Is Effective Altruism Evil?

Alright. Welcome to whatever I'm gonna name this podcast. I'll figure it out eventually. yeah, we should brainstorm that at some point, Niklas. Yes! Okay, I am gonna steal your name. I think that's what I'm gonna call it for now, Riffs and Rambles, because this show, being the first episode for anyone that does listen to this, will be basically me talking to the most interesting people I can find. I have been incredibly blessed and fortunate in life.

You can use riffs and rambles like we called our conversations before.

to just meet and become very good friends with some of the smartest, most idiosyncratic, like novel, interesting thinkers in the world, in my opinion. And I realized that we often have fantastic conversations, you know, one on one, at the bar over drinks, whatever, that I think would make great listening material for people that are also thinking about the issues that we're thinking about. And to be honest, this podcast is a forcing function for me as well. You and I talked about previously, Niklas.

I have been wanting to do more publicly and kind of t you know, talk more and get engaged more generally in the kind of cultural milieu. So this gives gives a forcing function for that. And then if nothing else, if no one ever watches it, that's totally fine because it also is a forcing function for me to catch up with and hang out with friends. So if all I get out of this is me and Niklas get to talk more, that's a win. so riffs of rambles it is. And Niklas, I'll let you kind of introduce yourself more fully, but for for those of you that don't know Niklas,

Niklas is an amazing man. he is the founder and CEO of Infinita and he is working to make the world a healthier and freer place through advancing biomedical and and biotech innovation both in policy and in the actual hard scientific fields. so Niklas, feel free to, you add or expound on on anything you want to there.

Yeah, thanks for the for the generous intro. Great to be on. we had already a couple of conversations on my podcast and always fun with our conversations in the bar, very aligned on our goals in life and what how we want the world to look like. so I definitely have been through partly through your work in Prospera being kind of brought into this orbit. Like I was a founder before already, right? So working on some

Software-related stuff, B2B, and have been a startup operator before for five years in a company that did large-scale data collection. I'm originally from Germany, right? So that's where I spent most of my life and grew up. but before I entered the world of startups and entrepreneurship, I actually du studied public policy and economics, right? So I started my career.

Like unpaid internships in Washington, DC, at places like the American Enterprise Institute and CSIS, and went to the Maxwell School for Public Policy and Public Administration in upstate New York. So I kind of always been drawn to America for a very long time, and that was kind of my entrance to it. And since and until now, with my current company, you know, we're focusing now entirely on the United States, starting in Montana, basically building a faster pathway.

For biotechs to get early clinical evidence. Right. So that's basically we got involved with Montana through the SB535 law, which is based on right-to-try-laws but goes beyond them and improves them in many important ways. And our goal is just to make biotechs move a lot faster. Right. So as you know, we originally tried that in Prospera, and I've been there for four and a half years, and I'm still a proud Prosper resident. so that was kind of our we won, and we had some notable successes there, right? So we had

you know, gene therapy, stem cell treatments there, whatever. But the question always came up, like how do we we need medical infrastructure and we need eventual FDA recognition. So that's when we eventu pivoted to the United States, but seeing prosper is kind of what provided us with a launch pad to get these ideas out into the real world.

Awesome. That's a a fantastic overview. And for anyone that's kind of curious or wants to learn more, where should they go look, Niklas? So I wanna make sure people can dive deeper into this if they want to.

So you can find my podcast, the Stranded Technologies Podcast. you can also find info about Infinita on infinita.co. And we also formed the first ETRB in Montana, that's a private review board on Montanaetrb.org.

Awesome. I love it. you should know something. I haven't even told you this yet. I'm not gonna name this specific office, but I was at a speaking with a senator chief of staff yesterday in DC and we were talking about, you know, special economic zone related stuff for obvious reasons. and he I was trying to get him to understand the concept of what I was talking about, right?

and he was like, so this would allow like the kind of stuff like they're doing in Montana for the, you know, experimental treatment for for people that can't get access. He knew about it and it was his reference, like conceptual anchor point to get to what I was talking about. He has no idea that you and I are doing the Yeah, he has he has no idea that you and I like know each other so well and are working on this stuff generally together. So I like, that's hilarious. I just told him, yeah, I know some of those guys. ex yeah, that's a good analogy. So it's definitely it's definitely working. I love it.

Amazing.

Yeah, yeah.

Yeah, yeah.

awesome. Okay. What I want to pick your brain about, Niklas, and kind of discuss and and we'll we'll joust about it. But I think it'll be super fun. so I wrote a screed, effectively. I'm very comfortable calling it a screed, against effective altruism. and let me explain a little bit my motivations and then what I think I'll do is just give like a I'm not gonna repeat the whole thing. I'm gonna give a super high level summary of my argument, basically. and then we'll dive into it. I wrote it because I went through a journey where I went I've you know been

reading and kind of part of online the rationalist cultural milieu for a while. And for those of you who are like, what are you talking about? What is effective altruism? What is rationalism? Are you people insane? I'll I'll I'll define everything in a second. It's niche online subcommunities basically that have suddenly become very real in the world. so I went through this intellectual journey of you know exploring existentialism and then specifically you know rationalism

and that kind of whole milieu of thought, which obviously leads directly into the effective altruist kind of milieu. I read a bunch you Will McCaskill's work, Peter Singer, and some of the other kind of Toby Ord, some of the luminaries of that entire school of thought. and found myself often coming to this funny conclusion where the ar the the logic of their arguments would make sense like in and of themselves, so they were internally logically consistent.

And then I found the conclusion always prima facie absurd. Or not always, but a of the time, like prima facie absurd. which was confusing. And I was like, all right, I have a just a general heuristic, which is if my logic feels sound and the conclusion is absurd, I have just given myself a reductor odd absurdum and I've I have messed up somewhere. Like one of these premises is wrong, something is wrong here. and so that's why I kind of moved away from that entire mo you and have am on, you know, you know I talked about this last time, a little intellectual journey of my own.

but my point here for effective altruism is I have come to find effective altruism very specifically as a kind of meta-ethical philosophy. bad. I intentionally called it evil on purpose because that would make it more interesting for people. But I do think the conclusions are evil that people often come to. And I want to separate out by the way, just for the purposes of this discussion. I have many, many rationalist and effective altruist friends. They will hate that I'm lumping them together, but there's a reason I'm lumping them together.

the very high integrity, smart, well-intentioned people, like literally just the philosophy is wrong, I think, and then it leads to wrong conclusions. That's it. and I have been there before. one problem with being autistic is you will follow a logi an argument to its logical conclusion, good or bad. you will bite the bullet proverbial pro proverbially. So a brief overview of what what kind of we're talking about here for those of you that are unawares. Rationalism I would define as something like

using the the tenets of rationality, which for this particular discussion are are primarily related to kind of utilitarianism, consequentialism, and Bayesian Bayesian reasoning, to come to conclusions about what is and is not ethical, what is and is not good behavior effectively. Again, I know I'm painting with a broad brush. There's many very interesting thinkers in this field. We have to oversimplify or else this podcast is going to be eight hours long. So that's the overall milieu and there's a specific group within that rationalist world.

that call themselves effective altruists. Okay. The original core idea was both noble and valiant and correct, which was hey, if we are going to do altruism with well, charity specifically is what they were targeting, you know, philanthropy, if we're gonna do philanthropy and charity, we should probably make some effort to learn which charities, which charitable efforts, philanthropic altruistic efforts, actually work at reducing suffering and improving people's lives and which ones don't. Now you might be thinking, yeah, duh, obviously

But that was like a n d brand new idea when when the effective altruist came up with this. Before that, the entire kind of philanthropy industrial complex was basically just like sinecures for like elite th you know, elite aristocratic class people who were not good enough to compete in the market. And

Vetro signaling, right?

Exactly. Virtue signaling and they work good enough to compete in the private market, so their parents would give to a giant philanthropy foundation where they could get paid five hundred thousand dollars a year to do absolutely nothing but feel good about themselves and virtue signal all day, every day. so it was a horrible, horrible enterprise before then. so effective altruism has done an amazing job saving hundreds of thousands of lives in Africa and elsewhere, you know, with things like bed nets, etc. Although I'm gonna come back to that because that backfired spectacularly. and yeast don't talk about it enough in my opinion. but they did a great thing there.

Batnet's backfired? Interesting.

yeah. yeah, I'm gonna chase this tangent for a second because it's fascinating. So they basically did a bunch of research and found out, hey, if we give bednets to people in Africa, it will stop mosquitoes from biting them at night, which means far fewer of them will get malaria and other mosquito-borne illnesses, more people will live. That was in fact the case. they gave out a bunch of free bednets in Africa and the malaria rates measurably dropped. it had a positive impact.

But the law of unintended consequences reigns supreme always and incentives rule everything around me. Well, do you know what else bednets are very good for, Niklas? They're good for fishing. Turns out they make perfect fishing nets. And the bed nets that protect people from mosquitoes and malaria contain a variety of chemicals intentionally to ward off or otherwise kill the mosquitoes. So guess what starts happening? They overfish, you have a tragedy of the commons occurring, because obviously there's no property rights for the fish.

and the chemicals in there started destroying the breeding conditions and the hormonal and chemical environment for some of the fish that they were using to fish these beds for. The downstream impact of which was a lot of starving, a lot of people losing jobs, going out of business, entire areas just getting completely overfished. So at best, no one's done to my knowledge the like detailed analysis of, you know, in EA fashion the cost benefit here. but they're they're kind of

My point here, and this is very important, and actually one of my key points, their their shining example, in fact, is a wash at best.

Have you seen the piece? Mm mm.

Right! Dude, there's a Yeah, Mike Solana also talks about this a lot. Like we could just gene drive them to extinction and nothing would happen. Like we could just get rid of a problem. I that's exactly what we should do.

Yeah, maybe there's also unintended consequences, but you know, it's a robust case and the benefit of it is definitely extremely high, so worth considering.

Totally. That's exactly right. That's exactly right. So that's where effective altruism started. A fantastic kind of nugget you know, core curdle where it began. and then my fellow autist got a hold of it and it went off the rails, in my opinion. so first of all, this whole kind of project relies on utilitarianism and consequentialism, which I will separate in a minute, but for now, just for the purposes of walking through it at a high level, I'm gonna just combine for a minute. because they are two different things, but they're very related.

what this refers to is basically saying, okay, there is you we can think of this concept of mm happiness, fulfillment, something positive affect, if you will, enjoying your experience as as utility. Okay. So utilitarians think, well, if we can convert that somehow into a number, then we can just add and subtract, you know, what various actions that we may or may not do will add or subtract utility to ourselves or others.

And the ethical principle is very simple and straightforward. You should do things that increase utility and stop doing things or work to stop things that decrease utility. Okay? later kind of developments in the field, from people, you know, like Peter Singer, Parfitt, or a bunch of awesome philosophers have written this. They focus much more on David Pierce, who followed me after I wrote that article, which I was very proud of. I've I've that's that was that's one I've wanted for a long time. basically started focusing instead of on

Increasing happiness, so increasing the number of of utils in the world, decreasing suffering. Okay. because the argument I think it's correct by the way, at least this this sub point definitely is suffering in terms of like the magnitude of how bad it affects your subjective well being and experience, is much worse than adding more good things to your life that make it better effectively. Like suffering is worse than joy, is their argument in a nutshell.

which like neurochemically I and and neurologically we can prove it is is true in a number of ways. and even just at on an intuitive level, right? Like if I told you would you rather me s you know steal ten thousand dollars from you or give you ten thousand dollars? Well it's very obvious that that you know you would rather not be stolen from than you would be given this like the the idea of having that money stolen from you just hurts way more psychologically and emotionally than just getting the random gift, right? So that's the the kind of core argument there. So so now we've gone from kind of increased utility to

Actually, don't increase utility. We should instead focus on decreasing suffering. So stop the things that decrease utility. Okay. and then there's two more kind of very important premises that we add in here that result in what I think is effective altruism having gone completely and totally off the rails. the first is how do you do the accounting for the future and for now? So what I mean by that is when we're looking at the world and we're trying to decide, you know.

and by the Niklas, I'm gonna turn this over to you to fix any. I'm trying to steal man as good as I can, so fix any of the you know misstatements here. you know, we're looking at the world, we're trying to decide what is and is not ethical to do. Like what should I do with my life, right? As a good utilitarian, as a good, as a good effective altruist. well when I'm doing that math, because it's always math, to figure out, you know, what's gonna net reduce reduce suffering the most, how do you deal with the future? Traditionally, the future is discounted by some discount rate, right? Which is a percentage discount effectively. Why?

Very simple reason. We don't know the future. No one does. And it's very, very hard to predict the future. even the best super forecasters like are decent at it, but you know, they still miss a lot. It's just very, very hard to know what the future holds. So because it's hard to know what the future holds, traditionally people and we do this intuitively, we discount the future to some extent. Well, a lot of philosophers have argued, this is a core premise of effective altruism, but they've taken a hole, is that you there should be a zero discount rate on the future.

In fact, actually while I was doing research for that, Niklas, I didn't realize Tyler Cowan had even written a paper about this a while back, peer-reviewed paper, that there should not be discounting of future people. and we you can expand on that if you want. so that's this is a core point though. Basically you should treat all possible future humans as roughly equivalent to in your moral math, in your ethical math, currently extant people. There should be no discount rate for the future. Okay. And then there's two more moves that are important. One is expanding the suffering calculus to include not just humans, but anything that is conscious.

That has what philosophers call qualia, like a a subjective experience, right? and then what they call long termism. So if you t if you accept the premise that there should be no discount rate on the future, well, odds are there are going to be, you know, if we don't extinct ourselves, hundreds of billions to quadrillions of uncountable, you know, ten to the thirty power, like conscious minds in the future, not just humans, but just conscious minds capable of suffering in the future. Well, what that means is you become what's c what Yes call a long termist, okay?

which means your current actions, what you do now to be maximally ethical, should be weighted by what is the marginal impact of those actions on the well being of the entirety of the future quadrillions of of the shorthand they use, which I I think is actually a great turn of phrase, is like you the the future of the light column of consciousness, right? Just consciousness writ large. And where they end up as a result of that is now here's the core problem. Now you have an infinity sign in your ethical math calculation.

Now, EAs will argue it's not really infinity, it's blah blah blah, whatever. I for all practical purposes, it is infinity. Which means anything you do right now has infinite, effectively, for all practical purposes, infinite ethical weight. And there's a bunch of other premises I'll skip over. They come to the conclusion basically that that therefore the most important thing we can do is prevent existential risk. What's the biggest existential risk according to EAs right now? AI. So where do you end up? You end up with a philosophy that now holds the belief.

That the only and most important thing they could ever do is stop AI existential risk, which is very real, by the way. I I I I'm I'm gonna write about this tomorrow, but I'm not discounting that. It's definitely real. not not in kind of in the way they think about it, but it's definitely real. So they argue, you know, AI existential risk to kill all humans, kill all conscious life in the future is very high. Therefore, everything they do today should be focused on just stopping that. And importantly, there's no limit, effectively.

on what they can do to stop that. so I wrote a long kind of piece fleshing all this out in much more detail. if anybody wants to read it, they can go dive into it. Niklas, feel free to kind of push back correct it if that I'm trying to steal man first so that then we can discuss the the ethics around it.

Yeah.

Yeah, sure. So many pl things to address. just so you were kind of my own priors. I'm not a consequentialist, I'm a weak deontologist, right? So that basically means we have prima facie rights and there is a prima fascia, hey it would be wrong to sort of, you know, kill someone to steal their organs, for example, to treat five people. Right. There is a presupposition that you shouldn't do what's wrong, even

if there are so but weak deontologism basically means unless there's a very strong override. Okay, you shouldn't do that with like five people, but with like a million or like whole humanity, right? So you can't escape that there is sort of a moral weight of large numbers and large amounts of suffering. So I'm not rejecting at all the exercise to quantify suffering or sort of quantify utils. I think my position is actually that now I'm you know I'm going

along with some of the thought exercises around animal welf welfare and things like that. Right. And it's kind of we can talk about at some point. but I don't but of course there is a point at which you end up at, and I agree with your approach there with kind of ridiculous conclusions, where there must be like, hmm, there's something wrong before maybe, right? So the example we can discuss there maybe is the insects piece that you quoted.

Yes, I would love to use that as the four that was the whole reason I wrote any of the thing I wrote.

Exactly. Right. So I do I'm not whole I'm not an EA. I don't but I do stun I do not think it's evil. I I think it's in fact possibly not good, right? And it's in fact better than sort of more mainstream sort of ideologies and like especially like mainstream civics and things like that. And I think sort of the project to improve over philanthropy or whatever, could be I could we can end up saying it's not good.

I would critique EA as being ineffective in many ways, potentially, like for example that with the bad nets, and then not convinced that all doing all these randomized control trials for all these interventions in third world countries is getting us much bang for the buck. But I sup support that at least we collect the evidence, right? So then kind of much of my contingency of how net good or bad EA is or continues to be is around how honest are we with the evidence. So your peas about the bad nets is interesting and instructive there.

one thing I would just want to put out there is I always and I'm also in a rationalist adjacent circles. You know, we discussed a lot with Scott Alexander, and he's had a very strong influence on both of us in our careers. Right. So in fact, he may have brought us together through his article on Prospectus on Prospera. And I'm I'm a huge Scott Alexander fan. And for some reason, I've always had this total befuddlement.

Hundred percent.

With Eliza Yutkowski. Like, I don't get him. Like, I was like, anything that I read that's supposedly supposed to be important or have a big influence. I was like, what? I mean, this sounds like a bunch of garble to me. Right? I I didn't f wasn't able to really finish or, you know, get anything out of anything he ever wrote. Right? I find them totally confusing. Yeah, yeah, yeah. and I find sort of the AI Dumerism case to be

Same.

It's like rationalist rationalist Heidegger.

You know, closest to evil as you can get. Right. So, you know, I think there is a point to debating sort of AI risk or whatever. and I do agree with you that's sort of coupled with some of the sort of radical EA style consequentialism that can lead to evil conclusions. Right. So I would give it kind of the benefit of the doubt, right? Because to call something evil, right? So if you look at examples of what we'd call evil ideologies in history, the bar, you know.

Correct.

Should be a bit higher, I'd say, right? So the bar should be something like: all right, so you are organized around this ideology, right? So it's not just a sort of wild mix of, you know, rationalists, effective altruists, long-termist, AI doomers, and the coalitions are kind of always shifting. There should be right, there's an organization around the principles of that ideology. and then you have people doing bad things with that ideology.

Right, so that should be a pretty high bar to call something evil, right? So, you know, Nazis qualify, communists qualify, Reg jihadists qualify, right? So it is a bar that catches what we would want to call evil ideologies, right? So another problem you kind of get into is well, could you call Nazism or communism evil kind of before they do the bad things, kind of to prevent them doing the bad things? Yeah, I think there's a point to that, right? but I also think that kind of goes in conjunction with evidence.

Correct.

Of how those kinds of groups actually behave even before they get into power, right? And are they doing evil things? And they were kind of signs before, I'd say, right? Sort of, and they were kind of wearing on their sleeve what they're tried to do, you know, especially with like Hitler or whatever, set out his program pretty clearly before of the evil things he wanted to do. And yeah, okay, just take him at his word. This guy should no get nowhere near power. And anyone who sort of endorses that guy and what he's trying to do should get nowhere near power, right? That's evil.

Right. So you know, so the case for me is not that strong that EA is even remotely sort of in that category. I would focus kind of more on the dangers or in the base sort of an AI doomerism, right? Because I think the biggest existential risk is actually that we hand over this powerful technology to sort of a totalitarian one-world government. And I think the AI doomers are actually most increasing that risk without what they do. Right. So they're most like, all right, let's regulate this or whatever.

Correct.

Correct.

And that's kind of what we know we have we have the most proven like this is how we get existential risk. Like all the nuclear weapons and sort of any military technology that was deployed and had the potential to wipe out Yemen Dio still has was furthered or done by governments like that exercised control over it. So that's where would say like, okay, this is kind of the the biggest risk or the biggest evil that I would address.

One hundred percent.

And you know, two year piece, you know, there's lots of things to address, right? So I I would you know, you probably don't agree with your entire piece anymore, I gather.

Well, I mean you and I both, I'm sure you've experienced I know you have in fact 'cause anybody that thinks a lot does. I'll write something and then get a lot of very interesting like pushback, feedback, responses from it. And twenty four hours later I'm like, I don't agree with that paragraph anymore because this guy made a really good point that like this is wrong.

Yeah.

Exactly. I mean there's a couple of problems, right? So I I think some of your premises and how you describe it is wrong or directly contradicted, right? So you say, for example, in the beginning that the altruism is sort of the highest moral good. no, it isn't, right? It's increasing utility is the highest good. Right? And altruism is kind of just a means to that end of furthering that utility. It's not necessarily the only one. And also McGaskell and others

Don't really demand sacrifice. And they sort of explicitly address that, well, you don't have we don't demand sacrifice, right? So they're very clear about that. And sort of a so you know, and there's things like that, all right? Is a EA the same thing as consequentialism? Does Yukowski speak for EA or whatever? Right? So, you know, I'm fine with sort of giving it a more spin that invites a bit of controversy, but I think the kind of reasons to try

That you chose for that controversy, sort of, you know, the Bach argument, Yukowski, and also the moral concealment just don't seem to me like the strongest cases, right? So I think the strongest case would be a you know, in general, I would have the, I would make the argument, I I would have written the piece as AI doerism is evil. So I and sort of EI is getting a bit sort of caught in the crossfire, I think. But I actually have much more benefit of the doubt that.

Yeah. That's

Correct.

sort of just to go directly into it, sort of the you know, presupposition that sort of the suffering of other species has a moral value, right? I think that's quite obvious. Right. So would you be fine with your neighbor torturing his dog?

Yeah, of course of cour of cour no, I totally agree. Yeah, yeah. Totally agree. go ahead.

Right. So there is some, you know, the it's implausible that, you know, any creature has an equal moral weight to a human. Right. So the example I always give is if someone is like, you know, torturing or abusing their cow or whatever, right, I wouldn't take a gun and shoot him to prevent it, right? so but I would also not endorse that action. I think that's really wrong. Right. So

Yep.

So and you know, in factory and farming we do that all the time, right? So we just like pa paying someone else to do it out of our site, right? And so there is, I think, it's very plausible to argue that is a very grave moral error, right? And then there's obviously things like bugs, right? So with bugs, I think we're just like in very inconclusive territory. And I I I think the bigger problem is

and feel free to prop me a bit on that and what I think about box or whatever, you know, they may not even experience pain, is the impartiality premise. Right. So I think that's much more where the crux of the issue sits for me. Like, are you required to be impartial to like your life versus someone else's life or your children or future beings? And I think you obviously you know, you it's

That's a g the point I was gonna make as well.

not wrong to be partial to your family and your children, right? So anyone is partial to themselves. And it would be an it it would be a a kind of an impossible or undue bar. And that's kind of what leads us to much of the dangers, I think, also of communism, right? Or of Platonism, is that you're required to have this impartiality vis a vis kind of your own children or other human beings or whatever. And that I think is where sort of if we identify sort of

Sort of an evil premise or something that really leads to bad conclusions, that is where it would sit for me, right? But that doesn't mean that all impartiality is wrong, right? So especially when you're in a position of power or of responsibility, then some impartiality is kind of required as part of that, right? So we have complex rules around like conflicts of interest, and that's something I think a lot about all the time. and ways in which you know we have things like

Equality before the law and things like that, and sort of equal rights in a sense. So these are all requiring mechanisms that create some impartiality in the real world, right? But at the same time, we should just take new human nature it is that we have that that we just accept that we're partial to ourselves and to the ones that are closest around us. So that's kind of where sort of the prima fascia rights would sort of override or at least put into serious doubt sort of.

much of sort of the long-termist kind of conclusions when you quantify the kind of suffering you want to reduce or whatever. Right. And I'm not entirely convinced, and you can tell me, that EA specifically has a very strong impractical impartiality requirement. Right. So I mean if I recall reading or listening to Singer, it's like, yeah, I don't donate more than 10%. I don't, you know, ask other people to do the same, right? So it's just impractical.

w we shouldn't, you know, sacrifice for others that would lead to other too much, like it would lead to other bad consequences. but you know, he's just arguing we have like some obligation, right? Especially when you're in like rich countries. You know, I think the drowning child example is still has some pretty good moral weight, right? So as long as it's like and and it's encourages us to think, all right, where's the bar for us that it's practical to help when we can, right? So when it just costs us a bit of a, right, we ruin our suit when we

When we save a drowning child's life, it should get us to think, hey

Yeah, hey, do you wanna just just interrupt you for a second, do wanna 'cause I'm sure people haven't heard of that. I mean you and I have discussed this thought experiment before, but will you lay out the thought experiment really? 'Cause it is a very good one. I love this one.

Yeah, yeah, yeah.

Yeah, so the d famous drowning child example, probably the most famous thought example in philosophy. So imagine there's a child that's drowning in a well, right? And you happen to walk by and you have a new suit, you're going to work, and saving that child would be easy for you, but you'd ruin your suit in the process. Right? So shoot you have a moral obligation to save the drowning child. Right. So the common intuition by people is like, yeah, come on.

Save the child, right? So me that means and and Peter Singer is kind of by analogy saying things that don't cost us much but have the potential to cause very great reduction of suffering, we have some moral obligation to do them. Right. And you know what he's not saying is all right, if the person doesn't do that, does another person have a right to put a gun to their head and like force them to do it? Right. That's kind of where I think sort of

How far does that moral obligation go? Right? Does it go to the extent that, you know, we have a right to coerce someone if they don't follow that moral obligation? right. And that's I think, you know, Psinger doesn't have a strong answer. But just the the example on its own, is there a moral weight? or that we should we feel a moral weight, is I think a pretty strong intuition.

Totally agreed. Totally agreed. So the Drowning Child thought experiment is one of the kind of most famous in philosophy and important. because it really does a good job of kind of crystallizing the core issues here. I'm not gonna that's a fun rabbit hole you and I should discuss one day, 'cause I have some very deep problems with that the construction of that thought experiment. But it it's for our purposes here, it's it's it's very useful. Because it it r removes a lot of the complicating kind of assumptions that make this kind of utilitarian or or moral calculus

more complicated in the real world. So I wanna I'm gonna I'm gonna respond to a few other things you said, kind of drive by, and then I want to dive into one that you said because I didn't even write about it and I think it's an interesting thing. I actually did write about this, but then I cut it because the piece was like 15,000 words long and I was like, nobody's gonna read this. It's too long. so on the first of all on the mortal concealment point for those that aren't aware and again we'll have the link to my write up in the in the show notes here

and I'm gonna it's on my personal website, it's on X, and by the time this episode's over, I'm gonna have my agents turn into a podcast. You can listen to it on the on the site if you're or you know, don't wanna sit down and read. but the moral concealment piece is basically Peter Singer and some other philosophers that are in this kind of broader milieu.

And labels are hard here, by the way, because you know, it's it's a it's a bit like libertarianism. Like no two EAs will agree that the other is an EA because they have a syndrome of small differences, right? So I'm painting with a broad brush on purpose. but basically the argument was look, if you're you know a consequentialist utilitarian and you have some goal, goal X, whatever it is, that you are like fairly certain will reduce suffering in the world, but if you were to be honest with the world about that particular goal, it would therefore prevent the goal from occurring.

Well then obviously the thing you should do from from a utilitarian perspective is what they call moral concealment, or esoteric morality, which is a fantastic euphemism for lying, effectively. or or misdirecting, obfuscating, whatever, right? the I I give some examples in the piece, but I think that

is important and the reason I included it there, this is one of my regrets, is I did not include what I'm about to tell you in there. I had I had written it but again thought it was too long and so I cut it. is I feel like I am being gaslit by some EAs I know constantly. because they my my favorite example of this, and I am going to write a piece that's just this. It'll be shorter this weekend. they constantly I feel like I yeah I feel like I'm taking crazy pills. Because they will

And have historically, a many of them have advocated for things like banning open source models above a certain threshold. You know, Yadkowski and and others wanted to supported that California bill that put a flops limit on like how many how much compute you could even use to train a model in the first place. you know, a bunch of other crazy proposals. And then like time continues and, you know, you have an entire company. the the founder followed me after this article, which I'm quite happy about. I'm gonna DM them because I think it's cool company, called Obliteration AI, which is what they do is they take the largest, best open source frontier models.

And basically remove surgically the the safety filters that keep the models from doing like cyber security hacking type stuff. they do they sell access to this. You have to like KYC and stuff to get in, which I think is good. but they sell this so that you can actually test your software. Because if any of you have tried to use Claude or GPT to write a piece of software that where security is important and you ask them to red team it, which just means like, hey, you know, we made this software, see if you can hack so we can find security holes. They won't do it. like they would they'll refuse it. in Anthropics case, they'll just kill your session entirely.

so it is a real real use case. My point here is there have been quite powerful now open source models that are roughly equivalent to or better than somewhere between Opus four point six and opus four point eight, which for those of you that are not in the weeds of AI is just like you know, the AIs of the the frontier AIs of six months ago, let's say. and they there are open source models that are at that capability level now and cheaper. And a bunch of EAs at the time were predicting a bunch of horrendous things that would happen if models of that capability were open source, and quite literally almost none of them have happened.

and I've tried to point this out before and they're like, no, no one said that. What are you talking about? We didn't say that. Or they'll they'll claim they did not they did not advocate for or discuss you know banning open source AI, even though there are Darrio himself talked about it at one point. Many, many people have in the community. So I was writing that, I I'm I I realized after I published it, like, I should not have cut that, I should have cut some other stuff because I literally just left out the evidence for why I was I had a I had a a long section on this of like quotes, like directly quote, like they said this at X date and then they said the exact opposite at Y date.

And at Y date, they said they didn't say the thing they said at X date. so th by my point there was just more that the the the philosophy because of the utilitarian calculus and totally allows for and even even supports that kind of behavior. which I think just makes the whole thing my meta point there was like A, I think that's bad, like that that is unethical in and itself. And then B, it makes the whole thing kind of hard to reason about because you're like, well, you know, I don't know exactly, you know, what

If I'm going to debate you or argue with you about your thing, well, i I need clarity on what your thing is, so we can discuss it at least. but I I take your your the spirit of your your pushback there is good. I I I left out the main like body of evidence there to support that particular point, so totally conceded there. I wanna focus on the different part that I actually did not write about much, but I think is another very good pushback on this. So I s I was discussing my article with Perry Metzger, who is a you software engineer, he's been a very prominent like kind of like anti

EA anti Doomer guy, and he he's awesome. he's a pugilist, I will say. you know, verbally. He loves he loves being very aggressive publicly. but he's he's been anti EA from day one, from before anyone knew about EA. and he made a good point that there's actually an epistemological problem at the root of EA. And the epistemological problem is that to put it in a sentence, I'm gonna put it provocatively on purpose for purposes of our discussion here, they are way too confident

in the priors that they have and how they weight those priors in particular, which are at the end of the day like gut level intuitive judgment calls. They dress it up, right, in what looks like very, you know, objective scientific sounding, you know, probability function and you know Bayesian you know probability statements. And at the end of the day it's still just like intuitive gut level judgments. And then they act like those like, well this is just objective truth. So there's a deep epistemological problem here. It ties into all of our other critiques. It's scientism.

The scientism, right, as Hayek called it.

Yeah, scientism, perfect. So it ties into my critique at a very f like root level, kind of. That's why I'm bringing it up. I just didn't have space to this is the section I the other very long section I cut was on this. because this problem of unfounded faith in their underlying assumptions of how they come to their conclusions and and how they reason about the world, they're just way, way, way too confident in them. And this ties into the long termist piece because the fundamental critique of this zero discounting for the future.

is a very simple one that I think is like a a kill shot for it. now, you know, there's there's debate on the margins here. I'm not saying you should you I'm not saying the discount rate should be, you know, infinity either, obviously. You there should be some value placed there. My point is like the future is very hard to predict. It is the bednet example is a perfect example of this. Even under very controlled scenarios where things you s you think you have a robust model of the thing you're trying to intervene in, you do randomized control trials, whatever, even still when that thing hits reality, it doesn't work the way you thought it did.

And so my fundamental problem with with this whole approach really is that the the kind of research, evidence, and kind of mathematical backing of the ethical claims that they make and what people should and shouldn't be doing, rests on a scientism. I I had forgotten about that term, but actually you just hit the nail on it. It is basically like scientism applied to the mathematics of ethics effectively. So I'd love your kind of reaction to to this overall. Do you what what do you think of that epistemological critique of the approach?

Yeah, yeah, I mean in brief and we can dig a bit deeper, you know, moral concealment works also under weak deontologism, not just under utilitarianism. It o only doesn't work under strong deontologism, right? So which is like super like Kant or whatever. So Kant said like lying is not justified under any circumstances or whatever. That's like a super strong deontologist. But a weak deontologist would say, Hey, you know, if your friend is a Jew and a Nazi wants to find them

That's a good point.

To like kill them, right? So that is kind of where the conse consequences of your lying overrides the wrongness of the act. Right. So I think sort of a case a some case for moral concealment in some cases is pretty easy to justify until you're s unless you're a super strong deontologist. So it really depends on the specifics, right, in what circumstances is or isn't it justified, right? So I think you're really overreaching a bit in your conclusions. I'm not sure.

What Peter Singer specifically said, but it's also something that we discussed like off the record. Like, are we fully like wearing on our sleeve what we're trying to do, right? When we go out and do policy? Like we're not shouting at people that we want to like about what our sort of goals are if we could get completely our way, right? So we formulate policies that are in the overton window that can reach some form of consensus. Right. So is that a form of moral concealment? You know, we're we're obviously quite radical libertarians or whatever.

Right. So how you kind of assert yourself in public is might also just be like wise to get anything done. And this might also then be just justifiable under any ideology or any ideological group that is somewhat pragmatic once and needs to get things done. Right. So how to react to that?

Agreed. th all all I'll say at this point is yeah, all I'll say at this point is you're right. and not only are you right. this is probably the weakest point that I made. the only thing I'll salvage from it, is that the and and this depending on your perspective, this is bad or laudable, I suppose.

but that there's a like effectively the shorthand I use in my write-up is there's like a formal doctrine for this almost in the like they've written peer-reviewed articles about how and when and why you should do moral concealment. And the way this caches out for them as a result of it being kind of prominent in the overall thought milieu is the type of kind of what feels like gaslighting to me stuff that I encounter regularly. So that was

It was almost an a an axe I was grinding insofar as yeah, I g I've I'm tired of being gaslit about them claiming they don't want to ban open source models when they very, very clearly do want to ban or did want to ban open source models and now it's not popular so they kind of flipped on it. but no all all of that I've

Yeah. That's kinda what I also feel I every time that I listen to a politician in like some committee or I mean

Right. Right. You're like, what you you said the exact opposite thing three months ago. Yeah, total totally agree. Totally agreed.

Exactly. I mean, moral concealment is like total common practice. I mean, I don't think mor I I'd be more on the side of like less concealment. I'm trying to be quite open about how I see the world, right? and I think that's so there's no contradiction with me sort of giving a pragmatic policy that could like satisfy different goals. Right. So some example would be when Milton Friedman or someone

says like, hey, to fix education, why don't we do a voucher system as opposed to like getting rid of the Department of Education or something like that? Like he would prefer that morally and like where his ideology would lead him to or his insights. But like okay, a voucher system would still satisfy some of the like demand that others might raise on, you know, give helping people that otherwise couldn't afford it, it would be in need, right? So I wouldn't necessarily I would kind of try to not conceal, right, but sort of be

kind of open about sort of my premises or my priors, even with the possibility that they are wrong, and not necessarily express some sort of my strongest or most radical conclusions from that, because that's like unwise. It might also be wrong, right? Because and it's not feasible, right? So not sure exactly what the EA doctrine is in that case. but to me, kind of the proof is in how they would deal with things like now you need to update your belief about Bad Nats, for example, right?

Yep, that's exactly right. no

And then there's like dynamics of like do you wanna gain more power and are you willing to then use concealment kind of to gain more power as opposed to kind of updating your beliefs, right? So that's where I would start say that moral concealment starts being wrong, right? So when you prioritize sort of gaining power over sort of truth actual truth updating.

Correct. I I totally agree. And the last thing I'll say on this, and and this in alignment with what you're saying, kind of how I operationalize this as well, is like, you know, without going into detail, I have, you know, I'm working on some policy in DC with with both chambers of Congress, and have, you know, I've met with Republican offices, Democrat offices like across the political aisle. And, you know, in both cases, I w my kind of heuristic is

I'm not going to like lie or obfuscate or get anywhere close to it, but I will use different language that conceptually maps onto or will be perceived by the audience differently to describe the same truthful thing that I am pushing toward. Right. So to take a fanciful example, let's say, let's imagine, you know, I was advocating for we'll use your your policy actually, because it's you know

I think the case is actually just c almost cleaner and more straightforward there. You know, we're talking about medical, you right to try innovation. Like, you know, once something has been proven safe, like people should be able to use it without having to wait ten years and spend two billion dollars to get through the FDA process to use it. Millions of lives could be saved as a result, lives improved, etc. You might or describe that to, you know, a a right-wing, you know, Republican politician as, you know, improving the regulatory environment around medicine, increasing innovation, increasing America's leadership in the pharmaceutical and biomedical innovation industry.

which it totally does and but that is completely true. And then you might describe the exact same, you know, underlying piece of legislation to a, you know, a Democratic staffer, a Democratic politician, whatever, as, you know, creating new opportunities for, you know, people who might have rare diseases and therefore are significantly disadvantaged to actually access, you know, life-saving treatment for those things. that, you know, maybe your policy would set it up in such a way that, you know, there's some like affordances for l you know, low-income communities to be able to access these life-saving cutting-edge medications, etc.

So in both cases, you're describing the exact same thing, which is medical right to try, and and being a hundred percent truthful about it. It's just like how you choose to describe it i is the is the piece and the kind of mental heuristics. I'm also you and I have almost the exact same moral philosophy, I'm I'm you weak deontology is a perfect way to describe it with, you know, a little maybe virtue ethics and a little consequentialism sprinkled in depending on the thing. and so that's how I approach it, at least, is like, all right, I'm you know, I'm not gonna lie about what my what my my policy does, but I'll just

tactically, of course, choose how I describe it differently. I don't think that is moral concealment, although I could see someone making a if they're arguing against my piece in particular, I could definitely see them making a case of like, hey, you're doing the you're doing the thing, just you're you're justifying it you know, to yourself, basically. Totally conceded. That's a great point. so I'll leave that one there. Feel free to react to that and let's switch to the epistemology point.

Yeah, no, I think we can go to scientism and the epistemology.

Yeah, I think that's quite there's there's something about that for sure, but the scientism critique, right? So EAs often or they pioneered the approach of using something like randomized control trials in like foreign aid interventions and things like that. Right. So you basically go to you know to Africa somewhere and you give so you take some like comparable kind of villages and one of the villages you randomly assign, I don't know.

mosquito net treatment or whatever. And the another other one you don't. Right. So in grad school we actually went into a bunch of these kinds of cases or experiment. Right. So these were also done with things like cash transfers. or yeah, so like GIF directly, for example, was born out of these kind of randomized control trials that said, all right, just give people cash instead of doing sort of complex intervention where you give them stuff. And you know, it seems to me that there might be some good evidence for that actually.

Right, sort of if you think you you know that's the right way to do things. and you know, I ultimately find that that is obviously better than what we've done before with sort of the virtue signaling philanthropy, or sort of with the foreign eight that has actually made things worse, right? So like the Bill Easterly case and things like that. So that was kind of like a bit of a I think a I a response to that, all right.

Let's figure out what actually like works. But I kind of end up always feeling a bit underwhelmed by that. Right. So, you know, my kind of counterexample or case for that is when I read Charles Murray, for example, and like his experiences in Thailand, and kind of seeing how these interventions actually work or don't work when you're like there and when you're actually, you know, understanding the community and their response to it.

As opposed to putting them into some kind of lab experiment. Right. So, you know, and this is like Goodhardt's law, any metric can be gamed. And as soon as there's money flowing, there is just a metric or a number that people can optimize for or go after. So I do think that that is kind of, you know, in these RCTs, it's done by people who are, as you can imagine, quite scientifically minded and think this way, they get an objective answer.

But there is just so many ways in which you can get the research design wrong or like overlook an important variable, right? So not everything that is relevant to the intervention you can kind of internalize in the experiment and get rid of any effects by randomizing two groups. Right. So I see that for example. just to give you one example because it's very top of mind for me. You know, the FDA approves drugs based on these very strict and super expensive hundreds of millions.

of randomized control trials. And what are they supposed to find? Right? They're supposed to find whether a drug is effective or not or whatever. And they go at very great lengths and very complex study designs and have like some of the world's best experts designing those things. And then they're misdesigning one key incentive for on the sponsor side, which is which patients do get into those trials? What is the patient population that goes into the treatment or control group? Right? So there is

A very big selection effect because of what's called inclusion and exclusion criteria. Right? So when you design sort of a new cancer treatment, then you want cancer patients, obviously, to see the effect that it has on them. But you have a like a long list of things you don't want. You don't want these patients on these kinds of other medications or kind of have other core morbidities and things like that. So you end up with lists, especially in cancer, of like,

A hundred exclusion criteria. Right? So and this is because you kind of want a clean sample. You kind of want to have one cause, one effect, right? And the effect what you're treating doing the treatment on, like the end point, is just having the cancer. So that kind of makes sense in some way from sort of you have a clean data perspective and you're testing sort of one drug, one intervention. But what you're then doing is you're systematically excluding

You systematically including healthier patients. Right? So often it's like 95 or more percent of real patients are excluded. And these are that and that's just a reflection of the real world, right? So in the real world, cancer patients have a lot of comorbidities and have a medication history and just have a lot of factors that make the data and the process to get data from them just like harder. Right. So

So what end up with is a systematically more healthy sample than the real world population. So does it actually work on the population that's not represented? Right? That's not always that clear. There's actually a huge, huge academic literature that's talking about that. But it's not really feasible to tell the FDA, all right, you need to ask for that sort of generalizability in those trials, because then they would get even more crazy expensive. Because if you wanted to get that attached to attack that on to

The process of the RCTs, you end up with completely astronomical costs. You can't do that. So you're creating an incentive basically for the sponsor for the biotech to basically have like just the cleanest possible conditions for patients that are selected, right? Because any patient that doesn't fit that or you have unclean data risks your trial not to go through, and billions of dollars are on the line for that. And there's another kind of unintended consequence of that, right? So

That is especially prevalent in phase one trials, right? So in phase one trials, sort of basic safety studies, you have a phenomenon of what's called sort of professional trial trialists, right? So there these are people that make an income with those trials, which is fine, right? I mean, these are healthy volunteers. but what do they do to not be excluded? Sort of in the pre-questionnaires, where they lie.

Right. So they know kind of how to game the reviewers. They know the questions. They've done it many times. So they're lying, right? And that leads to bad data. So my point is, you know, the y you shouldn't have too much faith, even in sort of experiments that have been very expensive, very well controlled, and sort of check a lot of these boxes how we get to scientific objectivity. Right. So

sort of if you properly have a proper scientific approach, you you need to be, you need to have sort of a very low kind of confidence in any individual study and things like that. But I do kind of feel so the question is like, so this was a kind of a long ramble, how scientist is EA? I don't know, it seems quite scientist to me, right? So I do feel most EAs are kind of autistic about these things. But also to give one counterexample.

I think Scott Alexander, you know, he has been one of the most one of the best kind of epistemicists in that regard. And sort of pointing out some of these like major flaws, like beware the man of the one study or whatever. Right. So I'm not sure that's kind of a strong verdict of the ideology of EA as a whole, right? Maybe in how it's practiced.

And kind of whether it's a net good or not kind of depends on how they're updating their beliefs based on, you know, some of the things that you talked about with the bad nets and whatnot.

Yeah, the kind of medical example, I'm actually glad you went to the detail you did 'cause it you're it's like you know, near and dear to your heart and it it paints the picture super clearly. there's also like in this discussion as well, so for those of you that are are not aware, the reason Niklas immediately went to R R C T's, which are randomized controlled trials, and why we're focusing on that for the scientism angle, is because that is held up by everyone correctly, as like the gold standard of empirical evidence for for stuff. But

They because it's the gold standard, they massively overweight the conclusions that can be drawn from it. because even the gold standard, the world is just complicated and messy. And even the gold standard is a far cry from, you know, the perfect ironclad evidence that a thing's gonna work the way it did in the in the experimental trial, effectively. I actually I wanna double click on the medical one, because the broader point we're making here is actually best made in that field because it's such a we have so much data and there's so much been written about it. So I'm gonna I'm gonna kinda stay on the medical for for a minute.

One thing that always bothers me is there's two things. First in the study design, and this is true of all you know, study design in just like the sciences, social sciences, hard sciences, whatever, when you're doing messy things in the real world. You know, it's not physics where you can literally control almost every factor with experimental design. You know, you do stuff in a vacuum chamber, etc. There's no and you can replay things, you know, thousands of times. you can smash two atoms together thousands of times to see what happens and therefore you have good data. in the real world, the in

It it's a point others have made, but I feel like people just kind of acknowledge it and just keep doing what they were doing anyway. In the real world, you don't have the counterfactual. Like you don't have the like, we could run this experiment, we can roll back the clock of time, literally, the reverse the arrow of causality and run this exact experiment a thousand times to get good data. That's physically impossible. and so when people analogize these type of, you know, social sciences or like in the real world with real human beings, you know, RCTs with that kind of thing, they

f kind of forget or or or kind of i it's no longer top of mind how much is lost by the fact that you can't literally repeat an a one hundred percent identical in every way that is possible scenario, you know, hundreds or thousands of times to then get a very clean kind of data point out of it. So that's the the first point I live in reaction to. And then the second is as it relates to again in the medical world kind of experimental design. Even when you do have like a perfect gold plated RCT, actually there's two sub points here. The first is

It's not even p hacking. Russ Roberts, who's the host of Econ Talks, one of my favorite podcasts, calls this the and somebody else came up with this, but he just talks about it lot, the curse of the third variable. And what he's referring to there is usually in a ex in a design, he's talking about economics, but this applies to every you know study design. You have your you know independent, dependent variable, right? Like the the thing you're measuring and then the thing you're changing to see how it impacts the thing you're measuring.

and you usually within that, you know, you have your a control group where like you have your intervention, the thing you're measuring, and then you have as close as you can the exact same thing happening, whatever it is. Like say it's cash transfers, right? You have your group that gets the cash transfers, and then you try to find as similar a group of people as possible and don't give them cash transfers. You measure the outcomes for both, and that is your your dependent variable, right? but what I'm referring to like here with the hidden third variable, the hidden third factor, is there is often a thing that you did not anticipate beforehand or you or

You did anticipate, but you simply can't model quantitatively, so you just ignore it, the third variable that has a an as large or larger causal impact on your dependent variable than the independent variable you're measuring. So it is almost like for a lot of these, not for all of them, but for a lot of this kind of scientism approach, it is like almost prima facie impossible to draw any causal conclusions, even if you have a perfect study design with, you know, differences in differences, you know, RCT, whatever, like the cleanest intervention possible.

You can draw some concl it is illustrative, right? It is it is directional evidence, is the way I like to think about it, and interesting. but you cannot, and I don't care if you have a hundred studies saying the same rough thing, you cannot draw super strong causal conclusions out of it. Because what you don't know, and and what no one can know, is that while the researchers were designing the experiment, collecting data for the experiment, what other things came up, what other third variables, that third factor?

That they set to the side because it makes the story too complicated. It renders a null effect, which means it's no longer publishable. This gets into the whole p-hacking kind of discussion, which you can feel free to expound on, Niklas, for for the audience that doesn't isn't aware of how that works. and my point here in all of this is like even taking the gold plated standard, which is like medical trials basically, there's so much going on that we don't know about. Placebo effects. Placebo effects are insanely strong.

And in in fact, in some interventions, the placebo effect is stronger than the actual pharmaceutical or medical intervention. and so there's just so much we don't know, and and that boils down to we don't have a good working model, right, of exactly the every single mechanism going on in the body and in in medicine. and people basically what what in my mind, the way I think about this at a higher heuristic level, is we basically looked at physics and everyone has physics envy. Everyone was like, physics gets these great results because in physics

You can control like n not everything, but you know, ninety-five to ninety-nine percent of the variables. Like you you genuinely can get very good coverage of controlling the environment to just measure the change of the one thing that you're curious what it does. And you can repeat it. Neither of those things hold basically anywhere else, and they don't even hold in physics sometimes. So i you know, I'm not saying we don't know anything. We obviously do. The modern world around us exists, but I think the real tests for almost anything.

is when it meets the real world. When when the the engineering is applied physics, right? Is another shorthand I like to about this. And so if something from physics gets applied to engineering, like this is why like we know special relativity works. Set aside all the papers on it and the experimental design. You want to know why we know it works? Because GPS works. Because GPS would not work if special relativity that was not accurate. Because it was built on special relativity. So I I think you know what matters is the real world. What matters is what we get to by way of analogy, right? The the engineering, which is the applied physics

That then tells us for real, all right, what works, what doesn't, what do we actually understand about this, what do we not? Is our model accurate and and does our model cover the things that matter or does it not? obviously this this overall relates to the the EA thing because the root foundation of all EA calculations or any utilitarian calculations about things you're considering doing boil down to the evidence for and against the how you're waiting, what the impact of your intervention will be. But me and Niklas are both saying to put a really fine point on it.

kind of provocatively on purpose, is that that foundational evidentiary layer actually has some severe problems that render this entire approach not like not wrong, not trash. Again, the EAs have done some very good work. I agree the the philanthropic approach, though the the specifically the charity philanthropic work of EAs is just I think in my mind unambiguously net good. It's you know, this is way better than the old model of philanthrop philanthropy, obviously. My point is, they

come to wildly sweeping, provocative and conclusions that have profound impacts across a wide variety of of domains in society that I think at the very foundational level, the evidence just does not justify the level of consequence that they that they apply to these things.

Yeah. And maybe let me, even though I overall disagree with the piece, sort of partially agree on one thing, which is not exclusive TEAs probably, but there's a bit of cowardice. Right? So you know, it's just much easier to be like some researcher in your chamber designing some randomized control trial that some other field researcher then has to run in Africa.

compared to what I think is actually the real alternative that actually lifts people up, which is entrepreneurship. Right. So, you know, my critique is EA is ineffective in their approach because they're not doing what is actually effective, which is entrepreneurship, and sort of creating the institutions that make entrepreneurs thrive and flourish, right? And lifting yourselves up and creating employment and prosperity in others in the process. Right. So not go too deep into that.

But that's kind of what when you have to when you're an EA and you hold yourself to the standard of you're looking kind of at the world objectively and what works better and what doesn't. That should be your conclusion. Right. But I do feel that is I there's I agree with a bit this cowardice, right? So it's just easier to pretend that, you know, you don't have to go sort of the hard way and actually go to move to these countries like you know, I and we did in Honduras or people little places like that.

To actually work with entrepreneurs in helping them sort of lift themselves up and things like that, which is really hard work. Sort of you have to forgo comforts and you have to sort of get your hands dirty. You're in the real world, right? So that's kind of where I would say, all right, this kind of criticism is justified. It's justified in general of I think of scientism and of many bureaucratic and public institutions, or whatever. so let me kind of double-click on what I think is the sort of

The the sort of history of science or philosophy of science points that sort of leads to scientism or solidifies scientism is institutional sclerosis, right? As Manchester Olsen calls it. So I'm thinking of like Thomas Kuhn, how science changes through paradigm shifts. Right? Terence Keeley has also written about the myth of scientific objectivity, right? So

Science is done through institutions and embedded institutions, and institutions can get sclerotic and ossified, especially when they sort of have too much power or have a monopoly. Right. So and much of the scientific funding institutions like the NIH or FDA, they're sort of in that sense the high priests of scientism, right? Because they're virtual monopolies and they can prevent others from doing what they're doing, right? So they have a legal monopoly over these things. So I, for example, if we in Montana

Started to do like randomization of patients, that would be that would be under federal jurisdiction. Like we we're not entirely sure, but probably you know, be you know, breaking the law, right? So we can't do that. But we can do this is just a side note, but you know, in Montana, this is like within medical practice law, and courts have affirmed that repeatedly, that that's fine and sort of collecting outcome data. so we're literally kind of prevented from

Sort of if it's a good thing, what the FDA does, and I believe in many ways it is, sort of getting safety and efficacy data, I think that's great. I think the more of we have of it, the better. But if you're claiming sort of that monopoly and not letting others sort of try different approaches to come up or to achieve kind of the same goals, then you get ossified, then your metrics get gamed. And then you're leading you're you're

Doing things for the appearance of safety as opposed to actual safety. Then I kind of find myself very c very often, much of these critiques of hey, the FDA is doing this, but it's actually not giving us representative efficacy results. That's a long-standing academic literature. Right? So I do think there is independent thinking, even in science and in academia, but I think there's a bit of a lack of imagination and a lack of courage of what it takes to change these things.

Right. And that's as a result, like very hard, right? So more of these kinds of functions. So that's kind of what we as libertarians are often kind of like like the problem with the state or with the government is it takes over the things that are desirable and good, like education or healthcare. Right. And that's kind of, you know, they see it working and then they take it over because they're popular with people by providing them, right? And that's kind of where the problem sits. Like it's not

it's hard to argue that the government is doing a bad thing. They're not, right? I mean the education and healthcare is good and sort of testing new drugs for safety and efficacy is a good thing. Right. It's just like we have to make a much more complex argument that the way they're doing it is kind of not a good isn't it's not the right way and doesn't preserve what we think is actually leading to innovation and on sort of continuous improvements in actual science, which is sort of the Hayekian decentralization.

Right, sort of the using kind of local and dispersed knowledge and sort of having sort of residual decision mak individuals as residual decision makers, right? Sort of not beholden to decisions that experts make for make for them. Like they can make bad decisions, but it's easier to correct bad decisions, right? And it's easier to create institutions that help individuals make better decisions in a situation like of market mechanisms and processes as opposed to sort of public or legal.

Monopoly of the institutions that are doing the same things, such as the science and things like that. Right. So again, I think the when we look at EA there through that lens, sort of I think they have some of the same failure modes of sort of government-funded science and foreign aid and things like that. But they're also typically not government-run, right? And they're kind of are deliberately doing it through philanthropy and through private funding and through like Dustin Moskowitz or whatever.

So I think that's also if we are very critical of publicly funded science and government run science, that's also more of a point in their favor than not. Right? So

Yeah, I I agree. That makes sense. kind of in in broad strokes, but I guess my

Wha it always leaves me in the space of like I would just call it very deep existen epistemic humility, basically, is where I land on like a kind of all of these very important epistemological problems. This is not just an EA, you know, utilitarian this is just writ large. Like the world's complicated and messy and really difficult to like wrap your head around fully. it's like the opposite of this one of my favorite quotes of all time from a mutual of mine, Iganrobot on Twitter.

Every problem in the world is monocausal and is caused by whatever particular thing I'm obsessed with this week, right? I feel like that's the the kind of operating mode of of a lot of people. And where where I kind of land on all of the kind of epistemic problems with with EA and with scientism, it's just they they need a much heavier dose of epistemic humility, really. They they the the the conclusions they draw from these things are just just too strong. I think you're right, by the way.

that a lot of this scientism and the rest of it is actually a result of the the sclerosis of the scientific and and research institutions as a whole. because if you imagine a counterfactual world, I have actually thought about this a ton, but now that I'm thinking about it, it's kind exciting. If you imagine a counterfactual world where that's not the case, where there's a lot more competitive, you know, dynamics and market dynamics amongst, you know, scientific and research institutions themselves, that there were much more kind of dynamic, innovative enterprises in and of themselves.

then you end up in a world where I think like a just a lot more stuff gets tried. and what I mean by that is we go to back to my engineering analogy, right? Like SpaceX is the perfect example of this actually. You know, the NASA let me contrast the two. So for those are that are unaware, the reason SpaceX has like made a private space industry a thing and is the only company right now that can do so many of the things they can do and put so much mass into orbit, etcetera, go to the moon, Mars, whatever, is because they don't take the NASA approach. So the NASA approach

was kind of similar. It's a weak analogy, but like let me stress it for a minute. it's not a good one, but I'm gonna go anyway. It it it is more analogous to I'll put it that way, weekly, more analogous to the kind of scientism, utilitarian EA approach to the world. So the NASA approach was always, hey, let's do a ton of research and let's make a bunch of models of what we're gonna build and let's all put our heads together with a bunch of like insanely smart, you know, two hundred IQ people and they are insanely smart and try to think about beforehand. Like what could go wrong? How can we prevent it? What should we build? Whatever. and then

After we spent years thinking about it and writing plans and rewriting the plans and reviewing the plans and writing the plans, then we one shot build a thing and try to fly it and see what happens. SpaceX takes a what I think is a much more like epistemically humble approach that objectively on the evidence works far better, which is sure that they're not not planning and modeling, obviously they are. But the the time between like model plan and then test the prototype of the thing in the real world is several orders of magnitude shorter.

And and scoped, right? So by way of analogy, SpaceX has rapidly iterated on their engine, the the Raptor engine, okay? And the reason they were able to do that is because every time they would come up with a new design, they just immediately throw the thing on the test stand and blow it up. And they blow stuff up all the time. In fact, it's it's not even a bad thing within SpaceX to like blow stuff up. Because that you know what happens when you blow something up?

Huge learning opportunity. Now you know stuff that doesn't work. So you don't have to do that. And you know what needs to feed into your next design. But the sooner where this all nets out is like the sooner you can make your ideas come into contact with reality, the better for the world. and I think this is also why like entrepreneurship in general is the thing that pushes humanity forward as a whole, the market dynamics that allow for entrepreneurship. Because it's the same thing.

All entrepreneurship is having an idea and bringing it into contact with reality as rapidly and intensely as humanly possible. That's the only way entrepreneurship works, right? the longer you spend in your ivory tower, in your design room, in your engineering conference room, and the less time you spend out in the real world by way of analogy on the test stand, like blowing stuff up, then the worse off you're gonna be as a whole. And it's why SpaceX is crushing it and no one else is, because SpaceX keeps blowing up starships over and over and over. It's why you hear them cheer. Whenever, you know, the recent starship flights, stuff would go wrong, edges would go out.

That's awesome. You know, the whole SpaceX crowd and the SpaceX H2 is cheering, which are like, why the rocket just blew up? That doesn't make any sense. It's because as soon as the rocket blew up, they knew, sweet, we'll be able to immediately find what caused this and fix it so it won't blow up that particular way in the future. And we'll keep blowing them up until it works. And then you know it works. Not because your computer models proved it, but because you've tested it over and over again in contact with reality in a way that proves unambiguously that it works. You may not even have here's the other thing you may not even be a hundred percent sure.

Like which parts of your design, your engineering design, are like the things that are making it work or not. But you again, because you're in contact with reality so much, you don't have to know. You don't have to know. Like, you know, I I guarantee you there's no one at Apple that can tell you exactly what it is about the iPhone that has made it the most successful phone of all time. A bunch of people have a bunch of ideas. No one knows. No one at Apple, no one else in the world, no human is really a hundred percent sure. But guess what? It has consistently been in contact with reality for a decade and it is the most popular phone on earth for a reason.

we can all guess at what that reason is, but the underlying you know important truth is that it it came into contact with reality and succeeded, so something there is working, even if you don't know exactly what it is.

Yeah, 100% agree. I mean, the EAs kind of have a bad excuse have no excuse really for not sort of viewing entrepreneurship really as what is effective, as opposed to, you know, all these RCTs and interventions or whatever. So that's my ultimate biggest criticism of them, right? So, all right, you're ineffective. Right. So better try something else. I mean, I think you can apply many of the same priors or premises.

Mm.

Minus maybe the partiality thing, all right, be per be pretty careful with that. but other than that, sort of rationally thinking what's making things more effective or not, well, you should do that.

that's exactly right. That's exactly right. Okay, I want to discuss one more thing with you before we we call it, which is kind of related to this whole world, but a little it's related sp specifically to the kind of the concealment point. and this is a thing part of the reason I'm bringing this up is because as I told you before, you know, I am not sure exactly how I like feel about this, think about this, whatever. So let me kind of throw a a half formed idea at you and see, you know, see how it lands. I

Anyone that is engaged in an enterprise to radically like improve the world usually means by definition you like very strongly disagree with some you know way the world is working right now. That necessarily logically implies that, you know, if you you know march into the offices of, you know, a politician or a decision maker, a person with influence and power, and you describe your thing the exact same way you'd describe it to your friends that already agree with you, you'll are going to like just

fail in absolutely preposterous fashion. and I I think i this is just a where I land on kind of how to conduct yourself basically is what I'm referring to here. How to how to conduct yourself in the world when you have what will occur to the rest of the world as a radical goal. Okay. Even if it's a very good goal and you know if they understood it correctly, they'd agree with it, whatever. I'm just setting aside the you know value judgment of the goal. Just say you have a radical goal relative to the baseline. Radical is always relative, right?

relative to kind of the norms now. there there are two questions here that I find are not always, but sometimes intention. And I'm curious how you think through this. on the one hand, let's imagine you just want to like th there's like success, instrumental success, like achieving the outcome in the world. and then there is the ethical and moral considerations of like how you conducted yourself, how you achieved that outcome and at what cost, right? like what you did to make that thing

True in the world. And not always, but I find that sometimes, right, these things are intention. I have personally just made a a decision that as a weak not weak, but like, you know, kind of coarse heuristic, that's what we're looking course heuristic, because this varies from, you know, inst like particular circumstance to particular circumstance, is like when in doubt. I think the high integrity and ethical thing to do is just be honest. for the simple reason that and this is almost like instrumental in a way.

So there's a deontological reason, which is like being more truth-seeking and truthful with the world. It's just the way you should behave from a deontological perspective. And it just so happens, in my mind, like there's the kind of a coincidence here, where also from an instrumental, you know, consequentialist perspective, I think for reasons of reality is messy and complicated, the more like high integrity and truthful and direct you are, and the less you obfuscate, hide, misdirect, whatever, then the better your outcomes will be.

Because you're not hiding anything, people are not going to be surprised by things, you know, you you gain higher you have a better reputation, you gain higher social trust, whatever. So there's they happen to like almost perfectly kind of coincide here. but there are like a lot of edge cases here where that you just kind of have to in my mind kind of feel your way through. And it it even comes down to this has a direct implication on even on like at a strategy level how you conduct yourself, right?

Because there's one school of thought, and this goes for both you know, what you and I are kind of in embarked upon, but also just like any, you know, large scale political or cultural project specifically. there's always like a there's a forking path here. and this is the classic i in the kind of I don't have you know, I'm not involved in the abortion debate, but they just have a good like conceptual framework around this tension. They call it, you know, abolitionism versus incrementalism to map it to abortion, right? And these are two schools of thought within the like pro life anti abortion movement that like hate each other.

the incrementalists say, Hey, look, we gotta deal with the world as it is. everybody knows, you know, we're honest with everybody, everybody knows our end goal is to like always ban abortion all the time in all circumstances, whatever, something like that, you know, the maximal pro-lex position. But our institution as a whole is advocating for tries to do things that make marginal improvements. So where this this this real fight happened, and hilariously, I think the incrementalists were just proven at least empirically correct, because again they came into contact with reality quickly.

is the abolitionist after the Dobbs decision, which overturned Roe v. Wade. And now all that did for those that don't know is it made it such that the state legislatures can make their own policy on abortion instead of it being overridden from the federal government. That's all that means. so immediately after that the pro-life movement was like, woohoo, we finally did it. We overturned you know Roe v. Wade now we gotta pass a law in all 50 states that says, you know, no abortion ever at any time on no matter what, more or less. Something like that. And there was an immediate fight here where the abolitionists pursued that approach in some states.

And the incrementalists pursued a different approach in other states. The incrementalist said, Look, there's no way in hell the public is going to accept that. It's not going to happen. That there's no legislature that's going to pass that legislation ever, anywhere, at any time. So we are going to advocate for, fill in the blank, an eight-week ban, a sixteen week ban. You know, they draw the line randomly depending on like what the politics of the state will let them do, more or less. but you know, or or advocate for, you know, well you have to approve this or that before you can do it. Or maybe it's a

What they call a heartbeat ban, right? So, like if there's a detected heartbeat, you can't have it. But like there's a bunch different places they draw the line. But they draw the line somewhere, instead of the maximalist kind of abolitionist approach. And empirically what happened is immediately after Roe v. Wade, no state adopted the maximalist position, like not one. And many states adopted some version of the incrementalist version, depending on what the state politics would allow. And in the states where this is critical, there are a few states where the abolitionists tried to go the full abolitionist route.

And as a result, nothing passed at all. and then the default kind of becomes a much more kind of permissive environment that the kind of pro life people don't like. and again, you know, I'm not making any claims on that particular policy. My point is that like what happened is quite kind of instructive, I think. You know, and th this is true of anybody go embarking on this this kind of endeavor. and and the argument from the abolitionists is well, we just have to convince everyone we're right, basically. Right? Like we just gotta convince everybody we're right.

but I think there's just like a almost a a real politique, a just like, you know, group meeting the world and working within the world as it is, where you just have to say, like, that's not realistic. Like it's not going to happen. Okay. just like for whatever reason, the the facts of reality are such that that is just not a that's not in the cards. So we need to do as much as we can to advance our goals within the realm of what is actually possible kind of in the world. and

If that means, you know, a incremental solution instead of the like maximal solution, then you know, it is what it is. but then the abolitionists have a and I'm just using this argument as the analogy because it's it's very clean. The abolitionist pushback is like, well, great, you got the eight week ban, sixteen week ban, whatever. You're never getting more than that because that was the political compromise you made to do that. And now, like path dependency, like the kind of inertia of politics is as you know is so high.

That like i anything that gets instantiated becomes permanent immediately. And that is just like the default in politics moving forward. and at the time it was maybe really hard and radical to get done. But then once it's done, what have that's the new normal? That's the default. and it makes it very hard to then revisit it in the future and say, actually we did sixteen weeks, but now we gotta go back and do eight weeks. you know, th the the response is typical, we already had this fight. This is where we landed. We're not revisiting this, right? so I it's just kind of half formed, but I was curious your thoughts on this kind of just in in general, right? This kind of

Yeah.

Abolitionist versus decrementalist approach is the analogy I'm making here, but just writ large for people embarked on projects like ours.

Yeah. So a lot to be said about that. I mean, I would kind of invoke decision strategy, right? So I think we live in a world of kind of where you can employ different strategies and there is no one strategy that dominates the others. Well, th some maybe. I'll caveat that for later. but for example, when you're dealing with an IUD log on the other side, like a Hitler or Pol Pot or Mao.

A pragmatist is not a good the right response, right? So in such cases, it might actually be good to have another sort of ideologue on the other side, right? Because pragmatists kind of try to reverse engineer the other person's motivation in terms of how they think about the world, right? But ideologists that want sort of everything don't think about the world in this way, right? They have a commitment strategy where I think their internal moral code is also different, right? Because

As per Robin Hansen, we're all deceiving ourselves about our own ultimate motives. Right. So and we have some heuristics that lead us to certain commitment strategies, partly designed kind of by our temperament or how we assert ourselves in the world, right? So and I think different kind of, if you would call it leaders, have different kind of strategic optima. Right. So, say a Stalin, for example, I don't think was like the most charismatic person.

Right. So he had to get his way with sort of deception and with screwing people over or murdering people and whatnot and things like that. Right. So versus, you know, other people kind of are endowed with like it they can afford to be more honest. Right. So you know, if you know that meme of sort of there's that woman at a work desk and then there is like a very pretty man and a very ugly pig man with acne. Right. One is saying, Hey, you're looking good today.

It's like, you're so sweet. In the other picture, the ugly guy is like, You're looking so good today. Hello, human resources, right? So I I think it also you you have a bit of an optimum of where you're good at at a leader, right? As a leader, right? And you know, a lot of people aren't very charismatic and they chose they choose strategies that are more kind of exploitative, partly as compensation for that, right? Because they can't get sort of the

Yeah, I love that meme.

positive sum convincing other people, buy in the way others can. And politics is actually full of those kinds of people. Right. So also because politics are generally or often zero-sum games. So someone who has much more willingness to play dirty is just much better suited for politics. Right. So that's not a like you should do that. It's more of an indictment of the mechanisms that politics is creating for people, right? Or of like why zero-sum games are bad in general.

For that matter. and when you think of how you personally want to assert yourself, I kind of would work backwards kind of from what your strengths are and what goals you're working towards. And you should ideally choose something that is working towards positive some goals. And even though we're working in politics a lot, I think we're doing that, right? We're trying to get like politics out of the things, right? And that's kind of the so we have to we have this contradiction that I think Kabier Millet spoke about, right? So

Right.

All right, you don't play zero sum games, well great for you. They're playing zero sum games and they're winning if you're not sort of playing back, right? But you what do we play back on towards what end? Right. So ideally to create more positive sum sort of interaction possibility and less politics that intervenes in sort of how markets are working and things like that. Which is why I find Millet more kind of inspiring than sort of his critic like Hans Hermann Hoppe or someone like that.

Who sees it more like from the purest ideological lens. you're not really libertarian because you like run a government or whatever, or you do like monetary policy and make all these like pragmatic decisions, like to dollarize the country or whatever. So I'm much more sympathetic to the Malay approach that you're going to be more pragmatic. But hey, I'm glad that people like a hoppe or the purists or whatever also exist, right? And sort of the process of criticizing is sort of part of, you know, ideally, it's part of

what would be a healthier kind of ecosystem. And let me say one more thing about sort of positive sum games, right? So I think in a world with positive sum with with less politics, with much less zero sum, and maybe it's utopian of me to think that we can sort of exercise zero sum games from the world. Maybe we we ki maybe we can't, but I think we can reduce it much further. In a positive sum world, I think cooperation is actually a dominant strategy.

And when you think of, for example, of the Robert Axelrod evolution of cooperation experiments, right? So for any listener, those are like a set of famous experiments from the 80s where they tested basically so everyone knows the idea of the prisoner's dilemma, right? And in the prisoner's dilemma, it's kind of rational to not cooperate, and then both people are kind of worse off, right? So I don't think I have to iterate the prisoner's dilemma for everyone.

but that's kind of a bit of a justification why, you know, the world is kind of Hobbsian or zero-zum or whatever, right? Because in a state of nature, whatever, people don't have an incentive to cooperate because rationally it would be, you know, you they should defect. But then Axelrod came along and say, well, what if we repeat prisoner's dilemmas? What if we play the prisoner's dilemma as a repeat game? And the real world is actually more close to repeat interactions, right? And in repeat interactions.

What kind of strategy wins? Does the defection always strategy win? No. Right? It actually turns out the most successful strategy in the first iteration of the game is tit for tat. Right? So start with cooperating, and then if the other person cooperates, continue cooperating. If they start defecting, if they start to screw you, then you should also start defecting, because otherwise you'd be just exploited. Like they would gain more from your cooperation and their defection than from both sides cooperation.

And then in in the later set of more precise experiments, they even found that tit for tit for tat is an even better strategy. So cooperate twice and then defect. So cooperate and then they defect. Yeah, cooperate one more time. If they then defect, then you also defect. But you know, giving people another chance and generally defaulting to a cooperation is a dominant or strong strategy in sort of positive some environments.

But of course, the question is always like, well, there's always a situation, and someone like a Carl Schmidt political theorist would always say, like, well, okay, great and good, but what happens in a situation when something goes wrong and somebody has to make a call or a decision? Right? And otherwise, like, right, who makes that decision in that moment? Right. So, and when it comes to things like politics, which is zero sum, it might involve things like political violence or war.

Like we can play all these liberal positive sum games in the world that we want, but you know, it just what just matters is who's making the decision at that point in time. Right. And I find that's something that's you know, a conclusion that's kind of difficult to escape from. What happens then? Right. So I don't have further thoughts beyond this, other than that is something that, you know, I keep thinking about.

Yeah, same. And I think it's one of those things that doesn't actually have like a clear the what I've I've become more comfortable with this kind of ambiguity now, it's just like, look, despite my like artistic brain wanting this to not be the case, some things are just like very situational, case by case, it just require a well grounded and strong kind of heuristic foundation to work on. A kind of like when in doubt do X type of heuristic, basically.

And to just keep as like kind of top of mind, conducting yourself with with integrity and and trying to always engage in positive sum games, and then just let the chips fall where they may. i then you're the you know, the kind of classic adage about, you know, you the you achieve your goal your objective but you sacrificed your soul, you know, it's kind of gesturing at this this exact kind of heuristic based behavior, that I think will both lead to better outcomes and also like you know, leave your conscience cleaner effectively. It's like, you know.

Even if I failed, I know I conducted myself with integrity and you know it failed on the merits, and that's okay. Like some things just don't work out, and I'll try something else next time. It's not a big deal. so that's kind of where I've landed on it overall. And what I have found, now I'm curious, I'm curious if you've found this as well. and then I think we'll we'll call it after this. I I have found that this approach, which I like quite consciously I I used to get stress a lot, right, before meetings in DC where I was like,

you know, I need to research everything about this person and, you know, make sure that, you know, I know ever like I'll be ready for any objection and answer any question. And not that I was like hiding anything in the slightest, but just like trying to very consciously work through and think about like, I need to make sure I describe it in this way and not that way. And I gotta be careful if they ask this question and how I answer it. And I don't know, like

Probably six months ago I totally stopped doing that. Like literally all of that. I I mean I'll I will, you know, do a cursory have an agent do a cursory research on the person before it just so I know who I'm talking to, you know, what's their background. That's about it, really. I don't really do any of that other stuff anymore. and I just kind of go with my my intuition, my gut during the during the meetings. And like all brutally honest, almost like to a fault. and I find that that just like works way better, for now. And this is I think a temp maybe just a

like this may not have worked thirty, forty years ago or something. But I think the equilibria we're at right now, with with regards to politics, and it's part of the reason I think you and I both have you know found faster success in a variety of ways than most people would expect is because the default kind of the current equilibrium is is deception, is obfuscation, is zero sum, is like everyone's out to screw over and get everyone else. so what that does, anytime there's an equilibrium and you like

adopt a strategy consciously to counter signal that equilibrium, well now that's a more much more viable strategy. You know, so imagine two worlds. Imagine a world where everyone is extremely high integrity and conducts themselves with honesty and transparency and integrity, etcetera. Well in that world, yeah, the payoff to being a deceptive, sneaky Machiavellian asshole is way higher. Way higher. Because no one else is doing it. So you can just kind of get past the naive people and get what you want very easily.

If you flip that and say, the default is the the sneaky lying out to get you zero sum Machiavellian world, and then as a result, you say, No, I'm not partaking in that. I'm just gonna conduct myself with transparency and integrity and be honest and let the ships fall where they may, that comes as like what I have found is breath of fresh air to people. they deeply, deeply appreciate it. And then you develop a like social and and political reputation for like

you know, yeah, this guy may have some wacky ideas, whatever, but like he's very honest. He'll he'll tell you what he thinks, he'll tell you what he does and doesn't know. You know, he's not gonna lie or exaggerate or whatever about, you know, the support he has for his the thing he's doing or whatever. and that he's willing to, you know, do do work with you and like, you know, hi accept pushback gracefully and even like enthusiastically sometimes. hilariously, that ends up being just like the optimal strategy in a world where the equilibrium is like the zero sum Machiavellian guy. So my current kind of working theory is like

Totally access 'cause I would do this either way, like either if it was the optimal strategy or not. A and in fact when I I kind of started doing all this stuff at DC, especially like, you know, six months ago when I made this conscious like effort to do this, like I honestly fully expected this kind of change in my mode of being and how I was going to conduct myself would result in failure. Because obviously everyone says you know, you have to be sneaky and Machiavellian and whatever in DC to get to do things. That's just how it works. so I fully expected that and accepted it. I was you know, that's fine, but I'm not going to I refuse to play the other game.

Right? I'm just not gonna do it. And then found out on accident, that like, wow, just like being high integrity and and honest, is t is like actually the optimal strategy by like a pretty significant margin because you become like the only person that everyone knows is kind of a node that is honest and will tell you, Yeah, I do have this. No, I don't have this. Like a classic kind of example of this. I'm gonna abstract away the details because this is, you know, it's still kind of in flight and I just don't wanna perturb it yet, but

You know, I need the support of a particular, you know, high-level US government official for a thing I'm working on. And I do not have it yet. You know, I've talked to some of that person's underlings and people around them. I haven't talked to that particular person, and their support is critical for the success or failure of the thing I'm doing. So the current you know DC default strategy would mean when, you know, a Senate officer, a Senate staff, the chief of staff, asked me.

Hey, well, do you have the support of person X? You know, 'cause obviously their their support is going to be critical to get the bill. The DC optimal strategy would be for me to say something like, yeah, no, we've we've we've met with them, we we've met with their team and they're very interested and they're gonna support it. so yeah, we're you we're good to go there. When th nothing of the sort has happened, right? I'm either I'm somewhere between if I were to respond that way, like exaggerating and lying. somewhere on that on that line. Doesn't matter. Bad either way.

My actual response, this happened like three times yesterday. so I have like a real life example or day before yesterday, I have a real life example of one at top of mine. My actual response was, nope, I do not have their support yet. I've tried really hard to get to them, you know. I've talked to this specific person in their office and this person. they were both noncommittal. Like they told me personally they thought it was a good idea, but that they can't speak for the office and you know, that they'd have to run it up the chain. And then like that, you know, one of them just like hasn't been responding to my emails. I don't really know why. so no, I don't have it yet. If you have any ideas about how to get their support, if you think it's a good idea, let me know because I'm kinda stuck.

I literally said that in multiple meetings. And immediately they were like, okay. Well yeah, that makes sense. well actually so in two of the meetings, now I have a path to get to that person directly. Because at two of the meetings, organically and independently, recommended, Well, actually, now that you say that, yeah, this other person that I know, I'll connect you with, and they have that guy's ear.

Right. they can help you, you know, get to where you need to go. They'll at least, you know, hear you out. They but they could actually either tell you, you know, hey, the the you know, this will fly or this won't, and you know, help you figure this out one way or the other. so I'll connect you to person. So in the act of being totally honest and just saying, No, I do not have that person support. I've been trying, I haven't gotten to them, I got the path to the person. and they one of the meetings the person was literally directly said

By the way, I really appreciate you like, you know, being honest and not, you know, papering over the fact that, you know, person X doesn't support yet. Yes, that's a big weakness of the push you're doing. but you know, it's obviously something you could overcome. so I love your your kind of in closing, your your reaction to that. You know, my experience, I'm sure you've had similar kind of experiences. Well i that's where I've landed.

Yeah. I I mean, you really shouldn't play a game that you're comparatively not good at, right? So you know, lying is most people that lie don't fail at it. Right. And the ones that do succeed with lying are the ones that are really good at it. Right because they're really good at it because they're kind of

You know, it's kind of based on some kind of truth or whatever, or they sort of are able to maintain kind of strong control over who has access, or that the information is siloed, like with the right people. And they're working just purely off of the public impression, or even if some people know that that's a lie, right? So I think that qualifies like for a large number of politicians, right? So you do you gotta know who you're playing against, right? So you can say you don't play the game, but then

A lot, you know, there's a lot that you so I think the right strategy is to kind of go with what you're strong at. And for me, that's the same as for you, right? So I also feel like there is a sort of breath of fresh air that people feel like when I'm pretty pretty honest and open. But it is something that can be exploited, right? So keep tit for tit for tat in mind, right? So one example I'm thinking of is

You know, remember Leopold Ashenbrenner and the Situational Awareness Fund, right? Sort of, you know, you know, you do your thing, but then Ken Griffin is kind of waiting for you, the shark that's there to meet you. Right. So there is just a certain right, this is how the game is played, and sort of more experienced people or people with a bigger what I would call is tool set or toolkit of that they have at their disposal.

Yep.

Exactly.

Would be at a significant advantage to you. So you can overcome that by sort of choosing your kind of your game or arena, and that's easier in a positive sum world. But in politics, you don't get to choose, right? Because there's potentially a very specific decision maker in front of you that you know is playing, you know, uncooperative. So you know, it is not a so so I do think it is good to have people.

I think it's good that you're in charge as someone who plays the more cooperative positive some game. But I think it's good to sort of strategically pick people around you that can kind of see through when somebody else is playing a different game. Because that widens your arsenal and your toolkit. Right. And that's kind of all right, then it's your responsibility. All right, what is the outcome? And does it create something that's good in the end, even though you're kind of using a different kind of tool than pure cooperation? Well, ultimately that's where I think.

That's a really good point.

You gotta be courageous, right? That's kind of where courage is, kind of to go also into areas where, all right, if you there is a risk that you do the wrong thing, right? So remember sort of the cowardice of people that just do research in their ivory tower, right? So that could also be a failure mode when when you only select games that you can win by cooperating with everyone. I think the cooperation is sort of the dominant strategy you should build around.

But I I do think you do want to think how you can overcome sort of how you can what you would do in a situation of adversarialness, right? And especially in the United States, like even if you're purely sort of in the positive sum like market game, like there is sort of the game of like lawsuits, right? And lawsuits are automatically leading to an adversarial situation, especially in the in the US system, right? So where the incentives are in many ways very screwed up.

Right. So all right, you you you you have to play that game. Once you're in like in a lawsuit or in litigation or whatever, you can't be totally open in others, right? Because that will be used against you. Right. I personally hate these games, right? I wish it were different, but I think to be sort of a successful in what you do and sort of be around that kind of leader, you do need to look into different tools into your toolbox or toolkit. All right. That might lead you to you know,

Correct, it's a different game.

to places where it's like you're kind of uncomfortable doing this. But I get I think you do need to prepare yourself to to go into these kind of places. And right that's kind of also your ultimate test, right? So as a leader, right, can you go using kind of tools that are something but pure cooperation and still sort of come out with your integrity intact. Right ultimately I think that those are kind of decisions

That you can't premeditate beforehand, you just have to go in it and then sort of decide based on the situation you're in. And ultimately, you know, the part of the outcome and part of like how you asser asserted yourself along the way and did the right thing will determine kind of your final verdict of what you were able to to achieve as a leader and what kind of your resulting moral merit is.

Agreed. There are th it's a really good point that there are times and places where like you just have to play a different game by the the circumstances of the the situation effectively. and it just kind of is what it is for lack of a better lack of a better term. But what makes that kind of palatable and I think still like net positive and net good is that the the thing you're pursuing is good for the world.

And you're doing this eyes wide open. Like you're aware, like, yeah, I'm just I'm I have to do this this way because of the direct circumstances, and I would rather it not be this way. I hate doing it this way, but like I just have to, because the the lawsuit example is just a perfect point, you know. I hate being, you know, secretive and not talking to people or whatever, but like quite concretely, doing anything other than like not talking to people and keeping things secret when you're engaged in litigation is the only thing you can do because doing anything else w just means you instantly lose the litigation and lose lose the long term goals or whatever else you're working on. So that's a good point.

I also hadn't thought about, but I think you're exactly right, the keeping around a you know, this is why I think from some his history I've read, you know, Middle Ages kings we keep a eunuch around, right? it's for exactly like this reason. It was the guy who was like, Hey, that guy that just came to, you know, petition you about his la land is actually like trying to take your throne. And here's why. You're like, I had no idea. I thought that guy was just here to visit and talk.

you know, th this is like a there's a historical pattern for this. I think you're exactly right. Like you you need to and this is part of just almost like leadership and entrepreneurship 101, really, but it's one that people often forget, which is like surround yourself with people who are way, way better than you, smarter than you, you know, more adroit than you at the things that you suck at. You have to have no ego and just say, like, hey, I suck at thing X, and so I need to find a person that's really good at thing X to kind of buttress and patch that weakness.

I hadn't thought about kind of some of our the you know, the more kind of policy politics oriented work we do that way, but it's great you're right, a lot of that it kind of entrepreneurship lessons carry over just one to one directly. And that's a really good one. I hadn't I hadn't previously considered. Have the guy that can see the people playing the Machiavellian games. Not because you have to play them per se, but so that you are aware it's going on at all. and and that you can respond to it and adjust your strategy accordingly. I I totally agree. So

Awesome. Well I think we'll call it there. really appreciate your your time, Niklas. I definitely again, even if nobody ever listens to these, I want to make this like a semi regular thing, you know, maybe like once a month or every, you know, couple of weeks, whatever we end up doing, you know, we'll we'll fill it out. where you and I will be a a recurring duo, riffs and rambles. Maybe we'll see if we can get some other people to join occasionally and we'll just dive deep into whatever has been fascinating and and interesting to us and see if we can we could learn something along the way.

Always will be long totally a wrong photo right. And try to release this episode as quickly as possible. Like that's the most important lesson of podcasting and getting started. Like release it.

Okay. Good to know. Okay. I'm gonna do that. Alright, I tell you what, if I have not released this by seven PM Central, message me. I say that just 'cause I just had hilariously, that person I was talking about that I could not get to to get their support. I just got well that person just connected me with them. Like five so I need to go talk to that guy and send him something and make something for him real quick. But in any event, yeah, if I haven't done this by by seven, message me. and then we'll we'll we'll keep the ball rolling there. Awesome.

Yeah, nice.

Awesome. Appreciate you, brother. Take care.

Appreciate it, Niklas. You too.