Igor Kurganov on the Legacy of Effective Altruism

On this week's episode of Upstream, I'm joined by Igor Kurganov, founder of Raising for Effective Giving, to discuss the strengths, weaknesses, and future of effective altruism. Please enjoy this episode.

Cool. Thanks for your patience. Thanks for joining. We've been meaning to do this for a minute because our Moment of Zen show often craps on effective altruism. And so I wanted to bring on a couple of people or do a couple episodes with folks who are respected in the space who could provide some more context or provide the other side a little bit, just so we can steelman sort of what's going on here. So thanks for coming on the show.

Sure. Yeah, I'm happy that you're holding that opinion and putting action behind it. It's cool.

So let's, um, let's first get into your background and how you came to be influential in the space. And you could give background also about how the space has evolved since you've, since you've been in it.

Mm-hmm. So I, I was, I studied math and then dropped out to become a poker player and poker player poker, as I realized at some point, while really fun and teaches you a bunch of things, it's kind of a self-serving pursuit in the end. The entertainment value it provides is smaller than other sports or mind sports. So at some point, my partner and I in late 2013, early 2014 formed a fundraising organization called Raising for Effective Giving with the goal to identify, well, help poker players in their charitable contributions, basically, where we stumbled upon effective altruism in 2013. And that was the point where my interest in philanthropy was reinvigorated. I used to, like, during high school, work at a soup kitchen on weekends, etc., and like always had that kind of desire to do something, but then was pretty disillusioned with it in later life. So anyway, when I stumbled upon EA, I was like, oh, wow, this is a better approach to philanthropy than I had seen before, because it's attempting to do basically just the scientific method as applied to charity. And I thought that that had a lot of potential.

Yeah. Right. Scientific method. Moneyball is also a metaphor some people, or analogy some people use, the book where, you know, in baseball, they tried to sort of use statistics to sort of optimize sort of coaching and evaluation of players as opposed to just pure intuition. And so that's how effective altruism started. But it later evolved to not just be about how to do charity more effectively, but kind of a broader purview of how people should allocate their careers and time on this planet to do, to do the most good possible, basically. Right. And then even further than that, not just think about how to make life better for people on the planet today, but also how to make life better for future people. And so that's when EA evolved to almost this additional camp of not just charity and helping people in Africa, but also existential risk and thinking about things like pandemics and AI. You invented AI safety. and bioweapons and other things like that. Is there anything you'd add or edit to sort of the evolution of effective altruism?

Yeah, I think that's broadly right. So how I would summarize it is kind of that I think it was in 2008, '09, Will MacAskill and Toby Ord were kind of trying to do some philanthropy and were asking themselves how to do it. I think basically how I see EA is that initially it was asking itself the question of what should we do in the first place? Like, what does good philanthropy even look like? And only later it found actually some answers and now tried to put that into action as well. And I think EA is not yet good at putting it into action. It's much better at this, like figuring out what should be done in the first place, because that is also the culture that it developed over time. It doesn't have a bunch of entrepreneurs. It's mostly philosophers thinking about things, et cetera, right? And I think, yeah, as we will later be talking about probably the things where it goes wrong and where it goes right, I think some of the issues are kind of follow-on mistakes from that. By the way, I just want to insert, I just noticed that my shiner is kind of being seen a little bit. I was not in a fight. I had a skiing accident from a jump. I fell flat on my face.

Oh, no.

I'm not being beaten up. beaten up in San Francisco on the streets from talking about EA or anything of that sort.

Don't worry, I didn't even notice. That's funny. Yeah, well, so let's get into things that it does right and things that maybe that haven't gone so well. In terms of the things that do right, I'll start. I mean, I think it has done an amazing job of capturing the hearts and minds of a lot of young, really talented people. And forming a scene. And sort of tech for a long time didn't really have scenes of intellectual people trying to do the most good. And so effective altruism sort of captured that and has funneled a lot of interesting people who would've maybe gone into totally different industries where they maybe would've had less impact. into either tech or AI or philanthropy in a more effective way, or different kinds of research. So I think as a talent aggregator and allocator, I think it's done a great job. And then also it's popularized things like AI safety. Maybe you could say more about the specific positive impacts that it's had, although some are in the eye of the beholder, of course.

Yeah. So I think as you pointed out with the example of Moneyball, basically, at some point, we like the world has been going in the like 2000s, maybe a bit before, maybe a bit after through like a bit of a data revolution where we now had one, much more data and then much more ability to also interpret it, right. And the field of data science grew much, much more useful throughout that time. So it's kind of unsurprising that this would then also touch charity. And I think EA was just one of the things that realized that this thing should also touch charity. And I think this is also where— that's one area where I see EA to be very justifiable and very strong, basically the area of quantifying quantifiable services that people try to provide, call it services for like nonprofit work, right? So some questions like, for example, Hey, is distributing mosquito nets or this medicine that prevents malaria in some way or reduces the negative effects, what is more cost-effective? And you can do that by doing a randomized controlled trial, doing it in one village and not in another one. And these things are actually very quantifiable and you can apply this data-driven approach to it. And I think that is something that EA has done well. is to say, hey, why are we just like discussing the individual story of an identifiable victim where we, if we have the goal of achieving the most good in like vitamin supplementation kind of effort or malaria or schistosomiasis or whatever it is, then let's just look at the data. But I think that also, I want to here highlight also the limitation of it because like specifically, I think Quantification should be applied to things that are easily, like, actually quantifiable, right? And the thing where I think EA has gone a bit too far, for example, is that it gave— and that happens often, right? Like, people will find, oh my God, things have been bad and now I found this one solution and now I will just overapply it to a bunch of other situations. So people have been using, like, gone too far, I think, and, like, used this quantification approach to their daily life, like while they hang out with their partner. It's like, well, my time is technically worth like 0.2 disability-adjusted life years per day, and your work is worth 0.02. Therefore, you should actually wash the dishes every time and me never, or something like this. And that's literally stuff that— not literally, but paraphrased, I am confident that these sentences have been uttered by EAs. in relationships. And this is, this is going too far with it.

Eliza Yudkowsky tweeted the other day something like, couples should have an understanding that, you know, if they find someone 20% better, they might leave, leave the couple. But they should agree on what is the percent threshold at which point they would stay versus leave. So that's another—

Now, that would be an example of that. Exactly. I think like, so interestingly, it's, I think the, like, if I was to take the essence of what he's trying to say. And well, there is some good essence there. The essence being that if you truly love your partner and your partner has through happenstance found a person that is actually a much better fit to them than to you, than you are to them, well, then you're kind of restricting their total capacity to grow and have a fulfilled life, right? By like Locking them in. So like, that's somewhat true, but again, quantifying it is just fucking weird. And I don't think that it helps that much. And like, most people don't talk in that way. And I think some things, um, uh, we are, we are limited in our ability to quantify, and they should— they are sensibly in the realm of like common sense and heuristics and intuition, et cetera. It's because some issues are complex and we can't just bring it down to like a simple set of factors. Anyway, so that's one thing where EA is strong, one thing where it's weak. I think another area where it's strong, and you've hinted at that, is around what, like, early identifying AI risk as a real concern for the world, and similarly pandemic risks as well. Like, since I've been in EA and we've been funding like AI safety stuff from 2015 and pandemic prevention work since probably like '16 or '17 with our organization. And like EA has been doing it a bit longer. And again, others have been as well. It's not just EA, it's just that they then also put like active kind of activisty type effort behind it. But anyway, so I think that was quite prescient by EA actually as identifying real concerns. And I think that is a strength of it as well, kind of Looking at using reason to anticipate, yeah, what will matter in the future. Turns out that EA actually was pretty good at with those 2 particular areas, basically making like gray swans out of black swans, right? Like for most people, like this was super weird that the pandemic happened and that AI is now all of a sudden on the map. Wasn't for EA. But Also, again, has limitations. Like those things are, one, not easy to quantify. Two, there are tons of issues that EA is not necessarily strong at predicting. It's one thing to say that AI safety and pandemic risk, EA is good at pulling forward and understanding that they will matter. But it's another claim to say we are generally good at pulling forward and claiming which things will matter, and therefore those are the only things that will matter. And some EAs again have kind of like overstepped and done this like assumption of, oh, your area, which maybe make it to be like helping democracy be still functioning in the society or something like this, or journalism be like good and strong and truthful. Like just because EA hasn't focused on it, some EAs still think that they're not important, but might be that they're more important. Might be that they're the generator function underneath some of these issues.

Let's identify the criticisms of EA. It certainly had sort of a PR challenge over the last year, year and a half, partially because of SBF, partially because of the sort of OpenAI situation. But I feel like even before then, there's sort of— it's because as it was getting more and more popular, I think it had some people who thought it was getting too big for its britches or something. Why don't you talk about the criticisms that you've heard about EA and which ones, and maybe let's talk about which ones you're sympathetic to versus which ones you're not sympathetic to.

Yeah. So actually, prior to SBF and OpenAI and kind of the EAACs developing themselves as the antidote to EA, the criticism that EA received was that it's actually on, it's like a capitalistic, pro-billionaire, kind of like justifying those folks group. And then since, yeah, etc., it's, it's now become the, whoa, these are the degrowther lefty crazies. Um, thing is, both criticism— like, I can look at it again and be like, and, and take it apart, or like look at the where, where is like kernel of truth in each of those criticism. And there is a kernel of truth in both. Uh, and like, that's the thing with EA, it's not a, um, like one from up high told optimization function for everyone or something. It's like, hey, we believe a certain— there are a bunch of things we believe. What it more is a belief of is, hey, we should apply the scientific method to some of our opinions. But which opinions people hold within it are pretty varied. Not everyone is a consequentialist. Not everyone focuses on AI safety, even though those 2 directions are the largest ones. So the criticism that I get is, I think that EA is a little too okay with centralization. I think that is a danger that they haven't sufficiently taken into account in the past. And EA is a little too confident of itself, I think, has been over the years, similarly to what I previously discussed of Like, assuming that their answers is like the whole set, like, it's the truth answer rather than just like, they're good at their niche thing. And then, yeah, the whole like, EA is there to justify billionaires' bad behaviors. I think that's just BS.

Yeah, I agree. I think, yeah, the more charitable take on the smartest EAs I talk to, in my opinion, is that they are techno-optimists and they align with EAAC folks in many things, if almost everything, except for on AI, on superintelligence. They're worried about that, but otherwise they are accelerationists when it comes to nuclear, when it comes to other sort of self-driving cars or other technologies that could 100%.

And the thing that I want to point out there is, when that's the case, like, who do you listen to? Where is a higher signal to be found? It's like, there are the, like, EU politician degrowther type person. And now they're worried about AI. It's like, okay, cool. Like, you're always worried. Or there is like the pro-nuclear, pro-healthcare technologies, like people, and they're like, hey, but here, it's kind of dangerous. Like, there's much more signal with the latter group, right? And a lot of the A's actually come from the latter group. And I think that EAC has been making the mistake of oversimplifying and like, kind of like pulling everyone under one umbrella, which is— that's— I mean, it's helpful for our memetic potential, but it's hurtful for understanding what's actually going on. Like, degrowthers are very different people to people who are advocating for AI safety. Like what, you want to call Elon a degrowther and Demis Hassabis and like many other folks? That's like, clearly you have to update your model when you run into those examples. Or they're regulatory capturing. It's like, okay, so you're saying they're doing that. Like, are you gonna tell it to their face? I don't think that's what you think.

We'll continue our interview in a moment after a word from our sponsors. The tech world turns to the Brave browser for its unbeatable privacy protections. But did you know that Brave also has a private ad platform? Brave Ads offers first-party targeting, and it's been cookieless since day one, so you can relax while third-party tracking cookies disappear from the web. Today, millions of people turn to ad blockers to avoid being tracked and pestered online. But Brave's new ad model aligns incentives for users and advertisers. Users earn rewards for viewing ads, which they can save, spend, or pass along to their favorite creators. And advertisers score points for respecting user privacy, generating ROI without invasive So whether it's high-impact announcements on the new tab page or keyword-targeted ads in Brave Search, Brave offers diverse, private, future-proof ad formats for all your business goals. Join the future of advertising at brave.com/ads. Mention MOZ when signing up for a 25% discount on your first campaign.

And so if, if I was to try to characterize the, the substantive differences between EA and EAC, while realising that they have a lot in common, is, yeah, I think EAC has a more bottoms-up or sort of decentralised approach that really believes in sort of like the best way to ensure resilience is to have competition, basically sort of adversarial companies keeping each other in check, as opposed to the big concern is less about kind of runaway technology and more about sort of this apocalypse, this one central authority, less even so regular capture, more just like they fear the one world government that is in the name of creating things better for people, will in fact create things worse for people, this central choke point. And I think there might be people in EA who would be sympathetic to kind of this world government stance in the sense of, hey, we need to coordinate on some of these existential risks. We need everybody to engage in that coordination. Why don't we have a world government? And maybe they don't fully appreciate the risks that that would create. I think that's— and EA sort of has this almost evolutionary view of human nature that— or of society that sort of, yeah, natural selection. We need to have True market competition, it will rise the best ideas to the top. And there needs to be like constant checks that, and thus you need to give them the freedom to experiment and compete with each other. Would you accept that characterization or anything you'd respond to it?

Yeah, I think that's right. Like that both things. One, that a number of EAs kind of underappreciate the risk of centralization. But a ton of them also do appreciate the risk of centralization. Like Bostrom early wrote about the risk of, like, a singular entity that controls everything. Will MacAskill, in his What We Owe the Future book, talks about this as value lock-in. So like, one thing taking over and now, like, just putting their values over everyone forever. Just imagine, like, the Nazis getting AI. or something, right? Like, we are aware of that risk. It's not a— sorry, Iyak, it's not a new thing. It's unfortunately an underappreciated thing of the past. And it's good that it's being highlighted, I think. I'm quite worried about that as well. But it's only one of the risks. There are other ones as well, like the worry about, okay, cool, now you do give everyone therefore like full access to the most powerful model that includes virology and includes like creating smallpox and distributing it all over the world, like, with a plan of how to do that. Like, COVID could be 1,000 times worse. COVID was, like, bad, but nothing relative to what we can do with engineered pandemics. So the thing that I think is being done wrong by both, actually, is, like, difficult problems are difficult. Like, they're not meant to be simplified down to just, like, this, oh, yeah, because I— like, it depends on where you start from. So, like, someone will have as a moral axiom, Like, which is basically what IAC is doing. They take as an axiom that centralization leads to corruption, and therefore you can't have it. And if you start there, you'll arrive there. If you start with the kind of other take of decentralized power, when power can destroy the whole world, is basically dangerous and cannot be created, then you'll arrive at the centralized point. This doesn't mean— like, I can understand each of them. But it's difficult because both have justified points. And like, now we need to figure out what to do with that situation rather than just like defend one side to the, like, in an absolutist way. I think a lot is being lost there. And on the point of like, yeah, the market and how IAG believes that the market, taking an evolutionary framework onto the market and like that, that will take care of itself. It's, it's unfortunate. I think it's pseudoscience, honestly. It's like you can't— like, we have different sciences because different perspectives and languages and concepts are used optimally for these different domains of science. Like, if you're walking around and using quantum language to describe consciousness, that's woo. Like, that's BS, you know? It's, you're being a quantum— it's like quantum mysticism.

Yeah.

Is what it's called. And similarly, like when you describe evolutionary principles as applied to the market, like that can maybe like help with your intuition a couple of times, but taking it as an absolute correct description, no, you can use economics for that. That's what it's good for. That's literally what it's designed for. So yeah, no, I think some of those things around the idealist view on the market is unfortunately wrong. And in particular, like, markets fail. There is nothing that prohibits— like, we had asbestos being released and like part of the market, like, did we get— did the market prevent asbestos? No, because the feedback loop was broken. Like, people only saw it late. And the companies actually knew about asbestos being bad and they just like swept it under the rug. Same thing happened recently with PFAS chemicals being released and 3M paying the $10 billion fine. They knew of them being, these forever chemicals being bad. Still, they just kept going with it because the market has failures. And one of the failures is information asymmetry. So why wouldn't we expect the same to occur in other areas? We would, in fact. And there is nothing that prohibits— that's kind of the mistake of this, like, oh yeah, you build power here and then on the other side there will be a perfect and opposite counteraction or something. Yeah, but while that occurs, maybe everything in the middle breaks. Like, the whole world can go to— it's not the case that like, oh yeah, nuclear weapons are always in perfect balance. No, they can destroy the world.

Yeah, right. And so someone once said something like, markets sometimes have failures, or private organizations sometimes have market failures, whereas public organizations that aren't— that don't have market checks on them always have failures. And so Do you have sympathy to that statement where it's like, yes, the market can't clear the asbestos problem, but if you create a separate organization to try to clear it, that organization will have misaligned incentives baked into it in the sense that it will always— sort of the public choice, there's a principal-agent problem. It will seek to grow, expand its budget, in some ways try to not solve the problem so that they can gain more power to then solve the problem, maybe even perpetuate the problem. Yeah.

Oh, I would agree with that. While working in philanthropy for a while, like, the view I developed is basically that use markets for kind of pretty much any problem you can use it for. Like, really just like maximize market use because it is the best, like, information aggregation system, I suppose, that we have relative to the other one. Markets do fail. Like, one, they have externalities, then they have non-participation, then they have like when you have information asymmetry. Then they also fail. When markets fail, that's where the government is meant to step in. Don't use it beforehand. Let it fail first a little bit, maybe even, or anticipate the failure. But sometimes the government is failing itself pretty often, in fact, right? Or they just miss that markets fail. And I think then philanthropy steps in. That's really how I would stack kind of like the order of use of each of those tools that we have in addressing issues in in society. And because ideally, like, you wouldn't use philanthropy. Philanthropy is not a self-sustaining system, right? The market is. So even though I worked in philanthropy, it's like the goal of philanthropy shouldn't be to forever have this problem. It's like, no, you want it to be out. You want it to be gone. Yeah.

Yeah, it makes sense. And so let's segue to the criticism that you heard in Moment of Zen, because it's another common criticism I think that Antonio had, which was basically, talking about the problems of utilitarianism run amok, where one, I think he was criticizing, hey, they're actually not utilitarian when it comes to— no one's saying we should harvest organs or certain things. Yudkowsky did have a tweet once saying that maybe we should extend abortion up to 18 months, but I think that's a very rare perspective in the community. I think most people don't take it to its logical conclusion. They, in fact, sort of— Antonio's critique was that the EA sort of perspective often aligns with sort of the normie Democrat left perspective on a bunch of different issues. And how is it that that's the case? And so challenges of utilitarianism run amok, or that they're not willing to follow it to its logical conclusion and are kind of faking that it's mathematical, but actually it's just what is popular, i.e., the normal Democrat. And those are both somewhat contradictory critiques, but both critiques. And yeah, I'm curious if you remember any of the critiques that kind of rubbed you the wrong way or any response. Yeah.

So on utilitarianism, again, it kind of touches on the previous point of like quantify the things that are quantifiable. I think we can't properly use utilitarianism because we don't have enough information to do the math. So then, like, if you pretend— you will make very large mistakes if you think you have the full scope of the relevant kind of factors, but you don't, and then you just run with it. Obviously, you'll miss parts of the world. So I agree that we shouldn't just apply utilitarianism to like every situation, but we should apply it to like things that are quantifiable. So that also leads me to then that Like, yeah, they don't run the arguments to the end. Fortunately, SBF did run the arguments to the end. That's the problem. Like, if you are like a naive utilitarian, basically like someone who just believes that, yep, I can do the math and I'll just go with it and I throw out my common sense objections, my virtue ethics objections, other people's deontological objections, etc., then you'll come to some pretty messed up places in your ethics, I think. So I, yeah, I don't think we should do that. I don't think we should run to the end. That said, like, is it possible, like, hypothetically, if we had like perfect information about everything, which is probably physically not even possible, but could we then do better math and use more utilitarianism? Probably. But I don't think we are there. And on the other side, with the democratic point, if I'm not sure it's fully the case. But if it was the case that the, like, lefty Democrat side is very much in line with the utilitarian side, then that would be notable. I don't think that's quite true, though. Like, the EAs are arguing much more for markets. I suppose I actually, as a new person in the US, I'm only here for 2 years, I don't have as deep a knowledge of, like, the standard kind of, like, Yeah, well, the liberal Democrat view on things to make the comparison.

My guess is that less than 5% of EAs support Trump. That's a guess. Yeah.

I mean, you could say the same thing about, I don't know, like SF, I suppose.

Yeah.

Or like many other places. It's true. By the way, the Trump versus Biden thing is another example of like tough problems are just tough and it depends on which axioms you start with. I can sympathize with someone saying, hey, look at Biden, you just can't vote for him, therefore I vote for Trump. Similarly, I understand if someone says, look at Trump, you just can't vote for him, therefore I vote for Biden. Both are justifiable points to get to if you allow yourself to have like this strong axiom initially.

Yeah. Yeah.

And I think just many people started out, Trump was the first one on the scene, right? For like many, he was like the Prior to Biden, a repulsive character to many. So I don't know if they were able to change their opinions, but some did.

Yeah. And I guess I don't understand on which EA grounds would Trump be repulsive. Maybe they would say, they'd probably say something around the lines of like January 6th or something of like he's an actual threat to democracy.

But I'm not sure they would actually. I'm not sure they would point at January 6th. So for one, I think like EA doesn't have answers to this type of stuff. That's like exactly an area where like it should just be quiet about. And I think it is for the most part, to be fair. Similarly, the criticism I've made of EA in the past is like, I'm not sure EA would have funded like gay rights protests in the past or something like this, you know, because it's just not the type of thing it has thought about a bunch. So it shouldn't have an opinion on like, yeah, how does this particular effort stack up against other efforts? And it's much better for How does this malaria intervention stack up against this vitamin supplementation against this other disease? Like that we can answer.

But EA did spend $200 million on criminal justice reform. I saw there was one blog post that was lamenting why they did that. And it was interesting. One of the things they were advocating for was moral diversity in the— because sort of basically implying, in my read of it, that they did it for ideological grounds. And I'm not an expert at this, but my, my loose understanding is that Dusk Moskovitz has loosely supported Chesa Boudin or kind of like the DA of San Francisco, kind of like, and what's happened there has arguably had a net negative impact and that there was like an ideological blind spot in sort of like that would lead someone to support criminal justice reform in this way that divorces That is more of like a wish for how the world worked than actual understanding of how the world works, which is very unflattering and maybe offensive even to actually fully accept what that would imply. And so to me, it's no accident that the biggest leaders are sort of SBF and Dustin. I'll say about the SBF part, You mentioned utilitarianism run amok, but also this kind of arbitrage nature. Like, SBF made his money arbitraging, initial money sort of arbitraging Bitcoin in different countries. But then also EA, to some degree, is about arbitraging different— it's like, hey, I could give money to my family or whatever, the people in my town, but there's someone in Africa who's struggling more, and thus there's an arbitrage opportunity to take the money I was going to give and give it there, and it will have more bang for the buck. And SF is an arbitrageur economically, and EA is about some sort of moral arbitrage. And then Dustin, he's someone who donates to a lot of left-wing causes, basically, among other great things that he does. I don't mean to pigeonhole him. And then also really dislikes Elon, thinks that Elon's impact isn't that big. Or at least has tweeted or on Threads that— and thinks one, yeah, he's commented, hey, someone else would have done SpaceX or Tesla, or he had some post to that effect. So impact is not as big such that we could just give him a free pass and thinks that his actions on X are actually kind of dangerous or something or have negative effects, whereas some people would argue the opposite. And to me, it's telling that there isn't much moral diversity, I guess, or intellectual or political diversity within the EA crowd. Of course, they argue about all sorts of things. I don't mean— are very rational, more intellectually honest than almost every other community there is, to be sure. But there's nobody voting for Trump. And I think that's just very telling. Yeah.

I mean, so just on Chase Aboudin, I've read into him a bit. I mean, looks like a It looks like a terrible, terrible kind of pretty clear also why. It's like, it's a whole history of terrorism in the family, right? It's like the parents died, but then like he lived with his uncle. Oh, that's better. Oh no, the uncle also was like involved in some crap. Anyway, yeah, I have no idea how that type of direction got funding from EA-affiliated people. That said, again, as I said before, it's not like a monolith thing. It's like, hey, there are people with their separate opinions. It's their money. whatever they want with it, which includes them doing things you don't agree with. And if you were to ask people in like Berkeley, SF type, Bay Area EA, I don't know how many would agree with that type of criminal justice reform funding. I frankly don't know. I know that like the EA center where I came from, which is London Oxford, UK, like they don't touch that stuff. We've never even like delved into that there. And so, yeah, I think that was Dustin and probably Kerry's kind of pet issue that they wanted to do. And frankly, it's their money. I don't know what EA should do about it, or like, what's right there. In either case, on the other part around that, oh, yeah, the arbitrage opportunities. So Yeah, I mean, how do you look at it? So like, yeah, EA has definitely— is thinking about like better bang for your buck, which I think is slightly different though to the arbitrage point. But like SBF has learned arbitraging at Jane Street, I think, rather than at EA, which is where he was beforehand, right? Like that was kind of the mode of his work at the hedge fund there, having read the Michael Lewis book. So I think arbitraging is something that's not a bad thing. Like, it improves pricing. It improves information even in the system.

Yeah.

And I think also the kind of applied to philanthropy, the question of whether, hey, if you can help 100 people in Kenya for the same price as one person in like Mexico, Should you do it? It's like, I don't know. There is no truth-based answer to that. Morality has not like some— there's no universal law that tells you whether you should do it. But I'd like to give you the information that this is where the numbers are and then you choose. And frankly, for me personally, I would choose to help 100 people in Kenya ahead of one person in like Mexico or the US. But yeah, I think that's anyone's choice.

Right.

And one follow-up, you mentioned earlier in this conversation of Yes, some people take the quantification thing a bit too far, but it's almost like, why wouldn't they? It's like, you know, that's their hammer. The more things they make the nail, the more power they have, the more status they have, the more fun it is. And so it's hard to have kind of self-corrective measures. Everyone agrees to some extent of, hey, if we're going to allocate money to charity, why don't we have an ROI perspective on that? I guess not everyone agrees, but a lot of smart people think that that makes sense. A lot of people can get behind it. What people struggle with is sort of taking it too far in areas where it's really hard to estimate. It's sort of this false precision.

Yeah, I think that's fair. And the normal— I'll just talk about that. But the normal way how philanthropy worked and works still largely is people donate money, then the question of whether you receive— the organization that was donated to receives more money depends more on how well they can convince other people. And the convincing of other people doesn't work via showing them the data usually, but rather just by like selling a story, right? Like that's how it has been for the longest time. Whereas If you sell a product, then usually your customers are the ones that have the information. Here in philanthropy, the people impacted by it are not the ones that are donating to you. So the feedback loop is broke. So the one thing that, like, more data-driven philanthropy, including EA, is attempting, and EA is not the only one, like John Arnold of the Arnold Foundation is doing, has been, like, advocating for a bunch of this stuff separately, and others. are just like trying to, hey, we need to bring the numbers into that information and have that like feedback kind of into the further decisions of whether they should do— whether this process should receive more funding. And I think that, yeah, that is just like a very correct thought. This is how you improve philanthropy, I think undoubtedly, as long as the goal of philanthropy is To actually impact positively the people that are the recipients rather than make you feel good about what you're doing, which sometimes you want to just feel good about what you're doing and that's fine, then do that. But should that be 100% of your giving? Probably not.

What do you think people don't fully appreciate about EA that you want to share? One thing you mentioned earlier is, hey, there are actual debates within the EA community about decentralization versus centralization, and that's like a misnomer. people have about the community. What else do you think that an outsider like me may not be fully familiar with that you greatly appreciate?

Yeah, so I think you actually mentioned yourself at one point that I really valued about the community is that they are kind of quite intellectually rigorous, or at least attempt to be. Again, humans make mistakes, tons, right? And there are still status games and whatever. But Like, from the communities that I have seen in front of me, that included, like, people studying math at university, poker players, like a bit of VC entrepreneurial folks, EAs. The ones that are trying to get to the truth the most honestly have been, yeah, EAs and entrepreneurs, probably. Like, where entrepreneurs, though, have— are a little bit more like sports fans of their own idea than EAs are. EAs much more are happy to— I'm talking about the good version of EA here— EAs are much more happy to shift the task that they're doing dependent on incoming information than I have seen in other groups. And I think that that's, especially in an ever-changing world, that is actually a very important value. That, yeah, I don't know, by the way, if EA is here to stay. I'm not claiming that, but I think that they were very good at that. And that's a good way of approaching a bunch of problems. People have literally, when they have been working at an organization, and then someone made a post that said that their entire work is BS and it doesn't work. And people respond to that with, oh, great, cool.

Whew.

Good that I found out now rather than later. Okay, I'll do something else then. Like, it's a very trying to actually do the good thing approach. And then given that humans are doing it, you'll have a bunch of mistakes come from that as well.

Will EA be here to stay? EA has an amazing community and beloved by that community. It's got a great brand over the last 15 years or whatever, but it's got a PR problem now. after SBF and after the OpenAI stuff. To what extent it's deserved, up to the listener. But what do you think EA should do now? There's a strategy of apologize and denounce. There's a strategy of actually fight back and say, hey, we've done all this great stuff and you're going to just point out 2 people who failed us rather than we failing them. What do you think EA should do?

Yeah, I don't know if EA should remain as a kind of umbrella of all of these things. I personally, I've been inspired a lot by EA in my early philanthropic journey, but I've developed a set of, as you can probably hear, separate opinions in part to what many of them think. And I Though I would still support EA in many ways, because I do think that the community has fostered a bunch of positive impact. And again, if we look at the actual, like, where did the money go things, I think the vast majority of money has gone to one, human suffering alleviation today. So like global health stuff. And then secondly, like AI safety and pandemic prevention work that again, otherwise wouldn't be on many people's maps to the same degree. I mean, now with COVID it is, but even with post-COVID, people are doing like terrible, terrible interventions and not actually looking at the whole stack of solutions that we could be using to address future pandemics rather than like are so overfocused on vaccines. It's insane.

Yeah.

Anyhow, so what should EA do? I don't know whether, like, EA as a brand matters. I think EA probably should just, like, as far as it remains, it should probably just keep doing the things it's good at. And after you do that for a number of years, you regain a deserved reputation and then you go from there. I think on an individual basis, probably for a number of individuals, it makes sense to just talk about their issue rather than of, like, always rerouting it back to EA. I've had that opinion for a while that people should be doing that, just also in part because if, say, I'm talking about pandemic prevention and then I'm like throwing in, oh yeah, and by the way, that's because EA cares about that, then now the listener is like, wait, why? I was down with pandemic prevention, but now you're like giving me a whole host of other ideas that I don't know if I subscribe to. There's just no reason to. Just talk about the thing. and argue for the thing on the basis of that particular instance, like, topic. So I think that should happen more. And then the people that remain involved and, like, fight for EA, I think they should just keep doing things that they believe in. And if the world agrees with them, then EA will regain its reputation.

Yeah.

Though I do think that some of the criticisms are, like, not Like the 2 PR disasters of SBF and Sam Altman's firing, I do think that there is undue kind of emphasis on the EA factor in the creation of both of those stories. I think maybe it's controversial, but maybe crypto had something to do with the whole SBF thing as well. I don't know, is that crazy to say?

Well, crypto took a brand hit as well, I think.

Yeah, I know. And it's terrible. And like VC also took a brand hit, like a bunch of things took a brand hit. Frankly, everyone deserves a brand hit from this, right? Like, why not? Of course, EA had something to do with SBF. But also so did like Sequoia writing this like most favorable article anyone has ever seen. So does the entire media cohort of like 2022 and 2021 that was like raising the status up of SBF so insanely during that time as if he's the world's first trillionaire. Like, yeah, people I've talked to say that he was the number one person everyone wanted to talk to at parties. It's insane. On the basis of what? Like, everyone created this thing of him which fed and like led others to overvalue him, including him probably as well. Anyway, yeah, I think EA took a bit of a hit there that was part justified, part not. EA should definitely ask itself what it should do, and it should just do good things.

Yeah. Well, maybe let's wrap on that encouraging note, unless there's anything we didn't get to that you think is important to discuss.

Just as a last point, I would throw out that I think it's relevantly true that IAG and EA agree on a number of things, as you brought up. And I think that the disagreements are really important to figure out. I think it is, again, like, which moral axioms do you start with, I think is a bad approach to figuring out truth. You shouldn't just go with like random axioms of your pet issue. And I think that that's the part where IAG and especially Bev Jezus are becoming zealous absolutists, and it would be great if they were instead having like proper, yeah, proper approach to the thinking around it and weren't like quantum mystics. So if he ever wants to debate, I'm down for it. Um, I heard other people reached out and wanted to facilitate that, but, uh, yeah, that's, that's my last point.

We'd also offer to facilitate it as well, and that's a great place to close. Igor, thanks so much for coming on the podcast. Thanks, Erik.

Upstream with Erik Torenberg is a show from Turpentine, the podcast network behind Moment of Zen and Cognitive Revolution. If you liked the episode, please leave a review in the Apple Store. Turpentine is a network of podcasts, newsletters, and more covering tech, business, and culture, all from the perspective of industry insiders and experts. We're the network behind the show you're listening to right now. At Turpentine, we're building the first media outlet for tech people by tech people. We have a slate of hit shows across a range of topics and industries, from AI with Cognitive Revolution to Econ 102 with Noah Smith. Our other shows drive the conversation in tech with the most interesting thinkers, founders, and investors like Moment of Zen and my show Upstream. We're looking for industry-leading hosts and shows along with sponsors. If you think that might be you or your company, email me at erik@turpentine.co. That's erik@turpentine.co.