quite, AI and development is quite nascent, and this is a really exciting moment for it. and I am quite excited by the potential for AI to increase capacity and throughput at, almost every level of development. I think often, human capital is a constraint for a lot of different fields from everything from, do you have enough doctors to do you have enough people to run the government properly?
And if AI can act as a multiplier on people's time, for example. Increasing diagnostic throughput for every doctor by making them a few seconds faster with each patient. Then I think it could have, significant positive effects in low and middle income countries in particular where, that is really the key blocker.
You mentioned, a piece I wrote about radiology, that was really focused on high income country settings, where the question was will ai, replace radiologists and therefore make, radiology interpretation much cheaper and much faster? But in a lot of, low resource settings, the question is not whether AI will replace the radiologist, but whether you can access a radiologist at all and whether AI might act as a sort of better than the default standard of care option for you.
So I do think the questions are quite different in low and middle income countries, and I'm very excited about, the potential for AI to increase capacity in lower resource settings.
Ben: So that sounds like we should move quite fast and hope not to break anything as opposed to moving very fast and breaking lots of things.
Deena: that's a very good question. I think I mean I think that there is also obviously a reason for caution in particular because there, there is a sort of paradigm in which if you believe that AI as a technology will get significantly better in the next, let's say one to three years, implementing a, very imperfect solution Now could, this is an example, but could, lead to mistakes that therefore eliminate trust in AI that will then be very hard to win back. Even if in say, three years the models that you're working with are significantly more accurate or significantly better in a healthcare setting, for example.
So I think that is like one consideration, which is that, when you're introducing a new technology, you are setting the terms of people's impression of that technology. and in a lot of cases right now, AI is people's, most people's interaction with AI in some countries is simply as a tool that spams use to send them a lot of, scam text messages on WhatsApp and scam phone calls.
And I think it is important to consider like the. The relationship that you're creating between the technology and the people and to a, I think to some extent the more that you force technology in where people do not organically see a need for it, the more you can sort of risk creating this negative impression and those impressions can be sticky long after the technology itself has changed significantly.
so I think that's, maybe one reason for caution, particularly right now.
Ben: Sure. I hadn't thought about that. It's very hard to recover from a bad first date is what you're saying.
Deena: Exactly, yes.
Ben: And I guess we also potentially move through some one-way doors, as in you do X and it's quite hard to row back from X, but maybe there might be some two-way doors.
I find it's really interesting that the technologists have a. There are different clusters of them, obviously, but they have a different view, from what a lot of the economists and sociologists think around this. And it's really interesting that they're so far apart. But yeah, when you don't have that consensus, it is a little bit tricky.
Maybe we should do a little short round on overrated and underrated. So I'm gonna give you like a little topic. you can pass or you can say neutral but you can say, yeah, I think this ideal concept is underrated or overrated, or we should have more of it, or less of it or something like that.
Yeah. Are you good?
Deena: Yeah. Happy to do it.
Ben: Yeah, let's do it. Okay. so the comedy one, malaria nets underrated or overrated.
Deena: I would say in general, underrated, still, underrated
Ben: even though we've been going on about them for so long.
Deena: I think so. I think we are in a bit of a bubble in that, within the world that I'm in and that to some extent that you are in as well.
people take for granted this idea of malaria nets are the canonical, reliable, cost-effective intervention. But I think one that's not necessarily really like common knowledge beyond this sort of corner of the world. And two, there is still quite a lot of need for it. Like it is still, an intervention that can absorb more funding.
And I think, until we reach equilibrium with that, I guess by nature or almost mechanically, I would say underrated. Yeah.
Ben: So hundreds of millions of dollars of capital could still flow to mal nets and we're nowhere near, reaching that. Okay. Interesting. Good. Climate net zero. Underrated overrated.
Deena: It's a good question because this is such a polarized one. you have almost, I guess in the US at least about half the country feels extremely negative about this and about half the country feels very positively about this. And I don't know, I would almost say as it nets out neutral yeah, I think when I think about each half, I would say, has been overrated maybe in the past 10 years, but in the last, like year or two, I think that has, like the pendulum has swung and it's pretty sharply corrected.
So maybe I'll say neutral on this one. Okay.
Ben: What, so your, observation is that the US or the world is neutral in it, but are you yourself now neutral or, actually you were saying that maybe the pendulum has swung. Too far. although the world is maybe balanced that actually because of where the pendulum sung, it's a little bit underrated.
Deena: I think. I think, where the pendulum has swung recently, maybe, makes it underrated, if that makes sense.
Ben: Yeah. It's, definitely a tricky one. I would probably say it's overrated for environmentalist and it's underrated for the average person in America, but, yeah, that's, the difficult one.
Universal basic income. Overrated. Underrated.
Deena: Wow. Such a good question. I would say almost. Not discussed, if that makes sense. Really, it's gone. I thought this was
Ben: the Californian solution to everything. I've obviously been misled.
Deena: I think it disappeared. it had, it's had its moment in, the, presidential campaign a few cycles back.
But I, my understanding is recently, like UBI as a policy proposal has fell out of the popular zeitgeist. I suspect as we think more about AI, labor market disruptions solutions like UBI will have their day in the light again. and I'm quite interested to see how different the reaction is this time around.
So I guess I would say, yeah, perhaps for our current like labor market structure and where we are literally today, slightly overrated, but for potential AI labor market disruptions in the future, potentially a little bit underrated in that yeah, there's not, there's, there aren't a lot of, concrete policy proposals around that yet.
and I'm, I think I'm excited to see more thinking around it.
Ben: That's very, well hedged on that one. It's interesting to see that it's actually completely di disappeared from the debate from what you said, or in which case, that would make me think it's a little bit underrated, even if it might have been overrated before.
I was reading that in Ireland, they're gonna do a small version of this, aimed at artists. So if you can get on this list, there's a few thousand artists which will have a small artist, UBI. And to your point about job disruption, maybe this is the offset about, okay. Put all of your work illegally into l and s because it's out on the internet, but maybe we'll give you a UBI instead.
Okay. Long-termism or thinking maybe about the very, very far future, is this overrated or underrated?
Deena: I think in non rationalists and effective ultra spaces, as in for the average person? I think somewhat underrated. I think it's just very instinctively much more salient to think about things that are impacting us today.
Even if, when you truly consider it and when you think about, future generations of, the people you know and your family, you would care significantly more than you do today. I think the average person likely might care significantly more than they do today. if like properly framed about longer term impacts, if that makes sense.
Ben: Yes. But the rationalists themselves maybe think about it a little bit too much.
Deena: Yeah, I think, it's a a concentrated dose of something that is, not popularly considered in, in most of the rest of the world.
Ben: Sure. I think they might just go too far. They just, maybe if you get us to think about 100 to 200 years or 500 years, that's something useful.
Once you go out to say a million, all of the math seems to get really screwed because when we're very, not very good with large numbers anyway. Definitely. But very good. Yeah. Okay. Overrated, underrated direct giving. So this is direct giving of cash transfers.
Deena: Oh, good question. I, guess I would have to say similar to my mechanical answer on bed nets underrated.
Yeah. Build demand in general use. Exactly. still lots of demand and, and not enough supply. So I'll say underrated.
Ben: Great. Okay. AI existential risks. So a GI risk. Do you think this is overrated or underrated?
Deena: it's really interesting, like some of the other questions, it depends radically on who you talk to.
I think, if you spend a few weeks in the bay, it will feel like there's almost nothing else happening in the world. Whereas, if you spend a few weeks. Many other places in the country it's a non-issue or not barely something that's even regularly discussed. So I guess I would say, I wish it were more evenly rated.
If that makes sense.
Ben: Sure, that makes sense. Okay. And Cairo, do you think that's an underrated or overrated place?
Deena: I would say underrated. I think yeah, often, people don't don't manage to get out there themselves, but I think it's, a city with so much rich, with rich history and on top of that it is a city that is just incredibly alive.
It reminds me a lot in some ways of New York and I'm biased because I've spent, quite a bit of time in both cities, but it is one of the few cities that I think compares in terms of sheer, like volume of people and energy and, like the idea of the city that never sleeps and the 24 hour city.
Ben: Yeah, very intense. I was there once for a little while to go to a friend's wedding, and it did ha it did actually have that New York vibe of everything happening all at once everywhere. yes. and last one on this philanthropy in general, do you think it's overrated or underrated?
Deena: Yeah, I, again, am biased, but I will have to say underrated.
You go, you're gonna have to say
Ben: that otherwise you're in the wrong career.
Deena: Yeah, absolutely. but I truly believe it. when you think about the volume of charitable giving each year compared to, for example, like pledges that have been taken or the amount of money people, intend to give, I think it is quite low.
And yeah, I think it's underrated and I think I am excited over time to see more. Very high salience and very visible and easy, ways of giving charitably for people who, maybe are stopped by a form of decision paralysis or uncertainty. but who would like to give, which, I suspect is, is, quite a few people.
Ben: I'd accept that effective philanthropy is definitely underrated, but I still think it's a potentially a bit more of an open question than I was hoping for about overall giving, which seems to have, a little bit more of a, more patchy history. Although I guess that's why places like co giving and open philanthropy come about.
And so any current projects that you're working on that you're particularly excited by, or things that you have granted to or looked at and when these are really great projects and have now come by because you are, because you've been working on them.
Deena: it's a good question. one thing I really love about my job is over the past year I've been able to look at such a wide variety of cause areas.
I think I started the year thinking about antimicrobial resistance and, I'm, ending the year thinking about AI for good and I've hit, many, cause areas in the middle. I think I'm, I am quite excited about the potential for AI, for Good. In particular, I think it's a very nascent field and one where it's possible that a lot of the problems that implementers have are shared even across verticals.
whether you're doing AI for health or. AI for agriculture or weather forecasting? I think there are maybe some common needs and challenges. And I'm excited about, continuing to learn a little bit more about the field and thinking about ways in which we might be able to support that, though, that's quite uncertain right now, so I don't, wanna make any particular promises around.
we, don't have an existing program in the area yet.
Ben: Sure. And so how do you balance out between looking at something like, antibiotics, a MR and. Climate and AI or systems, pieces and things. And you said that the own internal sort of metric is this a hundred thousand in coefficient dollars, which you, seem to be able to, go to, and I guess you're looking for 10 x, five x, a hundred x, that type of, that type of return.
And I guess that's it. You just have to, whatever area you are doing somehow bring it back to these are the inputs and the costs and get to those sort of figures or how do you think about weighing it up and at the strategic level, I guess you've gone development health, ai, also existential risk and some other elements where you're talking about maybe not, some other people aren't looking at it so much and have ideas about tractability or other, but I'm interested in how your own thinking is shaped about what should we work on and then when we work on it, how are we really thinking about whether this is a good idea to fund.
Deena: Yeah, it's a great question. in large part because by nature our time is limited, which means that we can't investigate every cause area. And so making prioritization decisions based on relatively little information upfront, is quite difficult, but a necessary part of the process. I think we often try to split the process of investigating an area into, parts that will allow us to pivot quickly and frequently.
So we might start, for example, with a two day investigation into an area that focuses on one particular crux that could basically suggest that we shouldn't look into it at all. So for example, if there's an area where. It seems really important and really neglected, but we're quite worried that there are basically no solutions and therefore it is totally in intractable, in which case we shouldn't do a full investigation.
we might just spend two days specifically looking at tractability and solutions. so I think that helps quite a bit. And then in terms of prioritizing cause areas, for funding, we do rely quite a bit on this framework of disability adjusted life years and percent increase in income as the sort of north star metrics or heuristics.
and we try to consider three factors, importance neglected is and tractability. So importance represents the overall magnitude of the problem and how many people it impacts. so this can often be dollies in terms of, health issues, for example. but might, go by other metrics depending on the area.
neglected is a measure of how many other funders are working on it and how many dollars are chasing the problem with some sort of adjustment for, what in particular they're working on. And tractability usually focuses on a few specific things that we could hypothetically fund if we were to enter the area.
So we'll try to do direct models of the impact per dollar spent on a few hypothetical grants that we might make in the area to understand whether they might be comparable to the rest of our portfolio. so that is, often how we try to make comparable judgments across very different cause areas, though, of course that is very tricky and there's a lot of, uncertainty there.
Ben: That's a really detailed level of analysis. Do you think it's worth, if you are an organization or even a startup thinking of applying to funders or even thinking of submitting to you guys that they already start to think a little bit in these terms? Or is it just more useful for you to do the analysis on what they would give you and not necessarily say that they've already started to try and translate it into a market size or impact size or a probability of success?
Is it actually helpful if they've already tilted it into your language and thinking, or is it. Better for them to just do their purpose and their things and you do the analysis on top.
Deena: Yeah, certainly the vast majority of grantees we work with, will not have used exactly our model or not be thinking about things in exactly the same way we are.
So I think that's something that we understand. And so therefore, it's quite important for, us to be able to convert even very different materials into something that makes sense within our framework. I think it is often useful, when, organizations have thought about things in, somewhat similar ways though, if that makes sense.
When they've thought about, this is a really big problem. here, we know the shape of the field and what other funders are doing. we know why we think what we're doing is particularly, effective at solving this problem relative to other potential solutions.
things like that doesn't necessarily have to be. very, quantitative terms or in exactly our language. But I think the general mode of thinking of, here's why we think what we're doing is quite effective is, often helpful for us to know.
Ben: Great. And couple of final questions leading into the end.
If you are the Ben benign dictator of the United States or the world, you can think about this globally or the us. what one thing, or I guess if you've got loads of ideas, you could do two or three things, would you implement in the world today?
Deena: Very hard question. As in, any sort of like law or regulation.
Yeah. You're
Ben: the ruler of the world, so I guess it'd be a law regulation more than more than anything else. But yeah, you can make this law and let's assume people are gonna even follow it or policy. or maybe you could, go the other way, right? You could take away a policy which you think is so bad that actually the, most net benefit we'll get is if this thing disappeared rather than this thing came in.
Deena: Yeah. Is it, too niche and boring to say I would fix government procurement processes in the US and in other sort of high income countries?
Ben: No, but I guess the, challenge then is, so you think government procurement A is not a, is first of all probably a pretty bad, but b would probably have an impact which is much larger than the average person might think.
Deena: let me caveat that with saying I've certainly not done a cost effectiveness analysis and I've, definitely not even it was just a thing that immediately jumped to mind. I'm certain there are many other much more cost effective things you could do if you were a dictator of the world.
So this is like quite a small petty thing to be considering. But yeah, in large part, I think our ability to,
Ben: I mean in some ways the US government did have this in mind ish. When they came up with Dodge, they thought, there's something going on here. I know that was only a part of it, so not completely crazy.
Deena: I think it's, certainly at the core of a lot of our ability to get things done. And, maybe this is also maybe again, I'm, biased here 'cause it's something that I personally experienced. For example, when you take public transit and you think about that the magnitude of the budget of the MTA and yet how expensive it is to do any particular project or process to improve the subway system.
and I think this goes not only for transit, but for things like construction, for for almost anything that the government, procures or does or has an RFP for. I think it's a wide ranging problem that impacts a lot of different sort of verticals that we experience day to day, if that makes sense.
Ben: yeah. So when you talk about procurement, it's not just buying supplies it's the building side of things as well.
Deena: Yeah, exactly. So any sort of like contractor agreements or RFPs.
Ben: Yeah, I guess that doesn't have a single bullet to it in a way that say if you propose a global carbon tax might or something.
So if you had 10 minutes with the president and they were gonna do the thing, you would implement that, this is the thing you would say, you'd go fix your procurement, fix your building.
Deena: Yeah. I, don't know. I think it's in part a sort of a funny answer. and in part something that I think is like quite important.
I don't know how much that I would stand by, if I spent 10 minutes thinking about it, which I certainly hope that I would, if I, were, to be dictator for. For a day. I'm not sure that's what I would come out with, but I, do think I do stand by that. It's quite an important problem.
Ben: Yeah, no, I think that, I'm surprised you didn't go for health though, but you won't. Procurement maybe maybe health is even too complicated. Not as tractable 'cause it's too many bits to it. Okay. I'm gonna ask the question in another way, but bring it rather than at the policy government dictator level.
If I give you a billion dollars, which actually might sound like a lot, but actually we know is not a huge amount in the context of the world. And I say, you have to spend this, by the end of the call you have to decide how are we going to do this billion, what would you what would you allocate it towards?
You've got a billion and you can do anything with it. But it disappears unless we use it. So we can't like invest it and save it and do something even better in 10 years time. But what would we do now?
Deena: Yeah, it's a good question. It's a hard one in large part one because often the answer might be divided into many little grants, but of course if one person is trying to think about how they would spend the funding, then you really just have to pick a few organizations or just a handful of, places to direct the funding. and then the second problem comes in of you get pretty, potentially pretty strong diminishing returns as you give the extra marginal dollar to the same organization, if that makes sense.
so the first, a hundred million, they might spend extremely cost effectively, but the further you go, the further down of the list of projects they have goes and, the more you start hitting constraints, sometimes thinking about like these very scalable opportunities I think there are certainly some organizations that that could take a large chunk of that billion and use it well.
again, bias, but organization, give, I think, would definitely be on the list. give directly thinking about, Gavi, for example. that could use a huge amount of funding. yeah, thinking about like organizations that are, Able to take a quite a large budget and have limiting diminished return as they move down the next dollar.
so I think ideally I'd split it up to at least five or 10 organizations, but I think something like that.
Ben: Sure. The effective giving organizations like your own and, but then if you had a pet project, so you're not allowed to give it to an organization. And maybe this would be a smaller amount of money and maybe this isn't first order the most effective thing.
What sort of project or areas would you fund about? I think about this a little bit because one of the criticisms around some of this line of thinking is it never really works on the systems level. Although we've slightly negated that with AI and policy and, the things that they're like and sometimes I talk about, or I think about where we've had great social progress, like over hundreds of years progress, Slavery, women's rights, disability rights have come from this systems level change. And this is my art or your storytelling hat on. Because if you don't have a storytelling narrative for some of this, then you don't have that enabling layer at all to get change. But then, maybe that isn't that effective and we have to do it in a different way.
But I'd be interested in your thoughts of that. And did you have any other smaller pet project that you would want to fund?
Deena: Yeah, it's a, good question. I think abstracting away from. My day-to-day role here a little bit. I think one area that we've looked at this year that I think has seemed quite timely and important is health system strengthening.
And in particular helping, low and middle income country governments allocate their health budgets and their budgets more generally as effectively as possible. I think this is particularly important now given, the perspective aid cuts and potential reductions in budgets for a lot of these countries.
And so it feels quite salient. But I do think there is a lot, of potential benefit in allocating those budgets a bit more cost effectively. And also in terms of, helping with budget execution, for example, ensuring that the budget can be spent, and can, and the money can get out the door each year which is almost an interesting and opposite challenge.
Ben: Excellent. Okay, last question. Would you like to give, any advice to people, listening? So this could be life advice or people thinking about how to give money or perhaps thinking about going into a career, either in terms of innovation research or grant making research, or your thoughts on where you've got to today.
life advice from Deena.
Deena: Another tough question, I don't know that I could presume to give personal advice. I guess I will say one meta observation that has been interesting and helpful for me is that often some subset of people listening to a piece of general advice, should take that piece of advice and sub subset actually need to take the exact opposite.
Okay. and often it's reverse correlated with who is hearing it. usually the listeners of a particular, or the readers of a particular outlet are quite similar to. The person giving the advice. And so they have like similar inclinations. And so they're exactly the group of people that should be doing the opposite.
They're already leading too far in that direction. So maybe my, meta piece of advice is try, to collect advice from pretty disparate sources and from people that you maybe don't necessarily normally interact with and from different bubbles of the world. I think, it's very easy to end up in like a sort of, relatively small intellectual circle that thinks relatively similar in relatively similar ways.
But, I think one benefit I found of living in New York is that I am constantly exposed in a way that I might not have been in SF or DC to people who are in completely different industries who, have never, heard about a lot of the things that, that I spend my days thinking about, but vice versa have a lot of pretty interesting and pretty different robustly different experiences.
Ben: Excellent. So that sounds like if you are looking for the advice you want, you should actually ignore it and go the other way.
Deena: Yeah, in some cases I think that's right.
Ben: advice is overrated. Great. on that note, Deena, thank you very much.
Deena: Thanks so much, Ben. It was a lot of fun.