Geoffrey Miller: Evolutionary Psychology, Polyamorous Relationships, and Effective Altruism — #26

Show Notes

Geoffrey Miller is an American evolutionary psychologist, author, and a professor of psychology at the University of New Mexico. He is known for his research on sexual selection in human evolution.

For reference: https://en.wikipedia.org/wiki/Geoffrey_Miller_%28psychologist%29

Steve and Geoffrey discuss:

0:00 Geoffrey Miller's background, childhood, and how he became interested in psychology
14:44 How evolutionary psychology is perceived and where the field is going
38:23 The value of higher education: sobering facts about retention
49:00 Dating, pickup artists, and relationships
1:11:27 Polyamory
1:24:56 FTX, poly, and effective altruism
1:34:31 AI alignment


Music used with permission from Blade Runner Blues Livestream improvisation by State Azure.

–
Steve Hsu is Professor of Theoretical Physics and of Computational Mathematics, Science, and Engineering at Michigan State University. Previously, he was Senior Vice President for Research and Innovation at MSU and Director of the Institute of Theoretical Science at the University of Oregon. Hsu is a startup founder (SafeWeb, Genomic Prediction, Othram) and advisor to venture capital and other investment firms. He was educated at Caltech and Berkeley, was a Harvard Junior Fellow, and has held faculty positions at Yale, the University of Oregon, and MSU.

Please send any questions or suggestions to [e-mail] or Steve on Twitter @hsu_steve.

Creators & Guests

Host

Stephen Hsu

Steve Hsu is Professor of Theoretical Physics and of Computational Mathematics, Science, and Engineering at Michigan State University.

What is Manifold?

Steve Hsu is Professor of Theoretical Physics and Computational Mathematics, Science, and Engineering at Michigan State University. Join him for wide-ranging conversations with leading writers, scientists, technologists, academics, entrepreneurs, investors, and more.

Steve Hsu: Welcome to Manifold. My guest today is Geoffrey Miller, professor of evolutionary psychology at the University of New Mexico. I am very, very happy to have Geoffrey on the show. We've been trying to schedule this for a while. Welcome, Geoffrey.

Geoffrey Miller: Great to be here, Steve. Delighted.

Steve Hsu: Great. So, Geoffrey, I always like to get a little bit about the childhood and growing up experiences of my guests and I. I know you, you grew up in Ohio and then you went off to Columbia for your undergraduate degree.

Could you just flesh that out a little bit? Like maybe just say to us what the world looked like to you? back in, I guess that would've been the early eighties.

Geoffrey Miller: Yeah, I was born in ‘65. I had a very happy, bourgeois upbringing in suburban Cincinnati, Ohio, which is a relatively conservative place I went to. Wonderful, actually wonderful public schools that had really strong, standardized testing and tracking, which was a big benefit to me. I had a great time in school, and wonderful local friends.

My dad was a lawyer. Mom was a politically engaged homemaker. And I had a sort of unusual family structure, which is that my mom was one of 12 kids. and a lot of my uncles and aunts were very similar in age to me, lived just a few blocks away. So, we had sort of the benefits of both of the bourgeois nuclear family, but also a very large extended family nearby.

Steve Hsu: Wow, that's great. Now, when you were growing up, did you, did you have a sense that you were gifted or intellectually precocious?

Geoffrey Miller: I knew I was very good in school, but I think my parents instilled some really strong civic virtues that basically said, you know, the smarter and more talented you are, the more of a, a duty you have to sort of, give back. And not to be arrogant. Not to be narcissistic. And you need to be using your talents kind of pro socially.

And, that for me was kind of a good lesson because I'm, I'm naturally, you know, a little bit narcissistic and, having that humility was helpful. Also, honestly, being in a heavily tracked, public high school where, you know, there's honors programs, there's AP classes, there's a lot of academic competitiveness.

I was smart, but I was not the smartest kid in my class. My best friend was actually the valedictorian and probably at least five or 10 IQ points smarter than me. Likewise at Columbia University. I think it's really important for smart kids to go to highly competitive schools so that they get the humility to know, yeah, I'm not the smartest kid in the room.

And to be able to learn from, from others and to kind of get calibrated, so that you don't end up with the kind of intellectual arrogance that can be quite toxic.

Steve Hsu: I think I, I, heard you say in an interview that you were very happy to have gone to Columbia when you did because they were still doing the Great books curriculum.

Geoffrey Miller: Yeah, absolutely. It was really inspiring. You know, you first, you spend the first couple of years at Columbia University, and I think this is still largely true reading, the great books of western civilization and philosophy, religion, science, social theory, et cetera. There's a heavy emphasis on the Greeks, the Romans, the Middle Ages, the enlightenment, et cetera.

And you talk about these in small seminars that are run by really world class academics who are actually teaching and not just getting research grants, they're actually running seminars, and these are like famous people. That was enormously inspiring. And then, Columbia University had such an amazing smorgasbord of other great classes.

I took a lot of courses in the East Asian Studies Department, Chinese history, Japanese art, et cetera, just because I was interested in it. And you, you know, I ended up majoring in psychology. but I had very broad interests. I read a lot of actual deconstructionism, dairy DA and Fuco and leotard and all, all those guys who were very popular in the Columbia philosophy department at that point.

so, I feel like I got both a really good grounding and empirical psychology, but also quite a bit of exposure to the kinds of thinking that would go on in, in the modern era, in the 21st century to become kind of the foundations of, of wokeness. So, I feel like I kind of got both sides of that, in college.

Steve Hsu: We can get back to wokeness in just a little bit. Can you put your finger on what attracted you to psychology at that stage of your life?

Geoffrey Miller: Weirdly, I kind of got into psychology via my interest in contemporary art, which is a very roundabout passage. I was always interested in art. I ran an art magazine in high school. I took a lot of courses on the history of art. I got very interested in 1960s and seventies conceptual art, which is all about kind of, it's not really about visual aesthetics.

It's about creating certain kinds of concepts or ideas or connections between ideas in the heads of the viewer. Often conceptual art was purely verbal. You just put up a sentence on the wall of an art gallery and hope that the viewers will think about it. And that actually led me into cognitive psychology because I thought, wow, conceptual art.

How does that work? How do you have an aesthetic response to just words on a gallery wall? So I got very interested in memory learning, categorization, conceptual thinking and metaphor. I actually did my undergraduate psych honors thesis on cognitive processes, underlying metaphor and analogy, which I thought were fascinating.

Geoffrey Miller: And then eventually, you know, later in grad school that led into evolutionary psychology.

Steve Hsu: Was the language instinct already, a book at that time, or was that, was that later?

Geoffrey Miller: That was later at the language instinct was by Steve Pinker. I think it was the mid or late nineties. I graduated college in 87, went to grad school at Stanford, 87 to 92 basically. So that was actually before a lot of the popular f psych books of the nineties.

Steve Hsu: I see. So you were, you were kind of ahead of your time in thinking these thoughts. I think.

Geoffrey Miller: Well, I just. It's a pure stochastic accident. I happened to be a graduate student in cognitive psychology at Stanford. At a time when my advisor Roger Shepherd, who's famous for mental rotation experiments, shepherd, happened to hire two postdocs, lead to Cosmos and John Ruby from Harvard who came to Stanford and spent a couple years there.

And they were developing evolutionary psychology there in the late eighties. And I just happened to hang out with them and get very fascinated by it. And me and my friend Peter Todd, who's now a professor at Indiana University, got kind of maybe a little too obsessed with applying evolutionary biology to psychology and heavily inspired by

Lee Cos and John Tubby, but also by Steve Pinker, who was just getting into it at that point. David Bus from Texas visited Stanford a bit. and Gared Gigerenzer, was also a big influence on how we thought about judgment and decision making from an evolutionary perspective. And he was visiting Stanford also for a year.

Steve Hsu: Wow. So, you were very fortunate actually, that's quite a cast of

Geoffrey Miller: Yeah, it was, it was, well, very fortunate. Intellectually it probably set me, just in terms of pure career and like getting jobs, it probably was extremely handicapping, but that's fine. You know, I'd rather have a really engaging intellectual career, without necessarily being an Ivy League professor than, than vice versa.

Steve Hsu: Yeah. One, one of the things that I, I, I admire about you is that I can just, I can tell because I've followed your thinking for probably 20 years now, you're, you're not afraid of, of, you know, pursuing your own path, based on quality and what, what's really interesting as opposed to just following academic fashion.

Geoffrey Miller: Yeah. And you know, the truth of the matter is that now, with, you know, with the internet and Zoom and so forth, you can really have colleagues and friends who are extremely geographically remote. You don't really have to be in the same place as other intellectuals the way that you would've had to do throughout, you know, any time before, the internet.

So, it would've been quite handicapping, you know, to be a psychology professor anywhere other than New York, Boston, or maybe the Bay Area, until quite recently. But now it doesn't really matter where you are physically.

Steve Hsu: Yeah, I have the same feeling. I mean, I find myself able to run my research lab and, and even a startup almost remotely, with people scattered all over the world. So, it's definitely a new era. I wanted to ask you when you were at Stanford, so can you reconstruct what the world looked like when people first started thinking about evolutionary psychology?

Because, you know, in some sense, you know, if you go all the way back to say, Darwin, talking about mate selection and stuff like this, the idea that evolution had to, you know, have an impact and vice versa on behavior. It's kind of an old idea, but I'm, I'm guessing it was, it was kind of shocking when Tubi and Sedi started really getting along.

Geoffrey Miller: It was kind of shocking, but I dug quite deeply into the history and pre-history of evolutionary psychology. When I was in grad school. I read a lot of Darwin. I read a lot of them. psychology in the late 18 hundreds that had been inspired by Darwin. Long story short, you know, basically what happened was, psychology ever since Herbert Spencer, circa 1850, all the way through about 1920, was fairly Darwinian.

And people like William McDougal and William James at Harvard were quite inspired by Darwin. They thought a lot about instincts and the origins and functions of human, mental faculties as they called them, and how cognition and motivation and, and volition and our goals are influenced by evolution.

Then you get the behaviorist revolution in the 1920s that basically says, stop thinking about e. Think instead about learning, think about classical and opera conditioning and how that shapes organisms. don't think about internal processes in the mind. Don't think about emotion or motivation. Just think about learning.

And that dominated psychology from about 1920 till about, hmm, 1970. Then you got the cognitive revolution, right? Basically what happens is a bunch of grad students in the sixties are trained in behaviorism, but then A, they learn about computers, and B, they do psychedelics, and they get very interested in how the mind actually works internally.

And that's the origins of cognitive psychology. Also in the seventies, you get the sociology wars at Harvard, you get Iel Wilson. In 1975 publishing sociobiology and saying, Hey, let's, let's look at how evolution shapes animal behavior. And then you get the blowback, right? You get the leftists at Harvard like Stephen Jay Gould and Richard Lein and Leon Camon saying, no, no, that's terrible. That's fascism. You're, you're not allowed to study evolution, in relation to human behavior. Terrible, terrible. So, they managed to squash sociobiology temporarily, but then you get, you know, Richard Dawkins publishing the office Gene 1976, and you get Cosmos and two B at Harvard, who witnessed the tail end of the sociology war is saying, okay, sociobiology as a brand is dead.

It didn't pay enough attention to internal processes in the mind. How about if we meld evolutionary biology and cognitive psychology and study, the mind as a sort of information processing system shaped by evolution. and from the mid-eighties through the early nineties, they, COSD, Tobe worked on that.

I think they were very successful. They basically launched evolutionized psychology as a kind of socio biology 2.0. But it proved enormously attractive to graduate students like me who wanted a paradigm, an integrative concise paradigm that linked the biological sciences to the behavioral sciences.

Steve Hsu: And you, you, you, you sort of already hinted that, because this was kind of a new turn in the field, specializing in it, didn't help your career prospects. I.e., It was probably harder for you to get a faculty position than if you had done something that was more in vogue at the time.

Geoffrey Miller: Yeah, totally. I mean, I was basically in postdoc limbo for about nine years, including a lot of time in Britain and a year in Germany. Incredibly hard at that point to get any kind of academic job if you were, identifying as a Darwinian. because, you know, people still took the old critiques of Soo biology seriously, and they were like, oh, well, Stephen J.

Gould has taught us that anybody who cites Darwin is on the far right and is dangerous and probably a, a crypto fascist and thinks way too much about sex differences and intelligence differences, and God forbid even pays attention to behavior genetics. So, there are plenty of jobs I applied for where, you know, I know I didn't get them because there was one or more people on the search committee who had read Gould or read other critiques of evolutionary thinking and thought, no, we don't want to hire this Miller guy.

He's, he's too dangerous. But, you know, that's the price of admission. If you want to get involved in a new intellectual movement that's a bit controversial, you cannot plan on having a sort of easy, mainstream career that pleases everybody.

Steve Hsu: Yeah, I was kind of in the opposite situation because I had originally gone to graduate school to study string theory. But once I learned enough string theory to realize that it was very unlikely to be experimentally testable in my lifetime, I sort of switched away from string theory. But in an era when almost all the top jobs are going to string theorists, so I, I, I have a lot of sympathy for what you went through.

Now in terms of how the field of psychology regards evolutionary psychology since the nineties, did the situation get better and now it's worse because of ageism? Or how, how would you characterize that?

Geoffrey Miller: I would, I would say s you know, sadly, mainstream anti evolutionary psychology has largely succeeded in, in squashing and suppressing evolutionary psychology as a field. Psych gets quite a bit of public attention, and there have been some amazing popular science writers who have popularized many psych ideas.

However, if you look at the actual numbers of psychology faculty, there's probably less than 500 evolutionary psychologists worldwide actively doing research, versus there's probably 20,000. Social psychologists and something like 40 to 50,000, neuroscientists in psych departments. So, psych is a tiny, tiny, tiny field with kind of outsized impact, especially in terms of public thinking.

But as with behavior genetics, which is another fairly small psych field, its influence has been successfully quiet, limited by, by the left, basically in, in the behavioral sciences.

Steve Hsu: And would you say that situation's getting better or worse?

Geoffrey Miller: I would say it's at, it's been at kind of an equilibrium for a few years, although it's gotten worse, in the last five years or so in terms of, the incredible rise of wokeness and its institutional power in academia.

Steve Hsu: Yeah. The way I might characterize it is that since George Floyd, you know, even the basic ideas in behavior genetics or Evo psych are basically just off limits. Like you, you could almost be canceled just by stating the standard things that you've given an introductory course in those.

Geoffrey Miller: Yeah, that's right. And you know, there's a bunch of, ornery, middle-aged, and older folks like me who have tenure and don't care and have already been canceled and say what we think on social media. However, there are not a whole lot of super talented grad students coming into the field because they can, they can read the writing on the wall.

They know if they study psych it'll be extremely hard to get a tenure track job or to succeed in American academia. And that's sad, but that's kind of where we're at. You know, hopefully the pendulum swings back in a few years and people, maybe the left having a strangle hold on. What ideas you're allowed to research is kind of a bad idea.

Steve Hsu: Yeah, I think it's very bad, but you know, these things can last for long periods of time, so.

Geoffrey Miller: Indeed. Yeah.

Steve Hsu: Yeah, I don't know what to think. You know, in the case of behavior genetics, you know, because of the genomics, explosion of data and genomics, it should be a kind of golden era for behavior genetics.

So, if you, if you look at the polygenic scores and things like this that people work on in behavior genetics, the corresponding predictors that are used for diseases and, health risks and things like that, those are very well supported and, you know, get thousands and thousands of citations and things like this in the sort of biomedical literature, but are still kind of very controversial in behavior genetics.

Geoffrey Miller: Yeah. And the extraordinary thing is, you know, if you'd asked me back in the late eighties in grad school, will people still be doing kind of social and developmental studies that are not genetically informed, where they don't bother gathering DNA and they don't bother, sampling from, you know, families with twins or adopted studies?

I think no way, it couldn't possibly be the case that in 30 years people are still doing genetically uninformed studies and saying, oh look, this thing happens to this kid at time A and then there's this adult outcome at time B. Therefore A causes B. Right? And any behavior geneticists will know. No, there might have been underlying genetic predispositions that both, you know, caused A and B.

It's not necessarily environmental, but, but it's still the case. And most, behavioral sciences. that people think, oh, if we do a longitudinal study across time, that we can infer causality and we can ignore genes, which I think is scientifically unconscionable. And yet, it's really quite difficult, still to get grant money to do serious genetically informative studies.

And there's also a big taboo against getting involved in genomic studies or studying polygenic scores for psychological traits like intelligence, personality, or mental health. because, you know, people just get very uneasy thinking about genes.

Steve Hsu: Yeah, I, I agree with you. I can open up any journal, even nature, and easily find, you know, in any given month a very prominent study where one could read it and then say yes, but they didn't consider genomic genetics. And genetics could be a huge confound to their conclusions. Like the causation could be backwards or et cetera, et cetera.

so that's, that's, you know, I guess the, a negative aspect of the current situation on the positive side in terms of, you know, small groups of people doing good science, it does look like in many or most of the old longitudinal studies, you know, where they followed some group of kids, you know, for 50 years or something.

Steve Hsu: they're gradually genotyping everybody in the cohort. And so now the, if you want to validate some genomic predictor that you've built, you can find longitudinal cohorts where, you know, you can check to see whether the predictor actually predicted properly, what their, you know, socioeconomic status or longevity or, you know, whatever outcome, you choose, you know, happened over the 50-year period.

And so, at the scientific level, things are actually really amazing. but you know, I think within academia, 99% of academics are not aware of this progress.

Geoffrey Miller: Yeah. And of course, Americans have this few that, only the science that happens in the US really counts, that everything else is kind of a backwater. but a lot of them, the large scale global. Genomics consortia are extremely multinational and they involve American, European, Australian, you know, other kinds of, of, of scientists.

And I've written before about the rise of behavioral sciences in East Asia and South Asia. And, you know, it's kind of obvious that as long as scientific research continues in places like China and as long as they keep investing in behavioral sciences and they have some degree of freedom to explore what they want, particularly without the constraints of American wokeness, it might not matter that much in the long run what the limitations on American scientists are. It might matter quite a bit more. what kind of scientific culture is developed in China, India, Southeast Asia, et cetera.

Steve Hsu: Yeah, I, I completely agree with you. I, as you know, I've, I've worked with collaborators in China and currently I'm working with, something called the Taiwan Precision Medicine Initiative, which is a, a very big effort to, I think they've collected now a million people in Taiwan through the healthcare system and genotyped about half of them.

So, you know, the gap between the data, the size of data sets available in the West, and at least in East Asia's closing very fast. You know, I want to give you credit, publicly give you credit, Geoffrey, because you were way ahead of everybody else in understanding this. I think you and I met in 2010, in person at the Behavior Genetics Association meeting.

It's an, I guess an annual meeting, and it was in Edinburg. I think that's right. And you, you really, I think you had already spent a year, as a visitor at the University of Queensland, which is the home of one of the top genomics groups in the world. So, you, you really knew what was coming. That was, that was, I guess now 12 years ago, but you already knew what was coming and I don't think probably most people are aware that you were that far ahead of the curve.

Geoffrey Miller: I'd gotten very interested in behavior genetics in relation to intelligence, even in, in the late nineties. Because I happened to be working at University College London, I actually organized a meeting in 1999 on the nature of intelligence that brought together some psychometricians and behavior geneticists and evolutionary psych people.

And I really took seriously the idea that evolutionary psychology should. Build tighter bridges with behavior genetics and with intelligence research, and wrote a bit about that, you know, around 2000. Then I had this opportunity, in a sabbatical in 2008 to go work with Nick Martin's genetic epidemiology group in, Brisbane, Australia, associated with University of Queensland, and learned a lot and tried, tried to kind of, get up to speed with the, the quantitative genetics, the evolutionary genetics, the, the emerging, you know, polygenic score revolution that was just starting at that point and I thought, wow, this is, this is amazing.

We will soon have a lot of DNA samples from large numbers of people for various reasons. you know, ancestry testing. Medical, genetics, et cetera. We will have these polygenic scores to predict complex traits including psychological traits. And I thought, psychology is going to be kind of blindsided by this and that's sad.

And you know, by the time we met at that, be it genetics, meeting circa 2010, I was pretty stoked about all this. And, I haven't, you know, actively pursued any kind of research in this area. Partly because like there's younger and smarter people than me already doing this and they're really good and they can actually program an R and they can do very complex multi-variate analyses and they know; you know, the evolutionary molecular genetics way better than I do.

I can just kind of bean, an appreciative, observer and promoter and. Try to take seriously what they find and, and what its implications are for psychology.

Steve Hsu: Yeah, I mean, it is really a well-developed, I mean, if I, if I take the field as a whole, so the field that uses large genomic data sets and tries to predict phenotypes using what I would call machine learning, what some people might call statistical methods, that's now a huge field. And yeah, there are a lot of really capable young people pursuing it.

It's making tremendous progress, mostly in biomedical. unfortunately, the, the size of the data sets in which you actually have the right label, the cognitive scores or personality scores for the individuals in the dataset, that has really not increased that much in the last, I would say, yeah, we've had not, not so much, advancement, in the last five years.

Steve Hsu: I guess I should be more careful because the educational attainment people thought have, have pushed things ahead quite a bit, but they're, they're kind of hitting a kind of plateau now, because that's a very complicated phenotype, educational attainment. It sort of mixes up lots of separate things.

But yeah, I mean, the behavioral and cognitive side of it is the lagging side of it, unfortunately, and I think primarily for ideological reasons. But, on the biomedical side it's advanced tremendously.

Geoffrey Miller: Yeah, and I think there's also a measurement issue with a lot of these psychological traits. It's very funny. We have extremely powerful, you know, ways to collect DNA and analyze, single nucleotide polymorphisms, and get polygenic scores and, and that's all great. But on the psychological side, right, we're still often using intelligence tests that were developed.

50 to 70 years ago and haven't been really that well updated. We're using self-report personality trait scales also developed decades ago that are kind of okay, but not super reliable or valid. we're using mental health also mostly self-report scales honestly developed decades ago. And so, this is very silly.

Like if we spent several million dollars, I'm convinced we could have interactive adaptive online intelligence tests that are extremely reliable, hard to fake, valid, and that give us much more accurate, robust measures of human intelligence even than the Stanford Benet or Waste IQ tests. But we just haven't done that because it's taboo.

And likewise with measuring personality traits in a way that's valid and reliable. Google and Facebook probably have ways of measuring human, conscientiousness and planning and reliability. that would far outstrip, you know, the big five personality trait scales because they can actually tell, like how conscientiously people drive and plan and manage their calendars.

And they probably have access to our, our literal credit scores and how we pay our bills and all of that. But it's all siloed within these social media companies, aka data refineries, right, that sell our psychological data to advertisers. And they probably have better metrics of a lot of these psychological traits than anybody in a psych department does.

Steve Hsu: Yeah, I, I think you're a hundred percent right. Now the question is, you know, if you were, if you, you know, were the boss of Google and you could reach into the existing organization, I wonder whether they don't have, you know, some team of Psychometricians or. At least people who understand psychometrics, who you know, are actually, you know, maybe pursuing this as a 20%-time research project or something. If not, then it's, it's, it's just a colossal waste of a, a data resource.

Geoffrey Miller: Yeah, it's funny, I've, I've done a fair amount of consulting work with market researchers and advertisers and big, big companies over the years, and, oh my gosh, they have such colossal amounts of data. But on the other hand, they're incredibly naive about psychology, and a lot of them still believe in things like the Meyers-Briggs personality.

which were invented in the 1940s based on, on Yi and psychoanalysis and really don't have much validity. So even if Google had a team of pretty good psychometricians, I suspect it'd be quite hard for that team to get anybody else in Google to pay attention to them. Because probably most of the other Google employees are skeptical about psychology, not trained in psych.

Unfortunately, they've probably read Stephen Jay Gould’s, The Mismeasure of Man, which is just one of the worst pop science books ever written in terms of its lack of intellectual integrity. So sadly, the corporations that have the data that could be really, really useful are probably staffed by people who aren't.

Academically or ideologically equipped to do anything useful with that data.

Steve Hsu: Yeah. You know, the one I occasionally hear stories about people who, for example, they're data scientists and they're at a company maybe that, is, is giving out loans or something. And you know, as part of that, they get access to some data. For example, maybe they can get access to your phone, some, some aspects of your phone.

And one of the things that I guess it was claimed had a very strong correlation, with, you know, your ability to, or your, your likelihood of repaying a loan had to do with the average charge level on your phone. So, if you're very anal and you keep your phone charged near a hundred percent, you're actually much more likely to be, a good, a good borrower.

Geoffrey Miller: Yeah, that, that totally makes sense.

Steve Hsu: yeah, so, so at that point I thought, oh, maybe there is some very interesting data science going on looking for signals like this because the signals are worth something. at the, at the level of doing genomics, you're a hundred percent right, that they're just, there's no innovation in, you know, trying to do adaptive testing at scale.

I mean, you could push out an app to, you know, everyone who's 23 and me and try to measure all kinds of aspects of their personality and, and cognitive ability. But there's, there's really nothing going on. And it's, it's kind of verboten and you can get in big trouble for it. You know, people will ask you immediately, why are you interested in those things? You know, it's kind of a bad look for you to be interested in those things. So, I'm not optimistic about that.

Just to give a very specific example, in the UK Biobank, there's a fluid intelligence score and, I think it's based on just 12 questions, and it has been administered to a fair chunk of their half million people. So, there's maybe, I don't know, last time I checked there's maybe 200,000 people who have done this fluid test, and I believe it was, you know, validated and built-in consultation with a guy. Maybe you know, this guy, Ian Deri.

Geoffrey Miller: Oh yeah.

Steve Hsu: Yeah. So, I think it's pretty valid, but of course you're limited by what you can do in 12 minutes or in what, 12 questions. Right. So, you know, it's very far from the limit of what one could achieve with modern technology.

Geoffrey Miller: Yeah. And it's a real missed opportunity because look, if you've got, you know, half a million people who've got the DNA and the UK Biobank, hypothetically, if, let's say, you know, a fifth of them, a sample of a hundred thousand, were willing to do more like a one hour adaptive intelligence test and you could pay them like a hundred quid for doing that in an hour.

For the cost of only 10 million pounds, you'd suddenly be able to do incredibly powerful genomics on intelligence. Now, ideally, you'd want a bigger data set because I know you, you can't really get a good polygenic score with only a hundred thousand people. But you know, you could scale that up.

Geoffrey Miller: And the idea that, you know, we can have crypto exchanges that lose 10 billion sorts of overnight for, for that price, you could do incredibly powerful genomics on major psychological trades that could sort of revolutionize our understanding of education and employment and society.

Steve Hsu: Yeah, you know, you, you and I are very like-minded on this, and you know, I, I, I can disclose it. Over the past few years, I've been talking to a lot of billionaires, including crypto billionaires, trying to get them to fund. a project to collect the data set that you just described. the target size though is about a million people actually what you

would need. And, based on some fancy math, that's roughly the estimate for what you need to build a cognitive predictor, which is as good as the height predictor that we built. so, but you know, the amount of interest is quite limited even among people who get the benefit. There's just a lot of reputational risk for being associated with that kind of project.

And, so I, I, I don't know when it's going to happen. Very as, as you said, and as actually as you said, very presciently, like 12 plus years ago, it's very likely that the first time it'll be done will be in China or some East Asian country where they just don't have any hang-ups.

Geoffrey Miller: Yeah. And I think, you know, I, I taught at a Chinese university remotely online for about, a, a year, and a half, and I was. So impressed with the intelligence and dedication and hard work of the undergraduate students, but oh my gosh, the, the, the pain and anxiety that they'd suffered to prepare for the GCO exam, the university entrance exam in China w was enormous.

It's just years and years of their adolescent lives spent studying really, really intensively, to pass what amounts to basically an IQ test. Right? And if you could do a genomic score that was pretty good at, at IQ and you could just give a DNA sample in, in five minutes instead of studying for the gal cow for six years, and if it predicted your university, aptitude about as well, that would be a huge win.

And my sense is that the Chinese government is finally kind of understanding the social and emotional costs. Of really intensive, high stake, standardized testing. And maybe if they could have a genomic test that, you know, could fill the same role in terms of, kind of helping their cognitive meritocracy and figuring out, you know, who, who would be really good at University of Beijing versus who should go to more, you know, mid ranked provincial university.

that would save them an awful lot of time and effort and money and, and mental health costs.

Steve Hsu: The problem there is that, you know, okay, on the plus side of the ledger, for them, they haven't read Stephen J. Gould

Geoffrey Miller: Yeah.

Steve Hsu: but on the minus side, they also haven't read much about psychometrics in general. And you know, they're not opposed to it. But when you talk to me, I talk to a lot of East Asian scientists and engineers and technologists. They're not opposed at all to the concepts. And when you show, when you show them the data or the research results, they, they can, they get it right away, but it's not really part of their education. So, they don't really have a good feel for, you know, how innate certain things are. Or, you know, there is a very strong feeling in Chinese culture or East Asian culture, Confucian culture, that if you work really hard, you can improve yourself. or even that, even if you have very realistic feelings about okay, how there's a limit to how much of that you can do. Nevertheless, you almost have a moral obligation to work really hard and try to improve yourself, especially as a student. So, there's a whole different confluence of factors there that it's not going to be an easy slam dunk.

I've been trying there as well. It's not going to be an easy slam dunk there, but at least they don't have some kind of quasi-religious opposition to the idea that you can measure intelligence. And it might be partially heritable.

Geoffrey Miller: Yeah, I think the, the crucial thing is both in the US and you know, the UK and Europe and China, you have very strong vested interests like the educational establishment, that wants to convince everybody that, going to school and doing homework is the royal road to, intelligence and knowledge and learning.

And that, if you emphasize too much the kind of innateness of some of these traits that it would, it will challenge their funding and their power and their influence in society. And I think this is also, you know, a lot of the opposition in America by the teachers’ unions against standardized testing is, they rightly perceive it as, a threat to their influence, but also a threat to, how much people are willing to pay for educational credentials, right?

Because look, if you had companies like Google just using intelligence tests to select employees instead of, how prestigious is your undergraduate degree, then the pressure to get a prestigious undergraduate degree would drop, and that would hurt enrollment and that would hurt people's ability, you know, people's interest in taking out colossal student loans.

So, higher education in general is extremely strongly opposed to IQ testing, not just for ideological reasons, but there, their economic interests are very threatened by this.

Steve Hsu: Yeah, no, that's exactly right. I don't know if you're,

Have you ever heard of something called the C L A College Learning Assessment?

Geoffrey Miller: No, I haven't

Steve Hsu: So, I learned about this when I was an administrator. So, the college learning assessment was an effort by, it was a collaboration between universities and companies to build an instrument that you could administer to graduating seniors, sort of like the G R e, but, but meant for companies to be able to evaluate whether students had acquired the right skills for knowledge workers at leading corporations.

So, they said about building this test, and, and they did it with input from industry. So, it wasn't a bunch of psychometricians or party headed college professors, it was people from industry saying, well, we really want our, you know, our staff to be able to like, read an email, read a long article or something, and write an email memo or look at these graphs and summarize them in a meeting.

You know, the tasks are very, Practical. And, you know, unobjectionable, I think any, you know, ex C level executive at a big public company would say, yeah, yeah, that's, that's what we want. We want to know whether the student graduating from Tulane can really do that. Right? Everybody was very happy about this project.

And of course, you're, you know, with your background psychometric, you can just guess that, okay, this is just a, a very kind of clumsy way to measure someone's, you know, g or something, but they didn't know this at all. The people who built the C L A and they turned it over to the Rand Corporation for a big study.

Steve Hsu: And that big study included 20 universities ranging from, I think Michigan and m I t, to some directional state colleges and also some historically black colleges. So, it was really a super extensive study of the CLA. And what they found was, there are two alarming aspects of the C one is, the s a t score that you had entering colleges, college predicts your c a result as a senior pretty well.

And secondly, there isn't much change in your C score. If I give it to the freshman, it doesn't change very much. They're not, they don't seem to be learning a lot of c skills, in four years of college. So, this is just incredibly damaging. These results are just incredibly damaging to the whole, all the, all the kind of, friend, you know, happy lies that we tell ourselves about how higher ed works.

And so basically the C is just, I think, more or less buried after, after that it was not really adopted by a lot of schools. and it's a good story because I, I mean, I became aware of it and I, I, I actually had a discussion with the other senior leadership at Michigan State, so people at the VP level, deans, provost p. I, I, I just explained to them what I just explained to you, and of course they all were like, they were like the robots in Westworld where, when you show them their own wiring diagram, they pretend they can't see anything. They say like, there's nothing on that page. You're holding up Steve. I can't, I can't understand what you're talking about.

So, I basically encountered that reaction, from people who are running, you know, one of the biggest universities in the United States. So, yeah, the higher ed, higher ed is well aware of, of, where its survival, and best interests lie.

Geoffrey Miller: Yeah. And even, even within any given department, you know, like, my psychology department tries to measure, okay, how much psychology do incoming freshmen know when they start taking our intro psych classes versus how much do they actually know when they graduate from the intro psych class versus how much do they remember two years later?

They do actually learn a bit, you know, from intro psych, but man, they forget almost all of it within a couple years. And so, the retention of this material is extremely weak and, you know, economist, Brian Kaplan covers a lot of this research in his book, the Case Against Education, to say actually the value added by a lot of high school and college classes.

we are greatly over. How much is actually retained by students over the long term. And so, if you're saying higher education is crucial to developing educated citizens who really understand the world, well look, if they hardly remember anything by their late twenties that they learned as undergraduates, what are you actually doing in terms of contributing to their, you know, functioning as citizens or workers?

Geoffrey Miller: It's, it's pretty dubious.

Steve Hsu: Yeah, I agree with you. I mean, having slaved, I mean, actually I shouldn't say slave because I actually enjoy teaching, these courses. But I, I, in Oregon, I taught these, sort of physics for poets courses.

which, you know, we're not allowed to use any math beyond simple algebra, but I did many, many really beautiful physics demonstrations where you demonstrate basic principles like concentration of energy or something. and, or, or what temperature is really, you know, what, what temperature really represents and things like this. So, I, I thought, okay, I'm, I'm equipping these kids with these really core ideas, which, you know, they, if they, if they can just retain them, they'll be valuable to them the rest of their lives. But then I kept running into graduates that had had my class.

Steve Hsu: Like I would be at some, like at the cell phone store or something and some, but he would come up and say, oh, Professor Shu, I had your class, you know, conceptual physics three years ago. And then I would just, for fun, I would just ask them, like, what they remembered from the class. And it was like, almost nothing.

They would just say like, I, I, you know, they can't remember anything, but they said, wow, you were a really good lecturer though. So, yeah, it's, it's, it's, it's not clear what we're actually accomplishing.

Geoffrey Miller: Yeah. And you know, my, my hope is, I, I teach courses like, human sexuality and human emotions and evolutionary psychology. I try to emphasize applications to people's personal lives rather than their employment or their kind of overall worldview. I'm like, look, here, here are some typical failure modes of relationships, or here's how to talk to your boyfriend or girlfriend about this tricky issue.

Or Here's, you know, some, like actionable information about like, sexual anatomy or, or contraception or, you know, how to choose the sperm or egg donor or whatever. My hope is that some of that personally relevant stuff might stick better than if I was teaching, let's say, a neuroscience class.

But I don't have any particular evidence, that it actually influences

Steve Hsu: I think you're totally right. I think those students probably rate your course on relationships or, or whatever, as you know, among the best that they took and that was the most useful because people are genuinely interested in those subjects. Right. So, another one that my son took in high school was a course in high school called Personal finance, which sounds very Class A and non-intellectual and stuff.

But if they basically just teach a kid what checking accounts are and how credit cards work and interest and the stock market and stuff, it's extremely useful to people for the rest of their lives. but of course, it's not viewed as a quote core academic class at the high school.

Geoffrey Miller: Yeah, I think there's a lot of, there’s a huge mismatch right, between stuff that adults think, oh man, I wish I'd learned this in high school, in college. personal finance, investing, pensions, how to do a job interview well, how to run a household, how to take care of babies, how to drive safely.

how to, you know, make, and repair your own furniture. How to shop effectively, how to avoid conspicuous consumption and run-away credit card debt. Like all of that stuff would be enormously helpful. I think students would retain it better. but we're not teaching that stuff and I honestly don't know why.

Geoffrey Miller: I guess, you know, educators just get locked into these cycles of thinking. Well, it's, it's just so much more important to teach Americans a second language, which they will definitely forget by the time they're

25 than to teach them like, here's how compound interest works and here's how inflation works and here's why.

You know, crypto is more volatile, but potentially has a higher ROI than traditional index tracking stock funds. Like that stuff would be super helpful. but oh my gosh, it might actually require high school teachers to really know things and be able to do things rather than just kind of,

you know,

Steve Hsu: from observing my son, I think the way it works for his generation is that they learn all this stuff from YouTube. And so, like if I, if I, or say when he was taking the personal finance course really interested in a particular subtopic, his way of getting more information, just be to type a search into Google and find a video and watch it on that topic.

And it actually, actually works pretty well. I think there's just enormous surplus value being created., You know, people putting up that kind of content on YouTube.

Geoffrey Miller: Yeah. It's a colossal amount of value. Like if I ever have to do a car repair or just get a part replaced in my car, you just go to YouTube, figure it out. So, you can do kind of real-time research as needed. and I don't think the educational system has taken into account that that kind of lifelong learning is now the norm, particularly from millennials and Gen Z and that they don't really need to sort of memorize a lot of that content.

The other thing I'm struck by, I was just emailing my colleagues yesterday about the new open AI chatbot. and how incredibly good it is at answering questions. It's way more powerful than Google. And I was saying, look, the era of short answers in exams on, you know, online exams that you do at home for your psych courses, it's, it's done.

Geoffrey Miller: It's over it. You cannot give, short answer exams anymore because people can just use this AI open AI chat bot to write a far better answer on any topic in psychology than the student themselves would be able to,

Steve Hsu: Yeah. And you might even be able to cut and paste a multiple-choice question in and get the chatbot to tell you which whether it's A, B, C, or D

Geoffrey Miller: Oh, that's also true. Yeah.

Steve Hsu: Yeah. So, something's got to give. Maybe we have to lock you down inside the lecture hall to take the test or something. But,

Geoffrey Miller: Yeah. One of my colleagues said, well, we, we basically have to go to the. Chinese Imperial Civil Service exams were

Steve Hsu: Yes, we have your own

room,

Geoffrey Miller: You're locked in a tiny room, and you're inspected, so you're not carrying any tiny little, you know, cheating techs sewn into your clothing. That covers the highlights of Confucian philosophy and, yeah, we'll just have paper and pencil exams, maybe.

Steve Hsu: I, I wanted to jump into a topic that I'm super interested in and you're one of the world's experts in. and it feels like just watching you from afar there, there was a phase that you went through where this was your main focus, and you were working with this guy Tucker Max.

and so, you know, it seems to me that, you know, you know more than almost anyone about things like dating, game pickup, artistry. Tinder Hyper Grammy, would you mind talking about that for a little?

Geoffrey Miller: Sure. You know, I've, I've studied sexual selection through Mate Choice since the late eighties, I guess. And it hugely informed my early books like The Mating Mind. I was also witnessing the rise of pickup artist culture in the early two thousands. people like, David D'Angelo, who's now a friend of mine, his real name is Evan Pagan and Neil Strauss was writing that book The Game, and people like, mystery were running around giving young men dating advice.

And I thought, okay, a lot of this is. Edgy and interesting and provocative, and it goes against a lot of the dominant, kind of narratives in society about how men and women interact and what's really important in attraction. That's fine. But by about 2012, I thought, you know, a lot of this, these red pilled pickup artist guys are getting some really fundamental things wrong, and I think some of the advice they're giving is very misguided and not actually very helpful or actionable.

And a bit depressing and exploitative.

Steve Hsu: can I ask, is it fair to say that it's, it's counterproductive for their long-term happiness, but maybe is productive for their short-term, uh, sex having.

Geoffrey Miller: Yeah. I think that was a main, a main issue, but I also saw a lot of young men kind of misallocating efforts in terms of what they were trying to improve about themselves., a lot of them seem to have the impression that, okay, women only care about two things. how physically attracted you are and how much money you make.

And they were really trying to optimize those two things very heavily. and they paid some attention to things like confidence and a little bit to humor, but they viewed those as sort of like superficial and manipulative and not really carrying value for the woman. Whereas they seem to have the impression that if you're a muscular alpha male guy and you make a lot of money, that's genuinely useful to women.

So, I met Tucker Max in Austin, Texas sort of by chance. about 10 years ago. We got to talking. He had actually read a huge amount of evolutionary psychology already, even before we met. And he said, I see these young men, they're kind of lost. They don't know how to go on dates. They were actually taking Tucker Max's early writings from Circuit 2005.

They were taking them as advice books. And Tucker was like, they're not advice books. They're all the mistakes I made of getting drunk and having stupid hookups and taking dumb risks. And, you know, this is bad. We should get together and we should write a book of dating advice for young, single straight guys.

And that's what we did. We ran a podcast that had about two or 300 episodes. We published this book Mate in 2015. And I think we've helped a lot of young men get a little bit better, deeper understanding of evolutionary psychology behind contemporary mating. And also, I hope we've tried to counteract some of the more harmful misunderstandings in the kind of red pill nanosphere community.

Steve Hsu: So, but before we get into the more enlightened Miller Max era of this history, I'm super curious about, you know, p u and game now. When I was in undergrad, in Los Angeles at Caltech, which is at the time in almost all male institutions, there were a lot of us out there trying to meet girls in the sort of LA scene.

And it seemed to me a lot of the observations from these guys were similar to the types of strategies that Caltech guys had arrived at. And not that there's any direct connection, but there, there seemed to be a lot of similar similarities and. It's also true that when I was growing up, I, I, I did stage magic when I was a little kid, so I knew about Pat and misdirection and basically how to control the attention of the audience.

So, I was just really amazed when I saw videos of these PUA guys at work. I think for a while there was even a TV show that actually had wired up a bar in Los Angeles and was just watching maybe this guy mystery or magic or whatever his name is, at work. So, I'm curious what your observations are about just the effectiveness of what those guys were doing in that.

Geoffrey Miller: I think it's, look, it's certainly effective if, if you are a shy, unconfident guy who's very scared of women and you deliberately practice approaching strangers in coffee shops or bars and you get comfortable with rejection and you get comfortable having conversations with strangers, of course, that's absolutely.

hopeful. I mean, it, it, look, this, this even goes back to Albert Ellis and rational emotive therapy, which is kind of psychotherapy back in the sixties and seventies. It's influenced, a friend of mine, Nando Pelusi, who's a clinical psychologist, Albert Ellis said, look, we run around with a lot of irrational fears.

The way to overcome the fears is face them and do the thing you're afraid of. If you're a young man and you're afraid of approaching a woman in a public park, you just have to practice doing that. And he would give assignments to his clients like, look, don't even bother coming back to therapy until you have politely and respectfully approached at least 20 women in a public park.

Right. And gotten calibrated about what, what are your chances of somebody, you know, slapping you or calling the police or rejecting you in a. in a, you know, embarrassing way, and it really worked. You know, it's a kind of desensitization therapy and no doubt in a lot of these face-to-face situations where your physical appearance and your confidence in your charisma is the first thing that women can see.

Then a lot of these PUA tactics can work very well. The trouble is in the era of online dating where women are increasingly wary of any man they don't know approaching them in public, right? They want a guy to approach them, you know, through a dating app, not just come up and talk to them. A lot of the PUA tactics that work in real life might not work on dating apps or might, might not work if you're, you know, talking to somebody in your line of business or your school or whatever.

Steve Hsu: Yeah, it, it seemed like this p u a thing, which was I guess the TS or something. It was the last gasp of that era that you and I were you, you and I grew up in this era where you had to actually, you know, meet girls at bars and parties and stuff like this. And that was like the ultimate development of the sort of male skillset for optimizing that situation.

And then we entered this world of dating.

Geoffrey Miller: yeah, yeah. And you know, it was also op, the, the whole p u a thing was also optimized to a highly alcohol dependent dating culture that was centered around bars and clubs. and I see a bit of a generational shift where, like, with the legalization of cannabis, you know, cannabis is a little more important than alcohol.

with the rise of Burning man culture and sort of renewed interest and psychedelics, that becomes relatively more important than just like, You know, drinking with women and, oh, also, of course, in a lot of communities now nootropics, like people routinely using [unclear] and stuff like that at work. The way you talk to somebody when you are tipsy, it's very different from the way you would talk if you're, you know, high or on Midol or whatever. So, the dating culture keeps changing, but I think the underlying evolutionary psychology principles still say, stay more or less the same.

Steve Hsu: So, when, when you and Tucker Max got into it, I, I think I've listened, not certainly, not listened to 200 of your podcasts, but I've listened to a few. it seems like you were giving very good, balanced advice to young men that I, I'm guessing you have a sense that you helped a lot of young men during that period of time.

Geoffrey Miller: I mean, we wish that, you know, more young men had bought the book. We were sort of like, we hope this is a bestseller. Every author wants their book to be a bestseller and expects it and is sad when it doesn't happen. But I have met, you know, young men who come up at conferences or public talks I give, and they say, wow, Professor Miller, you really helped me.

I read Maid, I listened to the Maiden Grounds podcast. Here is my girlfriend, or here is my wife. And we wouldn't have met if, I hadn't read, you know, your, your book. So that's enormously rewarding. And even if it's only, you know, a few dozen or a few hundred, a thousand young men, we've helped, that's still, I think it's a win.

I think it's, it's still real. Relationships and real lives that are happier, and that's, that's very rewarding.

Steve Hsu: Yeah, it's, I think you should take real satisfaction from that, from. Helping people.

Geoffrey Miller: Thanks. Yeah.

Steve Hsu: Yeah. You know what I, one of the startups that I helped found is, this is one called, it's in Texas called Aram, and it uses forensic d n a to solve crimes using genetic genealogy. So, from very, very small d n A samples, like I think the smallest sample that we've solved the crime with is something like 20 cells equivalent of d n a.

you can, you can get the genotype, find a, you know, first cousin match. The detectives then can go interview people and map out the family tree and catch the bad guy.

Geoffrey Miller: Mm-hmm.

I haven't really been directly involved in it because the company's just, you know, the people that, on the team are just so good.

Steve Hsu: They don't really need my help on any of this, but I, every time they solve a case, which is like, at this point now, several times a week, sometimes, I just take incredible personal satisfaction from it because, you know, you're bringing closure to some. Family whose daughter's been missing for 20 years, and you're putting some serial killer in jail.

It's just, you know, I mean, it, it, it just blows away like writing some academic paper or something like that.

Geoffrey Miller: Yeah. Real life impact is uniquely rewarding. And it makes me kind of sad. I have a lot of very talented, smart academic colleagues who just aren't very, I think they don't really understand how much real-world impact they could have if they were just a little bit creative about connecting their expertise to real world problems.

And in the case of criminal DNA, right? The big win, the big win would come. Criminals in general realized just how easy it will be to track them through their DNA and if that actually deters crime, right? So, it doesn't even show up in the crime statistics, and the people who would get raped or murdered aren't getting raped or murdered because the criminals know, oh, oh shit, that, that, like DNA plus, big ancestry data sets, mean it, it's just way harder to get away with things than perhaps it used to be.

Steve Hsu: I, I think we're probably within a decade of that, of getting to the, I mean, barring something like people outlawing this kind of, those kinds of databases or something, we will get to the point where it really has a kind of order one impact on, you know, likelihood of criminals being caught and things like this.

And it, then as a, as a next order effect, influences the thinking of, you know, at least the somewhat more rational or smart criminals out.

Geoffrey Miller: Yeah. And of course, that's a problem. This a lot of criminals are really pretty dominant, impulsive, and might not be deterred by anything rational. that doesn't involve an almost immediate kind of negative reinforcer. But that's another story.

Steve Hsu: Yep. So just to finish up with dating, I'm curious, are you still following how this whole thing is evolving? I, I had Rob Henderson on this podcast, and I was kind of asking him about, you know, he's much younger than we are, about the Tinder world and hyper gamy and, you know, the effect of the dating apps now on, on, you know, typical behavior of, of young people who are.

Geoffrey Miller: I am following it. partly because we have a lot of friends who are still active with dating apps and active on the, the mating market. partly because my own daughter, who's age 26, lives in New York, you know, is doing her own like mate search and. We often talk about this kind of psych of the young men she meets and is evaluating.

My students certainly talk about it in classes. and I do a lot of podcast interviews where I kind of get into discussions about the current state of things. I, I, I think unfortunately, the rise of wokeness, particularly in dating apps has been really harmful. Like, okay., it used to be a really awesome dating app.

I actually used it about 10 years ago, quite a bit. And it was wonderful. You answer all these questions up to two or 3000 questions. You get matched with other people who give the same answers to these questions about education and values and lifestyle and sexuality and all that. And it was very effective.

Geoffrey Miller: There were millions of people using it. However, when Trump got elected or was running okay, Cupid senior executive seemed to have made a kind of decision that, well, basically we don't want anybody conservative using, okay, keep it. And we're going to kind of, include a bunch of questions that call them out, that embarrass them.

We're going to make it easy for young women to screen out male like Trump supporters and. so forth, and they really politicized the app so that Okay. Cupid became this kind of woke desert of virtue signaling, and it was no longer very effective at actually matching people. And I worry that that has sort of spilled over into some of the other dating apps as well.

Or you end up with dating apps like Tinder, which are just incredibly superficial and basically all about how you look in photographs and that don't actually involve any psychological matching.

Steve Hsu: I thought that Tinder and OkCupid are now owned by the same parent company. Do

Geoffrey Miller: Yeah. Yeah. They just, they kind of have like different market segments, I guess, where Tinder is like for visually obsessed, superficial people. match.com is a little bit more., like frustrated New Yorkers in the thirties trying to finally, be like, ah, I'm tired of the dating scene. I want a long-term partner.

I need to, you know, lock down a spouse asap. And Okay. Cupid is more for the kind of Gen Z woke, virtue signalers.

Steve Hsu: Yeah, I, I, I, it, it's, what was I going to say? I had some contact with the, some of the management of the, the company that owns, all of those dating apps. And we were actually discussing in a very casual way, like at, at any point, whether they would start using DNA information, for some of the matching.

And they weren't ready. They weren't really ready for it, but they were quite interested in understanding where the, where the science and the technology are.

Geoffrey Miller: Yeah. I think, you know, as much as people. See genetic information. If you get them in like intellectual conversation mode, it might be a source of amazing kind of competitive advantage for a dating app to actually include some of that, DNA stuff, just because it would be new and different and kind of, provocative.

And also, I think particularly when people are thinking about long-term mating and spouses and having kids and settling down, they naturally tend to get a little more interested in, in the genetics and the heritability of various traits. Obviously if people are choosing like sperm or egg donors, they tend to set aside Stephen J.

Gould and become super eugenicist and hereditary. They're

Steve Hsu: Yeah. Well, when, when, when the preferences are actually revealed, which is in the office of the genetic counselor at the I B F clinic, then everybody, you know, plenty of people who are, what is it? Woken the streets. They're eugenicists in the sheets or whatever, however you say it.

Geoffrey Miller: exactly. Yeah. and of course, that kind of hypocrisy is so normalized and kind of left intellectual circles that nobody really questions it. but it, it, it'll be interesting to see what happens when the polygenic scores allow, embryo selection, you know, in five, 10 or 20 years, how do the, anti-hereditary, folks’ kind of square that circle once they, they start actually using polygenic scores to make sure that like their kid is actually going to be smart enough to get it to Stanford.

Steve Hsu: Well, you know, right now the issue that we have at, you know, the company genomic prediction, which, which does end reflection using polygenic scores, is, you know, we have, like, there's one class of scientists who. Really world class statistical geneticists, and they're of course, very interested in what we're doing.

And they often write papers where they've done some calculations or simulations of how well it can work. And you know, typically there, they end up basically replicating results that we've already published. But then when they write the paper, they write it with a very negative spin. Or the journal nature or science wants them to have a very negative spin on the whole thing.

But when you actually look at the graphs or the actual numbers that are in their paper, they're almost completely in agreement with the stuff that we've already published. So, there's this already, there's a, there's a kind of schizophrenia there in the way the topic is treated. Now, people who aren't scientists but are very woken just, are against the whole thing.

But you are oftentimes they're against it because it exacerbates inequality, not because they doubt that it can actually.

Geoffrey Miller: Yeah. And., I guess my, you know, my worry is that what, what could end up happening is you get a bunch of tech billionaires kind of using the genomic prediction on down low to influence their kids and create kind of a really like smart, high quality, dynasty. and then ordinary folks are so kind of brainwashed by the dominant antigen narratives that they don't use the technology even if they like, could afford it, even if they were interested, even if it, even if it worked and, you know, that could exacerbate the inequality.

Also, I think, anybody who genuinely cares about reducing inequality should be embracing the polygenic score prediction methods because it's a great leveler. It gives. Ordinary folks, a huge amount of power potentially to increase the capabilities of, of their, their kids.

And it would be a shame if they were kind of ideologically handicapped to brainwash and to not, not use it.

Steve Hsu: yeah. This is the point which, you know, if you are actually a progressive, but a progressive who really wants to rely on interventions that work as opposed to interventions that don't actually, then you should make IVF and polygenic screening of embryos free.

You should just have it covered by the healthcare system and give more resources to families that need the resources. So, if they, if they haven't been successful within their family and going to college and doing well, you know, you should just give them more resources to do more cycles and just get better results.

And a few people understand this. Actually, there are some progressive, bioethicists and philosophers who actually have advocated for this. But by and large, I think this point is lost on most people.

Geoffrey Miller: Yeah. And you know, there are limited resources, but I think if it's a choice between a quote, forgiving college student loans versus making, polygenic score prediction available to everybody, that's a no-brainer. Like you just obviously give people genomic information and power.

and you know, if you combine that with having really good state-of-the-art adaptive IQ testing, then maybe people don't really know, need to go to college as much.

Steve Hsu: You know, if you, if you limit just to the health, you know, reducing health risks, risks for, you know, certain diseases like type two diabetes or schizophrenia, the return to the system, if you have a single payer kind of health system, the return on investment is huge. So, it's actually, it's actually in the long run, cost positive, a cost negative to the.

To make, I V F and polygenic screening free. So,

Geoffrey Miller: absolutely. It's just, you know, that benefit might come after 10, 20, 30 years rather than within the, the, the next election cycle. So, there's not as much political pressure.

Steve Hsu: Well, it requires long-term planning, this kind of thing.

Geoffrey Miller: Yeah. So maybe, maybe there again, China adopts it first.

Steve Hsu: I don't know about that, but, maybe, maybe Taiwan. So, the people I'm working with in Taiwan are actually in the healthcare system, and so they're very eager to deploy the results. not necessarily infertility, but maybe in adult screening, like identifying outliers in risk for certain conditions.

I think that's one of the key wins. So far, no health system has pulled the trigger, but the UK system is studying it and Finland, they're studying it in Taiwan, they're studying it. So, at some point, some countries. Something like a single pair system is going to do this. so, things are definitely moving.

let me turn to maybe our last topic, or not second to last topic, but this is the one I see you interviewed about a lot, which is polyamory. And so maybe just say a little bit for, I don't think my audience is all that familiar with the whole thing, so maybe you could just give a very brief introduction.

Geoffrey Miller: Yeah, so I got interested in this topic, both personally and professionally. So, I've taught some courses on alternative relationships. part of the rationale for that is just a large number of millennials and Gen Z are in alternative relationships where they're not normatively monogamous. That is to say they might have a boyfriend or girlfriend, but they're not sexually exclusive and they have an explicit verbal understanding.

These are our rules, these are our norms. you can have outside lovers; you can have other people you are seeing you're interested in. You can flirt with other people. You know, this is all up for negotiation. And there's some pretty good surveys indicating that a substantial minority of people under 30 do not see sexual exclusive monogamy as the ideal relationship.

Geoffrey Miller: Structure. So descriptively, this is happening. This is a big trend among young, young adults, and I think it's important to understand it. just so that like, you know, if we're teaching a human sexuality class and I've got, let's say a hundred undergrads, probably at least, 10 of them would identify as polyamorous, and maybe another 10 or 20 of them would say, yeah, I've been in an open relationship, or I've tried one, or I'm familiar with it, or I have a friend or family member who's, who's involved in it.

So, I think assuming monogamy as the norm when you're, you know, talking about these things or doing research on sexuality is no longer valid. I also think intellectually it raises a lot of interesting questions about things like, Well, evolutionary psychology traditionally has had a fairly monogamous orientation that says, well, humans evolve to form pair bonds.

Marriage is natural. sexual jealousy is strong and adaptive and natural. and I think the question arises, how much can emotions like jealousy or N VB manage effectively? a lot of [unclear] colleagues would say, not at all. Jealousy is too strong. Nobody can possibly master it.

Geoffrey Miller: Nobody can desensitize to it. It's not possible to have a stable long-term relationship, unless you have extremely strong sexual exclusivity, norms, and expectations. And I think that's just empirically false. There are a lot of very happy, open, or polyamorous relationships where people do master.

Their sexual jealousy to a pretty high degree. So, I think there's a sort of interesting emotion management question, like how do they do that? What, what are the hacks? What are the tactics? And then recently we've had a lot of interesting debates about, what exactly is the function of monogamous marriage in civilization?

So, people like anthropology Professor Joseph Hendrick at Harvard and Professor Jordan Peterson, Toronto have made the argument, monogamy is absolutely bedrock central to civilization, and you can't have a successful civilization without sexual exclusive marriage as, as the norm. And I see these arguments showing up a lot on social media, especially Twitter, where people are like, Any challenge to monogamy is a challenge to western civilization itself.

And I've given talks where I've said, well, look, people do need long-term stable pair bonds to raise kids. That makes sense. But how sexually exclusive really do those need to be in an era of, you know, widespread porn and sex workers and many people doing open relationships successfully, et cetera. So that's kind of my, my slightly long-winded intro to polyamory.

Steve Hsu: When I see you interviewed about it. so obviously there's some intellectual interest. the subject itself, but do you view yourself as an advocate? Sometimes it seems like they're setting you up as the advocate for polyamory.

Geoffrey Miller: I'm not sure if to the extent that I'm an advocate, you know, I'm certainly an advocate of this is worth studying. If you're teaching anything about sexuality or you're doing clinical work with people, it's absolutely crucial to understand it and it's professional misconduct to not understand it. You know, in terms of personal life, I've explicitly said Polly is not for everybody.

I would not recommend this as sort of a widespread mass movement that everybody should pile into, you know, in hopes that this, this is like the route to, to, to sexual utopia. It's really difficult. It's a very steep learning curve. It's not for everybody. It's not easy.. It requires a lot of intelligence, emotional sensitivity, communication skills, good partners, et cetera.

So, you know, I identify as, as polyamorous, I'm in a happy, stable, open relationship marriage. We've got a young baby, so we're not exactly, you know, gallivanting around having a lot of partners at the moment. We're basically in parenting mode, not mating mode. but it's still, it's a heavily stigmatized way to have relationships.

So, I kind of feel an obligation that, like, if I'm personally involved in it, I do have a duty to be kind of open and, and honest and kind of own it because so many people who are into it are extremely shy about it and don't talk about it, even though they're doing it on the down low. so, I have the freedom of having been brutally canceled multiple times by the media that I don't really care, and I can, I can say what I want and pretty much do what I want.

And, you know, some people respond well to it. Other people think I'm just, you know, a degenerate enemy of civilization itself. and I hope if they read my, like my Twitter feed with an unbiased eye, they'll see like, I really like civilization, I like pro natal, I like, you know, a lot of things that they would probably value.

I just have a somewhat different, relationship style and I think polyamory is in much the same position that maybe, Gay relationships were, were at in maybe the early seventies when, they were just getting on the radar. There were just the beginnings of the gay rights movement. and, you know, gay and lesbian people came out of the closet.

Did Western civilization collapse? well, to the extent that it did, it wasn't because of gay and lesbian relationships. Right. That it's still, it's, those are a minority of people. there was no risk that everybody would suddenly turn gay just because we had gay rights. and I don't think there's any realistic, you know, risk that like everybody is going to turn polyamorous within a generation.

I think that's just not going to happen.

Steve Hsu: to, to me it seems like a very high degree of difficulty kind of relationship to execute on, and you know, someone like you who is obviously very thoughtful and mature and really understands a lot about human psychology, I can, I can imagine that you could execute on it without blowing up your, your marriage.

My own experience with this, going back to when I was in my twenties and dating and not ready to settle. I remember being in relationships where I would start dating a girl, a woman, and I would say, Hey, I'm not really ready to settle down. I really like you. I like to spend time with you. and being very open about the situation.

And then oftentimes the, the, the woman, my girlfriend would, or the, the woman would say, oh, I'm okay with that. I, and we, we could even negotiate, like being a nerdy physics guy, I would even like to say to her, wait, do you really, you know, do you really understand what I'm saying? Because I might, like Friday night , go out with somebody else.

Right. You understand that? Right. And I would even, so almost kind of get like the equivalent of a verbal agreement. Never a written agreement, but a verbal agreement that this was okay, but. In every instance when this happened, it always kind of blew up where ultimately, they really couldn't stand that kind of thing.

They were jealous. They did actually want to be monogamous. And so, I just never had it really work out and it always, it always ended up causing a bigger blow up and more guilt on my part, later on. So, I ultimately just thought, well, at least at my skill level, I can't really pull this off.

Geoffrey Miller: Yeah. And y looking back, like in my high school and college experiences, there were a lot of people who were kind of sort of being polyamorous, like they were dating multiple people. They were having multiple ongoing relationships. but they never talked about it. They weren't sort of radically honest about it.

They didn't negotiate their norms and expectations very clearly. They didn't really have any resources, you know, books and videos and stuff they could go to about how do I manage my jealousy, how do I. make sure I get the degree of commitment and, and respect that I want. So, polyamory as a culture, you know, has only been around since the early to mid-nineties.

It's very much an internet-based culture, but it has been working hard to kind of try to develop these best practices. I'm involved in a big international research consortium where we're trying to identify which best practices actually predict relationship satisfaction and stability, both in monogamous and polyamorous relationships.

We hope to have those papers coming out soon. and I would just emphasize to listeners who are like, oh my God, this polyamory stuff sounds terrible. Look, the, the choice that young people in the twenties face is not really realistically polyamory versus settle down in a trad life, monogamous marriage, the real choice they're facing is, Do they do ethical polyamory, or do they do a typical Tinder hookup culture where nobody's really honest about who else they're seeing?

Or are they still, you know, looking for mates even though they're kind of, sort of in a relationship where everybody's delaying, moving in together or getting married for years and years? Right. So compared to Tinder hookup culture, I think polyamory is generally better. But honestly, for a lot of people, you know, their best option would be like, find a really good mate in college.

Settle down, get married, start having kids in your mid-twenties, and don't wait till your like late thirties before you find a.

Steve Hsu: You know, in, in, uh, I, I heard you describe one model of polyamory in which there, there's a very explicit, prioritization or hierarchy where you say this, this is the person that I. Really, you know, I'm going to devote most of my emotional resources to, and I have a long lasting relationship with, and I just want to be very clear with you, you're, you're, you're someone I'd like to spend time with, but you're lower on the priority list.

and I, I think, you know, if, if you're really explicit, obviously maybe you discharge your ethical responsibilities. But of course, you know, whether that situation can really work out is still, you know, looks challenging.

Geoffrey Miller: Yeah, I think the important thing is, you know, you need to be confident that the person who is your so-called primary partner, like your main boyfriend or girlfriend or spouse, is actually very well suited to you and would not be easily replaced by somebody else. and it certainly helps to have mutual commitments to each other, like, living together, having a mortgage, having kids, you know, being intertwined socially and financially and so forth.

but once you've got that, you know, the great liberating thing is you can potentially have, secondary partners, other lovers you're seeing once in a while who don't necessarily have to live in the same city, who don't necessarily have to share all of your like, political values or aspirations, who maybe you, you enjoy their company for a day or three days, but where you know, you wouldn't actually be able to enjoy like, living with them full-time.

Geoffrey Miller: So, it kind of broadens the possibilities for people you can. interact with, have feelings for, feel romantic with, and you don't have to always hold someone up to this incredibly high standard of like, oh, if, if they're not the best possible marriage material for me, then they're of zero interest to me.

Right? That, that's what I think is very harmful. Treating people as kind of inferior and disposable if they're not perfect, long-term, mate.

Steve Hsu: I'm, I'm guessing you, given your crypto interest, you might have followed the F T X Alameda events, recently.

Geoffrey Miller: indeed. Very closely

Um, and I guess they reportedly were in a molecule in The Bahamas, and I think Caroline Ellison, the CEO of Alameda. Actually, wrote something favoring the, I think she referred to it as the Chinese imperial concubine system with some hierarchy of, of mates or something like this.

I, I, I, I, are they, are they students of your, of, of your teachings?

Geoffrey Miller: Oh my gosh, I have no idea. Like, I've never had any, like personal contact or met SBF or Caroline Ellison. I'm not in their social circles. I mean, the funny thing when FTX blew up was all the polyamorous people were like, oh my God, I hope they focus on the crypto and the effective altruism angle and not on the poly angle. And all the effective altruists were like, oh my God, I hope they focus on the crypto and the poly stuff and not on the effective altruism angle. And the crypto people are like, oh, this molecule, that's what messed them up. It's not the crypto thing. So, everybody was kind of hoping that the other subcultures would be blamed for what happened?

Steve Hsu: It seems like in the coverage I've followed, the molecule stuff is one of the most minor things. So, you know, number one is, you know, grifting [and] fraud. Number two is maybe [effective altruism] and then number three is probably polyamory in the coverage.

Geoffrey Miller: Yeah. I think that's accurate. That's what I've seen so far. I would say there's so many businesses that have gotten into trouble. Gotten into trouble about the kind of office romances that polyamory is almost irrelevant. Like this could have easily happened if the people involved were in, very respectful mono, you know, monogamous manila relationships. so, I think the poly is kind of a, a, a red herring,

I, I don't think anybody is, Attributing blame to poly for this disaster. Just, it's just another way to knock these guys by saying, oh look, they were weird in this other way as well. y Yeah.

Steve Hsu: I think [effective altruism] could have played a big role because one of the hypotheses that's going around even by [effective altruism] people is that SPF F was willing to, you know, make these 51 49 kind of wagers because from a kind of [effective altruism] perspective, it, it might be, might be justifiable to take on just a huge amount of risk in the way you run your company so that, you know, potentially you could double your money and do even more good in the future. So, that seems possible to me.

Geoffrey Miller: Yeah, there's been a huge amount of discussion within [effective altruism] and I'm, I'm, I've been involved in effective alt autism for about six years, and I've taught courses on it, and I'm very active on the [effective altruism] forum and oh my God, I mean, we've, we've written maybe millions of words back and forth in the last few weeks about, you know, was the [effective altruism] to blame for this and what role did it play.

And I think my take is that Sam Bankman Freed had an extremely weird and extreme version of Expected Value Theory that he was sort of applying to his business decisions. That would not be endorsed by literally anybody I know in [effective altruism] that that was just so far out on the edge of, of risk seeking that like, it, it doesn't make any pragmatic sense that he was, either thinking that way or making, business or investment decisions based on it.

Steve Hsu: So, the, the story that I'm, I'm certainly no expert on this in this area, but I, but I did follow, the FTX story pretty with great interest, the way SBF described why he left Jane Street in the first place and started trading, bitcoin. was that he had become convinced that he needed to be on the efficient frontier for risk taking in his own life so that Jane Street was too safe. And if it's safe, then by efficient market theory, you're giving up a lot of upsides. And so, he thought he had to push himself to some place where he really was taking on a lot of risk. And if you fast forward that to the way he was running FTX, that, if, if he continued with that philosophy, then he would also have to basically run FTX in a very aggressive way that he would actually be risking, you know, collapse blowing up.

and that, that does seem consistent actually with, you know, a lot of it could have just been negligence. Like they just didn't really know how to run the accounting and back-office functions and stuff. And so, they just got into a terrible situation. But some people are attributing, you know, actual sort of Machiavellian intent to a lot of this.

Geoffrey Miller: Yeah, I think the great tragedy here is if Sam Bankman free had just taken a few more courses in evolutionary theory rather than economic theory, he would have, he would've stumbled across this thing called Optimal Foraging Theory, which analyzes how animals make investments about finding food and avoiding starvation.

And when optimal foraging theory first started in the, the sixties and seventies, all the biologists thought, oh, animals are obviously going to maximize expected value, expected calories they're going to forage in a way that, you know, make sure they get the maximum expected number of calories as they go around the world.

Well, it turned out animals weren't doing that. Instead, what animals were doing is minimizing the risk of starvation over the cold night or the cold winter., right? Because animals can only store so much food or fat or, or whatever they need when they're, they're chili and they're hibernating. And so, everything from hummingbirds to grizzly bears, actually makes these optimal foraging decisions to avoid ruin, not to maximize expected value.

And what SPF did not do, was, you know, minimize the probability of ruin. If he'd actually done the math on like, okay, let's roll FTX forward. Let's roll Alameda research forward for a number of years. What is the likelihood we go bankrupt making the kind of extreme high-risk bets that we're making. I think that probability approaches one, like there was no way they could have actually run those companies the way they did without.

Going bankrupt or being in a position where they had to dip into, you know, use funds or get into massive legal trouble, they engineered a business that was absolutely doomed to fail sooner or later. And that's what happens if you try to maximize expected value and you ignore the risk of ruin.

Steve Hsu: Yeah, I agree with that analysis a hundred percent. But it is possible that S B F understood that and, and still went for it, And, and I think that's how he describes the reason why he left Jane Street, because if he had just stayed at Jane Street, he would've become fabulously wealthy, but not a, you know, not a billionaire.

And, but when he went off to trade Bitcoin, there was a good chance he was just going to be ruined and not, not succeed at all. So, he might have just been repeating, consciously, repeating this, you know, wanting to be at the risk frontier even if, if you have a chance of being ruined. the, the other aspect of this, which.

You know, maybe weird, even within [effective altruism] circles are, you know, he does have a background in physics and some people believe in the many worlds’ interpretation of quantum mechanics and in that interpretation, sure, he blew up in this world, but he might have been maximizing utility across the multiverse. So, there are other SPFs who became the richest man in the world and most influential men in the world or something.

So, this is all speculative, of course. I don't know what he was really thinking, but it's fun to discuss.

Geoffrey Miller: Yeah, I think that's actually a fascinating issue. I was talking with some locals here in Albuquerque about that very, that multiverse perspective. Like, okay, are there worlds in which SBF you know, in five years, a trillion dollars to do really good things and to save the world from bad AI and so forth.

And did he believe that, like, it doesn't matter what happens in the failed, failed worlds, as long as there's a few worlds in which humanity succeeds long-term and spreads throughout the galaxy, blah, blah, blah. But I think the problem is like his risk assessment was simply empirically off that it was obvious that the way Alameda Research and FTX were actually run, that they were doomed to fail.

Geoffrey Miller: And there's like almost no timeline in the multiverse where he would've actually succeeded doing the kinds of things that he did. And that's the alarming thing that I, I think, was a fundamental failure in their kind of risk management, whether it was multiverse or not. The risk of financial ruin and reputational ruin is just extremely high.

Steve Hsu: Yeah, I agree with you. I, I think even if he was knowingly really taking on as much risk as one could reasonably take, they probably screwed up in the actual execution of it. And, you know, without intending to, you know, blew them, you know, took on much more risk of ruin than they thought they were taking on, or, or that they, they were aware of.

So let, let me, let's close out with one, something which I think you just very did very recently, which is some work on AI alignment, which is, I guess, solidly within the [effective altruism] world. and I think you, I think you were encouraging these, I think you said something like, these young AI scientists need to learn a little bit more about psychology or evolutionary psychology in thinking about, how to, how to succeed in AI alignment.

Geoffrey Miller: Yeah. That, that's, it's a really fun, well fun and horrifying topic to work on. And of course, I come in and I'm like, well, actually, the field in which I happen to have expertise is going to be absolutely crucial to solving this colossal real-world problem. but I'm, I'm not entirely without some background.

I mean, at Stanford in grad school, I did quite a bit of work on neural networks and genetic algorithms and machine learning and did a postdoc at University of Sussex that was all about evolving, neural networks to control autonomous robots. so, 30 years ago I was pretty into machine learning, and then I kind of, you know, went off into this ab psych thing and lately I've returned.

Thinking about AI because I think it is a crucial issue. AI progress is very rapid. There's a lot of interesting ideas about how to try to make AI safe, but I think a lot of those ideas are just incredibly naive about how human behavior works and how human conflicts of interest operate. and in particular, you know, a lot of the AI alignment people run around saying, we must, we must align AI with human values.

But they seem to mean something very peculiar by that human values in general as a sort of lowest common denominator of what all humans would reasonably want if they were perfectly rational and farsighted. And I've tried to make the argument in various essays, for [effective altruism] forum. Look, humans can barely align with each other.

Parents can barely align with their own children. Do you know, coworkers can barely align to get a company to function. Faculty can barely align in faculty meetings to do what's good for the department. so, the idea that AI should be easily able to align with sort of all humans generically seems very naive both in terms of the general game theory of how conflicts of interest work and in terms of the psychology of how tribalism and self-interest works.

Steve Hsu: Yeah, I agreed with your take. I, I think these guys are, I, I, you know, I would say that I've been telling this to marry people for about 10 years, that it just seems like, uh, you're not going to succeed, for, you know, for reasons that you just elucidated. But even if you, you know, even if you were to succeed in hard wiring a quote, set of values or priorities into the AI, you know, humans who have the same quote values might still choose radically different strategies to achieve those goals.

And that still might, you know, involve something bad happening to humanity So it just to predict the behavior of some being, which is much more intelligent than you and much more complex than you in terms of its cognition, it just doesn't seem like one could actually, guarantee that, it would always behave in a way that, you would be happy with. So, I've always thought AI alignment was, you know, of course you can't prove anything about this, but it just seemed very, very implausible that it was solvable.

Geoffrey Miller: Yeah, and I think there's a certain culture in AI alignment that is, very nerdy and a bit Asperger and worship's formalization and is very technical and sort of has a view that, if you're not an extremely skilled computer programmer, you shouldn't even be talking about AI at all because you're not an expert.

And I have some respect for that kind of viewpoint. But on the other hand, you know, there's two sides of alignment. There's the AI side and there's a human side. And if you seriously want to align with humans, you have to align with them as they are and not as you want to abstract them into being.

Geoffrey Miller: And so, for example, some of my essays have written about. Well, if you want to align with humans, 80% of humans are religious and are involved in organized religion and religious values are important to them. But guess what? People believe in different religions. You know, the four big ones are Christianity, Islam, Buddhism, and, and Hinduism.

And they don't always see eye to eye. So which religion are you going to align your AI with? hundreds of millions of people in each religion. and of course, the AI alignment. People hate talking about this because they're all atheists and they think religion is stupid. And why should we worry about that?

And in 20 years there won't be religion. and I do worry about that. AI alignment research can give the illusion that, oh, very smart people are working on AI safety, and they will solve it by the time we get really powerful, dangerous ai. And I think, that's actually not likely to happen and we won't solve alignment.

So, my personal hunch at the moment is that we should, actually slow down AI research deliberately and, and seriously and just take a little bit of a pause before we develop it. And I'm not talking about a super long pause on evolutionary time scales. I think if we just wait, you know, just a few centuries, just a few, a few generations, that would be reasonable and maybe we can solve AI alignment problems by the year 2,500 or so.

Steve Hsu: Well, I think what you're saying might be wise, whether it's feasible in a world where the Chinese and the Americans are competing for tech supremacy, doesn't seem very promising to me that that's going to happen, but, in, in fact, we're probably going to rush headlong into it the way we rushed headlong and into nuclear weapons, and possibly with very bad consequences.

Geoffrey Miller: Yeah, I mean we should, we should wrap it up soon cause I got baby the needs attending to, but I, I actually do think despite the fact that we got a geopolitical arms race for AI between the US and China, that there are things that could happen that might lead, the general public in both countries to stigmatize AI so heavily to morally stigmatize it, that it becomes, as disreputable to work on AI as it currently is to work on behavior genetics or.

Steve Hsu: It could happen, it happened in Dune, right? Do you remember the, but Lori and Jihad, that they had a near death experience as a species. So, that's, that's how that happened. But, yeah, but I want to let you go because, obviously the baby and the family come first. It's been really a great conversation and, hope, hope we bump into each other in person sometime.

Geoffrey Miller: That'd be great. Yeah. I appreciate it, Steve.

Steve Hsu: All right. Have a great weekend.

Geoffrey Miller: You too.

Steve Hsu: Okay, take care Geoffrey.