Rachel Wallis is pregnant with her highest-IQ embryo, with a 99.99th percentile polygenic IQ score. She and Wolf Tivy discuss what this means, and how in the 20th century our concept of “humanity” narrowed into an ideological straitjacket.
Transcript
Intro (00:00)
Wolf: Hello again. Welcome back to The Students. I’m your host, Wolf Tivy. I’m here again with Rachel Wallis.
Rachel: Woo-hoo.
Wolf: Last time she was on the show, we talked about whether you should do polygenic embryo screening to improve the quality of your offspring. So, it was a bit of a debate. I took the side of the traditional way of doing things. She took the side of the more futuristic way of doing things. Now, this time, it’s become a lot more practical.
She’s pregnant.
Rachel: Yay.
Wolf: So congratulations, Rachel.
Rachel: Thank you.
Wolf: She’s pregnant with the first of her super babies. I don’t think it’s the first super baby in the world, but it’s among the first. So, that’s very exciting. And so this time, I wanted to bring her back on and have a discussion following on that one about not just the why and the whether, but the how. So, Rachel, welcome back.
Rachel: Woo.
Wolf: Thank you for coming on. Why don’t you give us a little bit of the lowdown on the story of how your baby came to be, I guess, and what you selected for.
How we selected the embryo (01:17)
Rachel: Yeah, absolutely. So, depending on how online you are, people may have seen the viral tweet from Gene Smith or perhaps from Cremieux, looking at the IQ chart of a bunch of embryos and being like, “Ah, this family has selected that top one.” So that is actually me and my husband.
Wolf: Right. Congratulations.
Rachel: Thank you.
Wolf: You’re famous.
Rachel: Twitter famous.
Wolf: Statistically famous. Yeah. What was on the chart? Remind me there. And yeah, what does it mean?
Rachel: Basically, for polygenic embryo screening, you can look at IQ, you can look at various probabilities for different disorders like Alzheimer’s or diabetes. And so let me get into a little bit of why my husband and I selected this particular embryo and how much weighting we put on each of these different factors. So basically, when I look at the last 50 years of medical science, I think we have made enormous strides in basically everything that is not to do with the brain. So, people don’t realize this, but we haven’t exactly cured cancer, but we have effectively cured a lot of different types of cancer. And so an example of this is that my uncle got metastatic melanoma in the early ‘90s.
Wolf: That’s very New Zealand.
Rachel: Very. It’s not an unusual story for a New Zealander. But so he, at the time, basically got six weeks to live, and that was about correct. And so he died at 46, which was extremely sad.
Wolf: Oh, no.
Rachel: But about 10 years ago, a new drug came out that basically cures this highly aggressive form of skin cancer. And so people with his diagnosis now, I think have 85% five-year survival rate or something. And this is true across so much of the disease sphere. And so when I project out what problems are my future children going to be dealing with by the time that things like cancer become relevant to them in their 50s, I actually think that we’re probably, if you use, I think it’s the Pareto principle, I think it’s very likely that a lot of this stuff is just not going to be a big issue for them. The area where this doesn’t hold true is basically anything to do with the brain.
So our treatments for things like schizophrenia or Alzheimer’s or dementia haven’t really made much progress in the past 50 years. If you’re—
Wolf: Some notable fraud contributing to that.
Rachel: Yeah. And so there was a huge thing with that, fraud in Alzheimer’s research that cost billions of dollars and hundreds of millions of lives, which is very upsetting. But so if we haven’t made that much progress in those, chances are, I doubt we may not make much progress by the time my kids are facing these issues. And so I preferentially weighted stuff like Alzheimer’s, schizophrenia, bipolar disorder. And so when I looked at my embryos, my husband and I were very lucky in that our overall risk of those things is pretty low.
And so my kind of heuristic was that if an embryo is under the population average for one of those things, then I’m basically like, that’s a yes. I think—it’s a good—
Wolf: Yeah. Don’t bother totally minimizing things that aren’t really even a problem for you.
Rachel: Yeah, exactly. And so, we basically just ended up going with the embryo with the highest IQ score because that embryo—also just didn’t have a very high chance of lifetime Alzheimer’s or schizophrenia or stuff like that.
Wolf: And it was also something of an outlier, right? I remember seeing that on the chart that you roll the dice, you see what you get, and then if one’s an outlier, it helps make the decision.
So that’s interesting. Okay, so you basically selected for IQ—secondarily with the notion that you would screen out anything that was suspect on other brain-related disorders—even though that didn’t come up.
What do these predictors actually mean? (05:04)
Rachel: And so basically, we have this embryo, and if you look at the chart, I think the top embryo is 3.62 standard deviations to the right-hand side of the curve. And so people asked, “Okay, so is this saying that this embryo is going to be 3.6 standard deviations of IQ?” Which I think is 154. And the answer is unfortunately no. The biggest—
Wolf: Why don’t you tell us about that. How do those correlations work? How do you actually predict, let’s say, real IQ from... Or what’s the distribution of a trait given a certain level of outlier on the polygenic scores?
Rachel: Yes. And so one thing I would say is that there’s kind of a naive way that you can do things, which is if you have a predictor of a certain power, and then you have parents of a certain IQ and a certain number of embryos, you can predict about how much more are intelligent than the parents, their top embryo should be, in theory. So given me and my husband, basically any couple, if you get 50 embryos, you would expect your top embryo to be about 10 to 12 points higher than the parental average. And so that’s an easy sort of naive way to do things. If you want to get into actually predicting specific IQs, it gets much more difficult.
And so the various factors at play are my husband and I’s PGS scores, so where we are on that chart, basically. Then our actual scores in real life, because the PGS score is not perfect. In fact, it only accounts—for about 20% of the variance, and we can get into kind of what that means a bit more. But—
Wolf: Right. So you’d expect basically some kind of regression to the mean of about 80% of the variance of the polygenic score—where the mean is the family background.
Rachel: Yes. And so regression to the mean is a concept in genetics where even if you breed two outliers together, typically you will see regression to the mean in their offspring. But this isn’t the mean of the entire population. You’re right, it’s the mean of the extended family group.
Wolf: It’s the mean of the population they were selected from.
Rachel: Yeah. And so you have to account for regression to the mean. You have to account for the power of the predictor. And so an interesting thing is the predictor gets better if you can calculate the mean of the extended family. And so then there is a question of if you’re doing polygenic embryo selection, does this mean you go out to all of your cousins and say, “Hey, can I have some spit in a tube so that we can better do our PGS score?”
And the answer is you don’t have to do that because we have 50 embryos. They give us an understanding of what the mean looks like.
Wolf: Oh, you get a pretty good idea of what the distribution is.
Rachel: Yeah. And so that one—
Wolf: Well, and if you’re just picking the highest polygenic score anyways—it’s not like with more information about the population it actually changes the ordering there. It just changes what you would predict given a certain level of polygenic score.
Rachel: Exactly. And so getting into the actual prediction of what the embryo is going to be is hard. I think our best faith prediction for her is going to be probably around 139, and then it could be slightly lower or higher. I think if you look at the bell curve, she’s 93% likely to be higher than 120, and so the break-even point of 50% is 139, and then—50% to be higher, 50% to be lower. But that has quite a lot of variability, but—she’s—
Wolf: Because the polygenic score simply isn’t that determinative—at least as far as we’ve measured. Actually, why don’t we get into a little bit of the technicals of why it’s not super determinative, because we know, or at least we believe based on twin studies and so on, that it’s something like 50% to 80% of the actual variation of intelligence—for example, is genetic. But we’ve only actually measured the polygenic score impact up to 20%, it sounds like. So is that remaining part mechanisms we don’t know how it works, or is it just we haven’t found those genes, or is it a gap in our ontology, like we’ve got the wrong idea with the twin studies, or we’ve got the wrong idea with the—polygenic scores? What accounts for that?
Where does that remaining variance come from?
Rachel: Yeah. And so an interesting thing about this figure of that we think 50% to 80% of intelligence is heritable, that doesn’t necessarily mean that it is a particular gene causes, say, a increase in the efficiency of the brain mechanism that makes a person smarter. It could also include concepts like genetic nurture. So you could have a theoretical gene that maybe makes mom more nurturing, or more attentive to the child, or more likely to have books in the home, or more interested in literacy, and that’s—
Wolf: Yeah. Another example of that is if your genes make you more attractive, maybe people treat you better, give you better—resources or whatever, and so you become more intelligent for that reason, or some other—whatever other sort of dimension that society is discriminating on.
Rachel: Yeah, exactly. And so heritability isn’t just talking about genes. It’s sort of like a weird stack of mechanisms that are still quite mysterious to us. And then the interesting thing is the other 50% to 20% isn’t necessarily environment in the way that we think about it. It looks like shared environment really has very little impact on people’s adult IQ.
Wolf: Yeah, that’s always one of the most surprising things is we can’t—measure the impact of parental style or household or any of that.
Rachel: Yeah. It appears that that 50% to 20% is not, as you would think of it, schooling or nurture. It’s more something like—
Wolf: It’s just random
Rachel: random variance. The entire... Yeah.
Wolf: Well, of course, there’s always going to be random variance, but if it’s actually just random.
Rachel: Yeah. But so the entire thing is quite confusing. And so when you have these scores, it’s interesting that the scores have far outstripped our understanding of the mechanism of what they’re actually describing. We don’t have a good idea of why the particular genes that we’ve selected as being potentially part of the score, what they do. So that’s an interesting thing.
Basically, the current predictors that we have are a lot less powerful than they could be because they depend a lot on the quality and amount of data that we have to make the scores. And basically, I think we talked a little bit about this in our last episode, but the majority of the data that people use to calculate these scores comes from the UK Biobank, and so you have the genome, and then you have a not very good IQ test. I think it’s six questions.
And so you can—
Wolf: Yeah, so weak data.
Rachel: Yeah. We have pretty weak data. There is data available that if the companies had access to, it would make the predictors way, way more powerful. So there’s the Million Veterans Project, which is—
Wolf: Right. There’s the military data.
Rachel: Yeah. And so if that was available to the companies, I think you could at least get Probably an extra 10% variance or something ridiculous. I’m very much guessing on that. But we do have the power to make stronger predictors. And—I think—
Wolf: And aren’t there people working on that, getting that data released? That’s what I understood.
Rachel: I think so, but I think it’s a very low political priority, basically.
Wolf: Yeah. Understandable.
Financial costs and benefits (12:53)
Rachel: But that does raise the question, I sometimes think of an alternative scenario, where say it turned out that there was a, as yet unidentified, but now identified type of super lead that happened to be in school rooms across America—that was causing a decrease in IQ of 10 to 12 IQ points on average for each kid. And then this came out. It would be a massive scandal, and we feel like we have—to get the super lead out of the schools and like—
Wolf: But if the super lead is just data ignorance, people don’t seem to care.
Rachel: Yeah. And I’m just like, if you said, “How much would it cost to get the super lead out of schools?” People would be like, “That’s so heartless.” Surely getting the super lead out of the schools is like, there is no cost that is too high. But when it comes to, well, actually, we could spend a billion dollars getting data that would make this stop.
Wolf: It sounds like we already have the data.
Rachel: Yeah. We don’t need to actually to spend—
Wolf: We just have to release it
Rachel: we just have to release it, make it publicly available.
Wolf: Yeah. Mind you, of course, then people have to go and adopt the technology and all this, and we get through all the questions like last time of—how many people are actually going to do this, whether you should do it, et cetera. But—
Rachel: there are some really interesting cost benefit questions. And so I’ve tallied up basically how much this entire process cost me. So how much did it cost to get 50 embryos? And it was about 100K. So about 50K went to the polygenic embryo screening company, and then it was like 50K to do the entire IVF process.
And so if I wanted to go out and get more embryos, I think the marginal cost per extra embryo would actually be less because you have this big fixed cost of basically—doing the thing. And then you’re sort of like, would it make sense economically for every family in America, if they wanted to, to spend 100K to get on average an extra 10 IQ points for each of their kids?
Wolf: Well, those numbers are also going to come down.
Rachel: Really, they are going to come down really quickly. But I think it’s just like, I would love to see an economist do the math on this and be like, would it make sense for—
Wolf: Yeah, I mean, I—
Rachel: the government to say subsidize this and say free IVF for different couples of course.
Wolf: Yeah. My understanding is a marginal point of national IQ is worth an enormous amount in terms of GDP per capita—and so on in terms of its impact.
Rachel: And I’ve done some small calculations on how much it’s worth to the actual family in terms or the child in terms of increased earning potential over their lifetime, and also even things like—
Wolf: It’s much less than the public benefit
Rachel: But it’s still the upfront 100K is still way less than the expected benefit per person. Even if you calculate things like the marginal IQ point on how much you can get in terms of the scholarship to a university, you save enormous amounts of money, basically.
Wolf: Okay. But that’s—
Rachel: Also the weird thing about cost savings, so it’s one of the interesting things about IQ is that people get sick a lot less. And part of that is overall disease risk reduction, which is pretty crazy. But also, people with higher IQ get into less accidents. And this actually shows up in the data.
Wolf: Makes sense. A lot of accidents are caused by stupidity.
Rachel: There’s a really funny article by Cremieux on Twitter, where he’s looking—at the relationship between concussions and IQ. And you’re like, “Okay, well, it would make sense if you get a concussion that it affects your IQ.” And he’s like, “Actually, no. Basically, people with lower IQ are way more likely to get concussed because they get into really silly accidents.”
Wolf: The causality is the other way around.
Rachel: Yeah. And it’s like, “Oh, okay.”
Wolf: Interesting. Yeah. But we can talk about these public benefits and these sorts of marginal benefits to earning potential and whatnot. But that’s not why we’re doing this, right?
So I guess I wanted to dig into that a little bit, the why. But have we covered for now what there is to cover on the statistical mechanisms, how this is done, what exactly you selected for? Anything else you want to say on that front?
Rachel: I think done for now. I think that I should potentially prepare a more thorough investigation into that for our next episode and be like, “Okay, let’s get into population stratification. Let’s get into assortative mating and how it affects the predictors and stuff.” But—
Wolf: Right. With the real technical details of—of how do you make the super baby.
Rachel: I think I need some diagrams would be very helpful for the audience.
Cultivating virtue (17:16)
Wolf: Yeah. We’ll need diagrams. Okay. But why don’t we talk then about why? Why IQ?
Why do this whole exercise in the first place? And we touched on this previously, but—I think it’s a very rich question to ask. I kind of always come at things from this philosophical perspective. And why we do things we do, what it is that we care about, these are a lot of the big questions, and so I’m just curious for your take on your thought process.
Rachel: It’s funny because I think that I automatically fall back on when I’m trying to convince people of why this is a good idea, I fall back on technocratic utilitarian arguments of being like, “Well, the marginal IQ point is worth however many points of GDP.” And then I actually interrogate myself, and I’m like, “Is that the reason I’m doing this? No.” But then I try and investigate further and be like, “Why am I doing this?” And I think the answer is we all have intuitions and moral intuitions, and it’s actually sometimes those intuitions are mysterious even to ourselves.
But if I try and drill down to it, I think it’s something more to do with something like the good life. And it’s like I believe that virtue is real and goodness is real and like—
Wolf: And a lot more of it.
Rachel: Yeah. And I went to Great Books college, and so I was sort of steeped in the classical tradition. And the ancient Greeks were very concerned about how do we educate our—
Wolf: The quality of offspring.
Rachel: Well, yeah. And also, this was the central... It’s slightly different nowadays. We have this idea of what is education for, and a lot of the time the modern answer is sort of preparing people for the workforce. But for I think a lot of human history, it was moral education, like how do we create—people, men and women, who are courageous and virtuous and good.
And we had varying amounts of success, I think.
Wolf: I think—
Rachel: Moral education is really hard.
Wolf: No, I think classical education works pretty well on that. I mean, people came out quite brave and strong—
Rachel: Oh, yeah, I do
Wolf: at the height of it.
Rachel: Yeah, but you still, it was very much less than 100% efficacy.
Wolf: Oh, of course.
Rachel: And so like—
Wolf: Yeah, well, and this comes back into these matters of what’s genetic, right? And, well, it turns out one of the big findings of, let’s say, the last 100 years—is just how much the blood matters and the genes matter—relative to the school.
Rachel: Oh. Yeah.
Wolf: It’s like, yeah, you can educate people. Probably, you should educate people. Our own education system has very mediocre outcomes relative to the possible variance that has been seen in history and so on.
But, genetics is also this incredible component of that. And like you were saying, the ancient Greeks knew all this. They knew about nature.
Rachel: Yeah. They did.
Wolf: And they knew that nature was the ultimate thing. And so—
Rachel: But they had no way to drill into it or figure out how to—yeah.
Wolf: Well, they had the analogy of animal breeding, right? And so it was like, okay, well, if we have someone who has great nature, they’ve got great phusis. They’ve grown impressively into an athlete or a warrior or something.
Rachel: But even they—
Wolf: We want them to have the kids, right? They had that idea.
Rachel: They did, but they also recognized the problem of reversion to the mean hits people really badly. And I think I can’t remember which dialogue it is, but I think there’s a dialogue where they specifically talk about the problem of great fathers having kind of mediocre sons. And this is like a really a big—
Wolf: Yeah. I think there’s something like that in, is it Theognis or something? Condemning people for marrying for money because it causes that.
Rachel: Yeah. But even if you have, say a great king and he marries a great queen of great virtue and intelligence, and they have kids. Like the—
Wolf: They don’t always turn out.
Rachel: Yeah. And mean reversion is absolutely brutal, and it’s particularly brutal for outliers. And this is true of outliers for intelligence—or strength, but also outliers in terms of virtue. And so you could have the kind of Ottoman solution to this, which is you have 100 children and then make them fight to the death.
Wolf: Yeah, but this doesn’t work very well because it’s like—
Rachel: It doesn’t work very well.
Wolf: again, it’s reversion to the mean. It’s like, okay, you got some outlier on some frankly pretty crappy test. And how much does that transition to any kind of lasting legacy? Well, not very much.
Rachel: Yeah. Well, the—
Wolf: This is also the problem with the sort of Bene Gesserit theory of good breeding, right? It’s like, oh, it’s all about these pairings and particular people. So no, this stuff happens at the population level. And we talked about this last time, I think. It’s about overall differential fertility patterns.
Rachel: But you basically, in order to have, say if you have two great outliers who are an outlier couple, you have to have a lot of children to have a child who’s probably going to equal their parents in whatever dimension you’re looking for. And so this is why polygenic embryo selection, to me, is so exciting because you can do the selection of 50 kids in a lab, basically. And people could say, well, you’re not actually doing the selection because you’re looking at one metric, and you’re not really doing anything to put them into the world.
Wolf: Yeah, and the metric is not... We don’t even have great data on the metric. Yeah.
Where is this going for future generations? (22:31)
Rachel: Well, but, the technology is the worst it’s ever going to be. And the technology is already okay. And so for me, I’m so excited about the possibility of what our ancestors for 3,000 years, at least, have been like, how do you create a more virtuous next generation? And suddenly we have this—
Wolf: Yeah. It turns out we have more science about this now, actually. You can—
Rachel: Yeah. We have a tool that they didn’t have. We have the possibility of actually achieving this.
Wolf: Yeah. Not only these particular tools, but the overall Darwinian kind of revolution in biology—since the 19th century has also given us much more technical understanding of how this stuff works.
Rachel: Yeah. But so for me, I’m so excited about the potential for each generation to be greater than the last. And I love greatness. I love excellence. It brings me so much joy.
And even if I look bad in the comparison, and I end up—with my kids and their kids. Yeah.
Wolf: Yeah. I mean, what we want to be that small and humble fish that crawled out of the mud. Right? And compared to some glorious descendant of ours—who accomplished much more than we could ever imagine.
Rachel: But so for me, this seems like very intuitive stuff. Of course. And yet when I—
Wolf: Yeah. I mean, right, it seems like the straightforward implication of the science of you think about what’s good in the world. It’s like, well, I like life. I like beautiful things. I like greatness.
All the human virtues, well, they come from somewhere, right? And we can make more of that. And it’s good.
Rachel: And we may even—
Wolf: And it seems obvious, right? But—
Rachel: And also in the pursuit of this, maybe we find out more about the nature of these things, which the philosopher in me—is so excited about that, where like—once we get the actual good data and be like, actually, it turns out that this gene—
Wolf: There’s something else
Rachel: that affects, I don’t know, risk perception or heart rate is actually correlated to risk-taking or something like that.
Wolf: Yeah. Wouldn’t it be interesting to push along this dimension with new and better tools until we find the interesting choices? Right? Where we actually do have to face a fork between intelligence and health or courage and stability or—whatever kind of thing it is. But we don’t get there by ignoring it.
Virtue and courage (24:48)
Rachel: And I do think that when you drill into it, I think you are going to get weird outcomes. So, Aristotle has this idea of the virtuous mean. So when he conceives the virtues, you know this, but just for our audience in case they don’t know, is that the virtue is a mean between two extremes. So in the case of courage, you have on the one side cowardice, and on the other side, foolhardiness of just—taking insane risks for no possible reward. And then somewhere in the middle is courage.
And then the reason that you say that the mean as opposed to the midpoint is because means can sometimes be skewed more to one end versus the other. And so I think that—
Wolf: Yeah, I think the other thing that people don’t appreciate about that concept is there’s sort of two... I almost prefer the kind of two by two square with, you can have both foolhardiness and cowardice at the same time. Or you can be well-calibrated.
Rachel: Yeah. Oh, it’s an interesting one the way you think about it. Yeah.
Wolf: The well-calibrated knowing what is to fear is the true definition of courage, right? And there are multiple ways you can make a mean. And really what you’re looking for is the synthesis of the virtues of the two extremes.
Rachel: Yeah. And I think—
Wolf: But yeah, it’s a complicated concept
Rachel: the Punnett square idea of virtue. That is an interesting one. Did you come up with that?
Wolf: Well, it’s just my interpretation of Aristotle.
Rachel: Yeah. I haven’t thought of that. I need to think that through. That’s actually really interesting. But so, and I think this is kind of true when you look at real-life people and real-life families.
So I love the Kennedys. I think that they’re a really interesting family, and I love reading about their history and stuff. And I think they’re an example of a family who either have a genetic or cultural predisposition towards probably more the foolhardiness end of courage. And this means that they have outlier outcomes, but also end up dying in ways that normal people probably wouldn’t die.
Wolf: But is it an excess? They’ve done pretty well.
Rachel: Yeah. This is true, but I think it’s like you look at them as a family, and they are so prone to risk-taking that this results in acts of great heroism. I think, was it JFK who swam his... There was a plane crash, and he swam various members of his platoon to safety at great personal risk. And you’re like—
Wolf: Yeah, sure. Right. Sorry
Rachel: that’s incredible, but they also end up going into probably one of the most dangerous occupations in the world, which is the American presidency. And you look at the death and assassination rate of that, that’s really high. But also they’re kind of willing to take the risks.
Wolf: Which they contributed to.
Rachel: Yeah. But they also do end up kind of hurting themselves and I think there was the light airplane crash, which was a completely unnecessary risk of one of the family members. And then you have RFK Jr., who gets himself into very adventurous situations. He got a brain worm, I think, from—his adventuring across the world. You see him sort of playing with roadkill and various wild animals.
I think there’s a Kennedy nephew who actually briefly fought on the front in the Ukraine, and they’re an incredible family, but you can sort of see how courage, and getting outsized influence in the world can come with a risk tolerance that also, in some circumstances, results in untimely death.
Wolf: Yes. Perhaps it’s worth it. I’m reminded of, I forget who said this, sometime in the late 19th century or early 20th century, they were sort of reflecting on the sports going on at Harvard or something. And it’s like, well, we have our young men playing polo, and occasionally someone gets injured. But what does it matter if we break a few necks for the improvement of the race?
Overcoming the “humanitarian” straitjacket (28:47)
Wolf: There’s this benefit to risk-taking. And you actually get this incredible benefit from it. But that also reminds us here that this perspective that you seem to have and that I share is in something of a minority these days. A lot of people don’t see this as the obvious implication of the science and the world. People are maybe much more concerned about the fairness.
They want to level things out, the Harrison Bergeron kind of attach weights to the strong people worldview.
Or insofar as you have people kind of with this desire to improve things or to push into some risky unknown, a lot of that is more acceptably expressed these days by building apocalyptic AI systems, right? The impulse to create the Übermensch is still there. We didn’t get rid of the impulse to create the Übermensch, but now it’s only expressed as over at Anthropic, where they just—put out this ad where they’re busy burning down the world and or whatever it is, terrorizing everybody with their scary AI system, and this is the development of, this is the direction that mankind has to go.
Anthropic takes the very apocalyptic kind of direction in their marketing, but a lot of people talk about the utopianism of it or the inevitability of it. There’s various narratives that fly around in AI. But point being, it’s seen as sort of this natural progression that oh, of course, humanity’s going to be replaced by machines that are going to be superior. We’re going to create the Übermensch, but it’s going to be the machine. And I find that interesting, right?
It’s like, okay, we’re human. Why is it that the feeling of improving ourselves or the impulse to improve ourselves has been kind of marginalized relative to killing ourselves for the sake of machines or something? That’s interesting.
Rachel: It is. I wonder... It’s like it’s, yeah I wonder if it has something to do with notions of fairness. And so something that when I bring up this topic to people, a lot of the time people bring up the idea of an arms race that this technology is potentially opening up of—the selected for and the unselected, and is there going to be a—
Wolf: Whereas with AI, it’s okay because everyone will die fair.
Rachel: Yeah. I do wonder if that’s psychologically part of it. It’s like the idea—
Wolf: No, I think this is precisely what it is. This is an idea I’ve been wanting to develop a little bit more and really explore is the concept of humanity is a straitjacket on man. It’s like the fact that we’re human means that we’re obligated to be limited in some way, right? It’s not just as a matter of material fact or possibility—that we are limited, it’s that we’re obligated to be limited—by the nature of humanity. Humanity is this concept that man must not go beyond, right?
And whereas the machine is allowed to go beyond the concept of humanity. And the machine is, in fact, allowed to threaten to genocide humanity, but it’s okay. People don’t have this reaction to it the same way they have a reaction to one person, a human going beyond humanity, right? And I find that so interesting. How did this concept of humanity become a prison?
Rachel: I feel like when you think of humanism, there are two competing conceptions of it, and one is humanity as is, is good and should be kept—
Wolf: Preserved. Right.
Rachel: And then there’s the—
Wolf: Preserved and bottled and yeah.
Rachel: Yeah. And then there’s the idea of humanity as becoming, in the midst of becoming itself. And I can sort of see why maybe it’s that these are two ends of a spectrum, and somewhere in the middle lies wisdom. And last time we talked a little bit about Chesterton’s Fence and the idea that if you go too fast and unthinkingly and without wisdom, that maybe you lose something that was essential. And so I think that’s the idea of humanity should stay as it is for as long as possible for fear of losing something.
But I’m just like, I don’t know. I always fall back on biological metaphors, but I think that if you look at a species, you kind of are always in this arms race with the world because the world itself is changing. And so to—
Wolf: Right. Yeah, you think you’ve got everything locked down. You think you live in this end of history, European utopia. But actually—
Rachel: The world is changing under your feet
Wolf: the rest of the world’s leaving you behind. That’s what’s actually happening. And so you have to continue the struggle in some sense.
Rachel: And so I think any wise idea of what humanity is has to be one that is continually changing and evolving because that’s what life is. It can’t be static.
Transcendent humanism died in the war (34:00)
Wolf: Yeah, I think that’s a very interesting point about the two different versions of humanism. And I’ll put those somewhat in historical context, I’ll call them pre-war humanism and post-war humanism. Right? And now probably the dividing point is somewhat earlier than that if you really start looking at the causes of all this. But I think the concept of humanity as Eleanor Roosevelt’s concept of humanity, which is the one that I’m calling a prison—is very much came about with a large amount of bombs dropped in 1945 and with the Declaration of Universal Human Rights from the UN and all these sorts of things.
And in pre-war, or let’s call it Renaissance humanism or whatever, when I reflect on what is it that’s actually good about the human, why do we care about the human? What is the thing that makes the human interesting? If we take humanism, you go to a dictionary, you go to the encyclopedia, what is humanism? Humanism is the belief that humans are the most interesting thing in the world, right? Or have the most potential.
Or are the philosophical subject, right? The human as the being that is capable of reflecting on itself, as the being that’s capable of improving itself. And there’s all these things that we can point to as what is it that’s interesting about the human from an objective, transcendent perspective. It’s not just like, “Oh, I’m a human.” It’s petty chauvinism.
And when you look at those things, it’s like, okay, it’s this list of transcendent potentialities that aren’t connected to being a 10-fingered ape. And then you have post-war humanism, UN human rights humanism, where it’s like the human is an empty stomach that needs to be fed. The human is a person with inherent dignity that can only be threatened by progress, right? Or progress only amounts to elevating the sort of baseline and not kind of going beyond, and not the reflection, the philosophical kind of pushing through to some new frontier, not becoming more, not being the center point of historical development. And I think that’s maybe after the horrors of two big, extremely bloody wars, people are like, “Okay, well let’s just cool down for a little minute and try to make things nice for a while.”
I get it. I get it, right? But it feels like now we’ve reached the logical conclusion of that idea. Which is that, okay, I’m starting to realize, and I think people are starting to act on the fact that humanity is now this—prevention mechanism, right? It’s this limitation.
It’s this system of moral limitation that can be transcended through creating the AI Übermensch.
So I think a lot of the extreme investment in AI, obviously economically it’s interesting and so on, but I think economically, if you just do the calculation, there’s some way in which improving IQ or birth quality or something is vastly more economically interesting, than the median projection of what AI, as we know it, will probably do.
Now I think the outlier scenarios are more interesting in AI, but there’s this known very large piece of value on the table, right? But we don’t take advantage of that. It’s not where the money’s going, and it’s not where the government interest is going.
Rachel: And the question is why? And I—
Wolf: Yeah, I think it’s the moral strictures of the concept of humanity.
Rachel: And I think you’re absolutely right to bring up the wars because it’s interesting that pre-war, I think that the idea of a transcendent idea of humanity and of a perfectible man was like—
Wolf: And as something that this was the project of modernity, right? Was to pursue this.
Rachel: Early 20th century, and this was across the political spectrum. You have the rightists who are like, “Yes, we can do eugenics to make man better.” But you also had the communists being like, “We’re going to—
Wolf: Yeah, new Soviet man.
Rachel: “We’re going to create the new man.”
Wolf: Through technology or social engineering, we’re going to create a new man. You had the progressives in America. You had everyone, right? Everyone.
Rachel: Everyone was like, “We are on board. We’re going to...”
Wolf: You look at all the universities around here, like Stanford. You go down to Stanford. What was Stanford interested in? Well, improving a man.
Rachel: Yeah. And you look at all the propaganda posters and just like, I—
Wolf: Right. Yeah, yeah. It’s man as aspiration, right?
Rachel: The optimism of the Soviet propaganda poster. You’re just like, “Oh my God, these guys are-”
Wolf: Yeah, it’s these big, strong, handsome guys—who work hard and accomplish the future, right? It’s man as aspirational ideal. Not man as ball and chain.
Rachel: Whereas nowadays, we’ve actually got a movement of the most and the most recent statues around, is how can we have the average person celebrated as opposed to how can we—push man towards something new and exciting and perfected?
Wolf: Man as ball and chain.
Rachel: Yeah. And then this question of why did the world wars have this effect on people? And I think you and Ben Landau-Taylor had a really excellent conversation about how the mechanics of war affect society. And I do wonder if for a very long period of time, people might push back on this, but I think for a long period of human history, war was partially eugenic. That basically you could, the way that fighting worked, that if you were faster or stronger or braver or smarter, your chances of survival were a little bit better.
And so it was—
Wolf: Well, it wasn’t just a little bit. It was the bravest and the strongest who led from the front—would do very well—and get their lion’s share of the loot and the glory, and then they would have this enormously elevated social position at peace time.
Rachel: And then the—
Wolf: You read the Iliad, right? And it’s like one of the characters is explaining to the other, “Why do we, the princes, have such an elevated social position? Why do the people cater to us? Why do they bring us our food? Why do they do all this stuff for us? It’s because we, the brave princes, are out on the front lines-” doing the hard end of the fighting.” Right? And this was the social order, until more or less the mid 19th or early 20th century. And yeah, I think there’s something to that.
Rachel: There’s something that happened of that our weapons of war, basically mean that no matter how brave or strong or technical you are, you could be mowed down by a machine gun just as well as anybody else.
Wolf: Yeah. The machine gun overthrew aristocracy.
Rachel: And it’s interesting because I remember reading about like—
Wolf: And democracy
Rachel: how I think the English aristocracy used to complain about the invention of the crossbow. And also how the—Japanese used to complain about the introduction of firearms as opposed to sword fighting. And they were saying it’s a less noble weapon. And I was like, “Ah, it was kind of cope from losers” and like—
Wolf: Which it actually is.
Rachel: And I’m like, actually, I think they might have a point of how it changed—
Wolf: It is going to lead to a less noble society.
Rachel: Yeah, because it means that you basically, the archaic, noble masculine virtues actually don’t produce the same results that they used to. And so maybe there is a sense that as our weapons technology, it suddenly exploded in terms of the amount of destructive power, and then we felt that we used it wrong. Which I think is actually fair. I think that like—
Wolf: Yeah, there was definitely a lot of technological, reactionary sentiment, after the war leading up through the 1960s, that led to kind of everything grinding to a halt in the Western world in the 1970s.
Rachel: And I think maybe it was sort of like we had this immense optimism about how technology was going to change the world for the better. And then with the world wars, we kind of like, “Wow, maybe we’re not wise enough to wield these new powers appropriately.” And I think it’s a fair thing to be afraid of.
Wolf: Yeah, and I think—I think that’s, on the surface, yeah, that was the idea, right? But these ideas are always adopted by systems of power and pushed by particular live players and these transformations are somewhat engineered. And I think it’s out of scope here, we should definitely talk about the historical analysis of what happened—between 1930 and 1980. Or maybe 1970, whatever. But that was an enormous change.
You can draw those boundaries in different ways, but enormous change to society. Suffice it to say, before that period—Everyone kind of understood that the implication of Darwin’s theory and science and development and modernity was that we were going to improve man, and that we could improve man, and that we knew how to improve man. And we had some fights over the details. We’ve had a lot to work out, but now the science is even better. We know how to improve man, and it’s a lot less dangerous than the ideas that people were throwing around then.
But nonetheless, we’ve kind of lost interest. I think if I would summarize it, it’s that ideologically the idea of humanity has taken deep root, and the idea of humanity has become a prison. And this is a direction that I want to explore, but maybe we can put that aside as well. But point being, I think there’s something in the ideological water supply that means that we find it less offensive to build a silicon Übermensch than a biological Übermensch.
The silicon ubermensch (44:04)
Rachel: And I think that there’s, it probably is a little bit out of scope talking about the World Wars and the... But I do think that there is a psychological wound at the heart of modernity, which was—the Second World War, where we were basically like, we were going to be powerful, and we’re going to create the future, and then we used these enormous powers to do massive mass scale murder. And then there’s a feeling of maybe we weren’t meant to wield these powers. And how can we disempower each individual and ourselves and try and divest the power as away from each of us as humanly possible? How can we divest it into committees and into large scale thought processes?
Which is why I sort of wonder if maybe the empowered LLM is the end result of this line of thinking, of wanting to bureaucratize as much as possible. In some ways, the LLM represents the collective unconsciousness of humanity, maybe the collective consciousness of humanity.
Wolf: Yeah. It’s the collective official text. Right. Yeah.
Rachel: And so consult, if you’re thinking about doing something, consult the official manual, and now the official manual is more consultable than ever. We have the LLM—which is the amalgamation of all of the human corpus, and it’s like—the most average.
Wolf: Yeah. Lovingly curated by experts to make sure it never says anything off color.
Rachel: Ew. And so I do think that it’s less confronting to have the AI Übermensch than it is the biological—Übermensch, which I think is sad because... And people might be surprised by this because I’m Mrs. Transhumanism, whatever, but I do consider myself a humanist. I love humanity. I think humanity is awesome.
And so—
Wolf: But which humanity?
Rachel: Yeah. Well, and that’s the thing.
Wolf: Is it the one that is the sort of by nature, the center of the project of history, by definition, or is it the one that is the empty stomach that needs to be filled?
Rachel: Obviously the first one. And we’re the same on that. But I can see... And so I get a bit sad when people are like, “Let’s create the AI Übermensch,” and I’m just like, I hope that we can get the biological Übermensch online fast enough to kind of deal with the AIs.
Wolf: Well, let’s also say some words in defense of the people who think that this whole Übermensch thing is—is maybe a bit too scary.
Rachel: Many people are like, “I don’t want the Übermensch, biological or AI. Can we just not?”
Wolf: Yeah. But yeah, I dunk on them as kind of the partisans of this idea that I’ve been arguing against here, the partisans of humanity. But I think that is the dominant viewpoint among most people. But I think that nonetheless, when you look at that, the outcome of that viewpoint is Ted Kaczynski, right? And it’s this reactionary conception of history that basically says humanity is good, the biosphere is good as it is.
There is no process within technology or modernity or development that can improve it. There’s no hope in that way, so we just have to stop that. And when I go and read “Industrial Society and Its Future” or “Anti-Tech Revolution”, it’s like what I’m struck by is the lack of any positive direction to that. And so I think the reason people keep... Even though it’s kind of this dominant idea, this sort of reactionary concept of humanity, people find these ways out.
And I think that’s what AI is. It’s like people have found a way out of humanity, and that’s what they’re trying to look for there. And I think this is sort of necessary that the philosophers are going to find themselves in this unsatisfying end of history and say, “No, actually, that’s not what we want.” Francis Fukuyama himself predicted people would look at this utopian end of history world and be like, “Yeah, actually, I’m bored of this. This is stupid. I want out.”
And I think there’s something to that. It actually, when confronted in its fundamental nature, isn’t actually compelling.
“The choice between death or change” (48:30)
Rachel: I also think that if you truly care about humanity and the continued existence of humanity, or, and I’m using humanity both as who we are right now and also our offspring who might look very different from us. You want it to be as anti-fragile as possible, and if you do that, you need to look at our place in the universe, and it’s like, what if there is another super intelligent species out there that doesn’t want us around? What if there is—
Wolf: What if we’re alone?
Rachel: What if we’re alone?
Wolf: And no one else is going to do anything with all this—
Rachel: That would be so sad. Or the universe is actually a pretty dangerous place. There could be a massive pulse of radiation from a different—part of the universe that could fry all of us instantly. And it’s—
Wolf: Yeah. If you’re not growing and improving, you’re just accepting this unacceptable risk.
Rachel: You’re accepting death eventually, basically. And so it’s sort of like you have the choice between death or change.
Wolf: Yeah. It’s inherent to the nature of life to seek improvement. And so I think that’s ultimately what defeats the end of history, sort of humanity prison from the inside is just the, when you accomplish it, it’s this endpoint that then there’s no going beyond it. It’s this boring endpoint, right? And all you can do is say, “After me, the flood.”
Right? It’s, “We’re done here, and I just want everything to die.” This is the sense, I go and visit some of these boomer strongholds, where people think like this, right? It’s just like, “We don’t want further development. We just want everything to freeze and have our nice retirement, and then everything just die.”
And like—
Rachel: “Why can’t we just freeze California and never build any more housing or build any more power?”
Wolf: Yeah. And this is—
Rachel: “And just live as we have.” And you’re like, okay, well, it’s nice that you currently live in the Shire, but like you know the—
Wolf: It’s uncompelling.
Rachel: Well, yeah. And it’s also—
Wolf: It’s uncompelling
Rachel: it’s dangerous. Like the Shire, yeah, is a self-contained unit, is kind of nice. But it’s like there are dark forces massing, potentially outside your borders and you need to be able to have the strength—
Wolf: Right. So when—
Rachel: to contend with them. But I do take your point. Even if you’re not... Like, I am using existential threat arguments, but even if they weren’t an existential threat, I agree with you that I would also still want to pursue greatness for its own sake.
Wolf: Yeah. Well, life wants to grow. And so this, I think, to bring us back to our initial subject here, this is having children, having better children, having better children through technological means, through the miracles of science, whatever. It’s an expression of this idea of life as a self-improving process. And, yeah, it’s interesting that it’s so underrated.
I think it’s not as taboo as it used to be. I think these things are developing. People’s attitudes are developing as it becomes more possible, more concrete, less like an abstract fear, and more like a concrete opportunity. And so, maybe people will realize the potential of these things. As we argued last time, I’m the more conservative one, I think, whatever.
Let’s just do it the old-fashioned way. But nonetheless, if people need this Faustian technological kind of vehicle for their aspirations to reproduce, okay, fine. People have always needed religion to reproduce.
And, so okay, we’ll do it that way. But I think maybe it could catch on and almost displace the psychological impulse towards the AI apocalypse, which is sort of the competing possibility. And I find that very interesting.
Investing in the next generation (52:27)
Rachel: I also think that, something I want to add is I think having a multi-generational view of the world is good for you and good for the world. And so something I think about with my own family is that I am very aware of who my grandparents and great-grandparents and great-great-grandparents were, and their names and their stories. And they definitely actively sought to create a better world for me. And to be part of building a new country in New Zealand—and also making enough money that they didn’t spend, that they could then pass on to the next generation. And I think that growing up with that knowledge of they never knew me, and they knew that they would never know me, but they loved me enough to sacrifice for me, is just—this immense warm feeling of being part of this chain of people who are doing this.
And so, with my parents, they created a wonderful life for me. And it’s, I think in some ways difficult to improve on them as parents. But one of the ways I can do that is be like, not only can I give my child a good financial and cultural patrimony, but now perhaps a good genetic patrimony as well.
Wolf: Yeah. You can invest in a new kind of wealth.
Rachel: Yeah. I think that’s good. I think that it’s good that we are continually investing in the next generation and being like, “I want your life start to be better than mine.” And because people sometimes are like, “Oh, well, what if your child doesn’t turn out to be a super genius?” And she’s on the...
Okay, so she’s 93% likely to be an IQ over 120, but what if it’s a 7% under that? I’m like, “I don’t care. I’m still going to love her.” Like—
Wolf: You roll the dice and you do your best.
Rachel: You roll the... But it’s just like, I did this for her. It’s like, I want her life to be as good as possible, and I can look her in the eye and say confidently, “I really did try to make things as good for you as I can.”
Wolf: Right. Hmm. I think we’ve had a very interesting conversation here. Meandered through all manner of philosophical topics that are related—and historical topics. I don’t think that’s an accident.
I think that is where the interesting questions to ask around this stuff actually are—is in these more foundational questions. But maybe next time we’ll talk more about the technical aspect—when we have more results to show and more prepared statistical methodology and results and—and detail. I think that’s the technical part. The actual opportunity is also extremely interesting.
Rachel: Absolutely. And yeah, please comment on the videos to give us ideas of what you guys want to hear us talk about on the subject.
Wolf: All right. And with that, we’ll leave you until next time. Thank you for watching.
Rachel: Thank you.