This podcast is brought to you by the Albany Public Library main branch and the generosity of listeners like you. What is a podcast? God daddy, these people talk as much as you do! Razib Khan’s Unsupervised Learning
You know that genetics plays a huge role in our health and more people are using genetic testing to determine risk for diseases like cancer for themselves and their kids than ever before. So I want to tell you about Orchid. It’s the only company that does whole genome testing for embryos. Testing before your child is born. If you're doing IVF, that's the clear choice now, because now you can reduce the risk for 1000s of single gene disorders, including heritable forms of autism, pediatric cancers, and birth defects. Check them out at orchidhealth.com
Hey, everybody, this is Razib with the Unsupervised Learning podcast here. I don't usually say this, I'm just gonna say this real quickly, please rate and review the podcast on whatever platform etc, etc, etc. It's good for me, and it's good for the guests. So it'd be great if you did that. So we could get more ratings so that we can be recommended to more people. Because I think some of the content here is very good as they say. I am here today with Seth Stephens-Davidowitz. And he has written a book very, very quickly. And we're gonna get into that near the tail end of this discussion, but the topicality of the book is sports, it's NBA, it's really data driven, I read it, it's a really quick read, as they say. I mean, for me, a lot of it had to do with the fact that just so you guys, data nerds out there, there's a lot of data, there's a lot of insight. There are some books, I'm not gonna say they have a lot of fat, but yeah, they have a lot of fat, this book does not have a lot of fat, it like gets to the point. And you know, the the chapters kind of tell you where you're going. And some of the stuff I kind of knew from Seth’s twitter feed but, or elsewhere, but he really dug deep into the data, it was great. I really enjoy it. He is also the author of “Don't Trust Your Gut: Using data to get what you really want in life” I think that came out was that two years ago, Seth?
Seth: Yeah, year and a half ago, I think,
A year and a half ago. Okay, so that was his previous book, and he's gonna be writing a lot more books in the near future. And we're gonna go to why and how, those of you who follow Seth on Twitter, and you should already kind of know that, but yeah, so who makes the NBA data driven answers to basketball's biggest questions, and we're gonna go into the details of the book shortly. But first, I want to ask, why did you pick this as the first topic for your new process of - And I should say, you know, we were talking about affiliation. I think you were with the New York Times for a while. We did talk once, when you were there. But right now, I think, you know, you're a freelance data scientist. I mean, this is, this is how you're going to be writing books, presumably, using the skills that you have, as I think you're, I think the term is data journalist. I don't know, it's kind of like, not used too much anymore, for some reason. But it was a big thing. A couple of years ago, peak, Nate Silver. In any case, why did you pick the NBA for this? You know, first trial?
Yeah, well, I'm a huge basketball fan, basketball nerd. So it was fun. There are some subtle reasons. So I wrote the book in 30 days using AI tools. They're just like subtle reasons related to AI that I use, what used to be called Code interpreter, what's called data analysis from ChatGPT+, and basically, to use data analysis only works with datasets of a certain size, if the dataset gets too big, it doesn't really do the analysis, it has memory limitations, and you it'll write the code for you, but you have to run the code outside data analysis. So the NBA was kind of a good use case, because the datasets are pretty small, you know, they're only been 4500 NBA players in history. So you know, I think it's better to study that than to study like daily stock market data or something, or minute by minute stock market data where I think data analysis wouldn't be particularly useful.
Yeah, actually, I think most of most of my listeners will know how, how professional basketball works. But as you mentioned, in the book, there are some countries where basketball is not a big deal at all. And so the number of players for example, baseball has a lot of players. I think a lot of stuff to record happens in baseball, right? Whereas the NBA has a smaller number of, you know, like five players at any given time, I think Is it is it 11 The roster? Yeah. Okay, so a roster of 11 So what you're getting at here is that the data, the scale of the data is tractable. I mean, there is tractable data, first of all, and the scale of the data is of a reasonable size so that you could, you know, like, I'm imagining here, like, your code is like going through SQL lines or something like that, or sub database, you know?
Yeah yeah. It's just and it's, I mean, the other great thing about studying the NBA is there's so much data available, you know, any sport, this is true, there's just, you know, basketball reference has data on every player where they're born, you know, all their stats. You know, there's data on the recruit rankings of players, there's data on, you know, the Combine performance of players, there's just a lot of data out there. So tractable datasets are that are very useful to play around with and data analysis
So let me just ask you real quick. What team are you a fan of?
Seth: A Knicks fan
Yeah, okay. Okay. All right.
So, yeah, I grew up in the 90s. And the Knicks were super good. That was the Patrick Ewing era. They kind of spoiled me. You know, I thought they'd be good forever. But they, they've been sucking forever, but now they're good again.
Yeah, I actually I stopped following people who know me know this. I basically stopped following professional sports in the mid 2000s, mostly because of time, and whatnot. And, but I am a Celtics fan. I was a Celtics fan, but I was a Celtics fan. Well, I was really, really little in the 80s. So I don't remember the glory years in much detail. I just kinda remember. You know, but you know, when you're seven, you're not sports isn't the same, spectator sports, you know, I think it's like an adolescent male thing. By the time I was an adolescent, though, Bird retired. And those were those were dry years mostly. And Reggie Lewis died. So we kind of like lost it was like a lost half a generation, you know. And then of course, there was a championship around, you know, whatever, like the 2000s. But that was a whole separate thing. By that I stopped watching. It was like, for me, like basketball, it was like, kind of like, I paid attention. I was like, obsessive, where the Celtics would have, like, kind of their nadir, you know, over the last, like, couple generations. You know, now I think they're, they're better honestly, I don't I don't pay attention. But yeah, like, it's great. I love I loved watching sports. Maybe I'll do it someday. What I have time right now. I don't. But it's great that you still keep up. And you know what, I gotta say that we were talking about your voice earlier? You sound you could be, you know, I could be hearing you on ESPN.
No they said, some people say that they're surprised. You know, we talked about this earlier that they're surprised when they hear my voice. Like, when they read my books, or whatever they expect, like a soft spoken intellectual. You know, like, you know, a high pitched voice or something. And then, you know, they they hear me, I'm like, I sound like a dentist from New Jersey or something.
Yeah, lfg, right. It's just like, Yeah, that's great. So you keep it real, you know, to your roots. But in terms of, let me just ask you, we'll go kind of like more sequentially by the chapters. But what was the what was the biggest surprise of you working on this book?
I mean, the biggest surprise, not content wise was just like how powerful AI is. But that was before I - So I started writing the book, because I was using code interpreter. I'm just like, holy crap, this stuff is amazing. This is, you know, things that used to take me four months are now taking me four hours or whatever. So I'm like, but like, throughout the book process, I'm just like, anytime I'm like, I wonder if code interpreter could do this. And so frequently, the answer was, yes. You're just like, you know, just like little things, but they add up, you know, it'd be nice for this chart for every color to correspond to the color of the college. So you know, Kentucky would be dark blue, Michigan State would be green, Arizona would be red, and you just tell code interpreter that and in a second, it comes back perfectly. Like all the colors, it would take me forever to do that. So I think the biggest surprise was just how useful how amazing code interpreter is for data analysis.
Okay. Yeah, I want to I have a lot of questions. I have some questions about that, actually. But I'm gonna hold that until the end. I wanted to talk a little bit about the NBA. So let me just say one of the, I use the term learnings, whatever people use that term now it's normal word. You know, millennials have been popularized it. So one of the big learnings here, although, you know, this was as I was reading the chapter, I immediately knew where it was gonna go, because if you follow the NBA, you kind of know this. It's quite obvious. And also, partly it's just, I think my biological science background is it's known that humans these are almost all male, but not exclusively, above six foot three, start to get klutzy. Okay. Our body really is not - your mortality really starts dropping above 6’ 3”, from what I know, for these are almost all males. I have a friend, he's 6’ 6”. And I think I've mentioned this before, there's been a lot of podcasts. So I think I've mentioned this before, he's 6’6”, in grad school, we got to call his girlfriend got to call it my other friend, she got a call that he was in the hospital, he had got a lung tear while he was in the shower. And I was like, wait, what, like, does he have some condition? And I'm actually using a fake name, let's call him Rob. And so, you know, she was like, no, no, no, he said that his doctor has told him that there's always a risk of this, because he's tall and thin. And you know, people that are tall and thin they're told that there's always a risk that this is going to happen. And your probability of if it happens once, it's probably going to happen another time, at least before you die. And he could have, he could have actually died. If he had to go to the hospital and whatever, he had to have therapy. So he just your lungs do not scale to a certain size, the morphology of like various things, organs, you know, just like systems, your body, etc, etc. So, you know, there's a reason we're not all seven feet tall giants. You know, I think Andre the Giant if those of you who used to follow the WWF back then, you know he died early, like these big guys, they tend to die early. It's just how old is the body is not scaling. So one of the-
It also depends on whether it's due to a disease or not. So some super tall people have pituitary gland diseases. You know, so the tallest people in human history are like eight feet tall. And they all die very, very early. Because pretty much all of them. It's due to a pituitary gland growth disorder.
Yeah, I think Andre the Giant was in that class. Yeah. Yeah. There's other people that are just tall.
Yeah well, pretty much all NBA players are just tall, like it's just genetics. Their parents are really tall. There have been studies, genetic, there have been genetic studies. Shawn Bradley, he was a 7’6” player, and they studied his genome. And he had I think, 198 variants connected to height enhancing variants. So he just had a crazy draw of genetics, but there's like one player George Muresan, who was 7’6” in the NBA. And he was a grown disorder. His parents are like average size, and he just had a pituitary gland disorder.
Yeah, I remember. I remember his features. Like, let's just say you could tell that, you know, he's a little different. That's all I'm gonna say
It's so weird that like, you know what, one of the things in the book maybe you're getting to is the importance of height and making the NBA. So each inch doubles your chances of reaching the NBA throughout the height distribution, which is really interesting. So, you know, if you're 6’1”, you have twice the chance of reaching the NBA compared to if you're six feet, if you're 6’2”has twice the chance compared to your six one, kind of all the way out as far as we can measure up to about seven feet. What that means is you have like a one in seven chance of reaching the NBA, if you're seven feet tall, which is just completely insane. Like, I don't think there's any other pursuit, glamorous pursuit where one trait gives you such insane odds of making it. And what that means is Yeah, height is such an advantage. It's so weird to think that you can reach the top of one of the most popular sports in the world, just due to a pituitary gland disorder. But you can't you can in basketball because heights such an advantage. Yeah,
Yeah. So I was gonna get to that. So basically, though, the probability is way higher. So I think Epstein like David Epstein, like he had a book sports gene. And I think there was a quote in there like, you know, what percentage of guys that are seven. They're all guys, I think that are seven feet tall. are in the NBA. You got to number one out of seven. Which is that's Americans, right? Yeah. That's crazy, right? One out of seven are going to make it to the NBA. Obviously, it's like one out of millions. Well, it's actually one out of like, 3.8 million I think if you're like 5’10” or something,
Seth: if you're 5’10” or below yeah.
Which was like okay, was that that makes sense. But But the irony or not the irony, if you think about it is the short guys are much better athletes. Can you talk a little bit about that?
Well yeah, so you know, anyway, we can measure. The taller you are among NBA players. The taller you are, the worse the athlete you are. So you're slower, your vertical leap is lower. You have less agility. You shoot worse you’re worst performing in the clutch. And I think it almost follows directly. It basically follows directly from the fact that height is such an advantage in the NBA. So basically, you know, if you have a think about if you're, if you're a five foot, if you're six feet tall, and you have, let's say, a one in a million chance, or whatever it is of reaching the NBA, then to reach the NBA, you have to have one in a million ability on everything besides height. But if you're seven feet tall, and you have a one in seven chance of reaching the NBA, you need only one in seven ability to reach the NBA, which is not that crazy. So you know, it's above average, but it's not that much above average. So, you know, the average six foot NBA player just is an insane athlete, you know, their speed compares to the best sprinters in the world and their vertical leap compares to some of the best high jumpers in the world. And, you know, they're shooting and coordination is, you know, as good as it gets. But the for seven foot tall players, it's not that way at all, you know, their vertical leap is just a little bit above average. And their speed is kind would be below average on a high school track team and they shoot, you know, they shoot worse than a lot of high school - than most high school basketball players. So they're not that extraordinarily skilled people, they're just tall, you know, because the the height is such an advantage. And you see that you know that George Muresan the NBA based on a pituitary gland disorder and you know, their stories of, you know, Shaq, Shaquille O'Neal, Phil Jackson once said that, Shaquille O'Neal would have won 10 MVPs. If he just worked hard, and somebody told Shaq this, and you think he'd be really offended? Like, how could Phil Jackson, my coach say that, say this. And he basically admitted that it was right, and that he never really practiced hard. And that it’s probably true that he could have won 10 MVPs if he just worked hard. So I think you know, among these giant players, you know, they might not have to practice that hard. A lot of them don't particularly like basketball. There have been - A dirty secret of basketball is that a lot of the players particularly the tall players don't particularly like basketball, they're just tall. They don't practice as hard. They're not as athletic. But the height is such an advantage that kind of all the normal rules of what it takes to reach the top of an athletic pursuit are off the table.
Yeah, I mean, because you had a comparison to volleyball players. There are countries where volleyball is popular, and that is, you know, obviously competitive in some ways to basketball. But with the net, the height matters volleyball, professional volleyball, but the most well remunerated volleyball player in the world makes $300,000 which look with inflation today. That ain't that much.
Yeah, well, I don't know if it's the highest pay, but he's the highest Leaper. Yeah, so volleyball is the only the only sport, other sport, that really uses height the same way basketball does. And the average volleyball player has pretty much the same body type as a small forward in basketball. So 6’8” 6’9” not incredibly bulky, extraordinary Leeper and you do see that in countries where volleyball is really popular, I didn't realize just how popular volleyball is. In certain parts of the world. You know, Iran, volleyball is five times more popular basketball is volleyball is more popular than basketball and Brazil and Italy and Russia, Bulgaria, Puerto Rico. And in these countries, there are fewer basketball NBA players than you'd expect and particularly fewer forwards than you'd expect. So Brazil is only produced two NBA forwards over the years, so much fewer than you'd expect. And I think a lot of the 6’8” -6’9” athletic men in Brazil are playing volleyball instead of basketball, which is a mistake from a financial perspective, because salaries in the NBA have just exploded that You know, even the worst players are making a million a year guaranteed and no others. You know, very few sports hang like this
Let me correct myself. I use chat GPT to to look it up. It looks like Zhu Ting is a woman. And she's making 1.6 million, the highest paid and the highest paid man is 1.4 million Willfredo Leon, I don't know how you want to say it. So the the other high ones are, you know, north of a million but below 1.4.
And that's like the minimum in the NBA. Where as the highest NBA is 50 million.
So you're talking about like two distributions were at the high end and the low end they meet and that's it, but they're almost disjoint Yeah. Okay. Yeah. So, so in terms of like competition for I mean, yeah, if you're seven feet tall, and you're doing something else, you would be insane. Like, you just, I mean, to some extent you just need to show up. Because, you know, yeah,
You know, just so - I mean you don't just you need to be one in seven ability, although that's just reaching the NBA so even if you don't reach the NBA you might be able to be a player in Europe or Australia, or one of the other leagues, they don't pay quite, they don't pay nearly as well, but they do offer you the chance to make a living playing a game. I know you're obviously, you know, genetics is your thing. And I'd be curious what you think about this, I kind of end the book talking about it, that the future may involve more genetic testing to guide kids towards where they have the highest genetic potential. So I talked about how Sean Bradley had, you know, more than had almost 200 genetic variants related to increase height, and you know, Shawn Bradley's 7’6”, and presumably, and he ends up way taller than you'd expect, just based on his parents height, presumably, Sean Bradley could have known that at a very young age, his parents could have known that he was going to be seven, six, and he could have practice shooting, you know, gotten great coaching on shooting and other fundamental basketball skills at a very young age. You know, I could imagine a future where people you know, now the way it's, it tends to work with these seven footers is particularly if they don't grow up in the United States is, you know, they're 12 13 years old, and they have a growth spurt, and people tell them start playing basketball. And, you know, so a lot of Hall of Fame centers didn't play basketball until they were teenagers. So Hakeem Olajuwon and Joelle Embiid, Dikembe Mutombo. And that probably does limit their skills compared to if they started practicing, when they were much younger. And, you know, I think with genetic testing, it's probably silly to wait until someone is 12, or 13. And they have their growth spurt for them to start practicing basketball.
Yeah, so we can just we can we can jump into that right now. Because it's an interesting topic. And you know, I'm assuming some of my listeners would want to know about it. So basically, most of you guys know that there are characteristics like intelligence that are polygenic height is obviously polygenic. Seth mentioned in the book 80% heritable, that's a pretty good estimate. So that means 80% of the variability in the population is due to variation in genes. Height is not, you know, I'm not gonna say it's polygenic. But I'm not gonna say it's not as polygenic as say, intelligence or bio behavioral traits. But the effect size of the genes per gene is like about 10x, larger than would say, intelligence. So it's technically a more tractable trait to imagine doing a predictor right now than say intelligence. So I think with intelligence, the best predictors, using just genes of intelligence in the population Hit like 18% of the variance, which means that like, the majority is still not accounted for, in purely genomic predictors, where it's like you go in, you identify position in the genome out of the 3 billion, and you use that as your independent variables to predict the variance right. With height, I think they're coming, like closer to like north of, with whole genome sequencing and stuff like that. So I just looked it up, as you were talking. So it said, right now, genomic methods can predict between 10 to 40% of the variance, okay. With, there are people with whole genome methods that are claiming that they're coming close, like not too far from 80, actually. So that means the heritability estimate using twin studies, which you mentioned in the book, I think, which is just usually inference based on correlations across families. You get those and the genomic methods, which is, you know, kind of like going straight from the genome and doing a predictor are closing. So is it feasible? Yes. I mean, it's almost certainly feasible within the next within, let's say, within the next five years, yes, like, it's feasible. I know genomic prediction has a paper out from like, five years ago, they have a very good individual predictor. So some of these statistics, as those of you in the audience, have any of you know, our population wise predictors, their efficacy is much lower on the individual level, but - but, you know, assuming that you have the parents genomes, you have their realized heights, you can like set up some sort of system where it's like, you can predict with very, very high accuracy at a young age. What I mean, reasonable precision, okay, that the odds of this person being x height, are going to be like, you know, 1000 times greater than any random normal person, right. And so, when you're talking about Sean Bradley, that makes sense because the further you are from the expected value of the population or even the mid parent value of the parents, the more likely it is you have a weird i don't want to use the word concentration but the weird configuration of representation of alleles, right? And those are gonna get picked up with these predictors. So even if the predictor for example, like people say, Oh, well, IQ predictors are only like, explaining, like, 15% of the variance, but like, bro, like, if someone is like, I don't know, just crazy, four standard deviations, that's gonna get picked up, and you're gonna have a sense, even a polygenic trait where you can explain most of the variants that, you know, the person is probably exceptional. I'll tell you, I think like, I think I've mentioned this before, I have a whole genomes of considerable number of people in my, in my family, and I have like, snip arrays on everybody. I've already done a polygenic index on like educational attainment, and I'm just gonna say the right quarter is exactly what you would intuitively expect. So these are these are much better than people right now think, just like looking at the raw numbers, you'd be like, Oh, no, but the issue is, a lot of times you just want to predict the rank order. You know, you just want to get a sense across the general population, not the exact value. So I think yes, like, for height, a lot of a lot of what you're talking about in this book is actually, so athleticism is a multi variable, characteristic. So one of the things that I want to mention is, if you watch NBA players, you can tell that some of them that are very tall, are clumsy, and some of them are not, like some of them, there are, there are guys who are six foot 10, they move, like they're six foot one. And those guys, they're like, you know, the, they're kind of at the big forward position. But you know, they're like Magic, they’re point guards, you know, they could do things like that. These are very, very exceptional people. And it's not just random, like they obviously, they probably have some genetic dispositions, they probably practice a lot. Although the practice a lot, the focus, like this has to do with like, I think clutch shooting and other things, all that like that's like in the head, and some of that is probably heritable, too. So, you know, that's a separate issue. But all of these things are multifactorial, but in basketball particular, I think it is kind of easy to look at it from a data analysis perspective, because there's huge variable that's like overwhelming, which is the height. Whereas in other sports, I don't think there's, I mean, you have to have a floor in speed with football and other things. But then there's like slow positions, like, I don't know, like, center, whatever. My point is, a basketball height is this huge variable is highly heritable. So I think it's overwhelming a lot of the signal even though there's other interesting things out there. It's such a clear signal that allows you frankly, your narrative is much easier, it’s a short, compact book,
I mean, I guess you'd compare, if you're really trying to predict someone's height, you wouldn't just use the genome, you'd also use their bones or whatever, right? Because they're also predictions based on your bone structure. They, when a certain age people, they have a some sense of how tall you're going to be. So you could kind of put everything into the, into the system. But yeah, it is interesting to think about that, you know, all these, I mean, it's, it's a, it's a, it's a rare issue that, you know, there aren't that many people grow up to be 7’6” but it is true that among people that grow up to be 7’6” their life would probably be a lot better if they started practicing, you know, if they got coaching in basketball by the age of, you know, five or six, and sometimes they do apparently, you know, Yao Ming, was picked by China, like basically does more of this where they do try to measure I think they use the bone structure, and they try to measure who's going to be their giants. And they put them in a school. And I think Yao Ming, who was the son of two basketball players, and they knew he was likely to be enormous, was trained in basketball from a very young age in China, and you will see is is different than a lot of seven footers. He's a he was an 84% free throw shooter, whereas the average seven footers about a 60% free throw shooter.
Yeah. Well, I mean, so Yao Ming I mean, from what I remember, and I was still kind of paying attention to basketball when he first came on the scene, and he's gone now. But Yao Ming, I mean, his parents, they were set up. Yeah. Like he was. He was, candidly he was bred like, they bred him, actually. So let's just jump to, you know, what, I'm gonna, like, change the order, hey, this happens, you know, let's just play by ear. Right? So you're talking about like people basketball is a very heritable sport. And the one thing that you found that I thought was interesting, again, like, what I'm going to say is a lot of the facts that you report on in this book, after the fact, they totally make sense. You wouldn't necessarily know you know, what's true ahead of the time, but I'm just like, okay, that makes sense. You know, so it's great. Free throws and the heritability of basketball like let's talk a little bit about that.
Yeah, so heritability of basketball is I actually, I looked at what percent of same sex siblings or identical twins in different sports. Uh, and I got a Olympic data from this guy, Bill Mellon who collects data on the Olympics and then data for the American sports are just readily available. And what you see is that the NBA is just dominated by identical twins in a way that other sports aren't. The one other sport that is that I didn't include in that chart. Also similarly dominated by identical twins that it goes to your point that athleticism can be very heritable is Track and Field. Track and field is also very dominated by identical twins. But basketball is much more dominated by identical twins than football or baseball, and numerous Olympic sports from swimming to gymnastics to equestrian riding. And you know, that's a dead giveaway, just how genetic basketball is. I actually do my own calculation. That basketball probably is around 80% genetic largely because if height is so important, it's about 80% genetic and a lot of other things, traits, you know, the body in general, it's not just height, it's arm length, and hands width which are also extremely genetic. Athleticism can be extremely genetic, a vertical leap seems to be pretty, highly genetic, maybe about 70 or 60 to 70%. So basketball is just, you know, extremely genetic as seen by all the identical twins. I do that because I'm a fan of Stanford basketball team. And the Stanford basketball team goes through these periods where they suck for a while, and then they get a - they recruit a pair of seven foot identical twins. So first, it was the Collins twins, Jason and Jaron Collins. That was the Lopez twins, Brooke and Robin Lopez. And they have two NBA seven footers, you're in pretty good shape. So they were great when they had both those pairs of twins, pairs of identical twins.
Yeah, it's fascinating. The heritability thing of the different sports is fascinating, because I think it gets to the point of, it's kind of being driven by the heritability of this, again, the height variable, but also you were saying like, things like wingspan and other things, basically. So when you look at the genes that are correlated with height it’s bone morphogenesis genes, okay, so those make sense. You know, it's the morphology of like how your bones develop. And this is a very, assuming you have good nutrition, which, you know, look at the size of average Americans, we got good nutrition, good, quote, unquote, but whatever. You know, it's gonna be driven mostly by genetic variation. And so that's just like the big thing in basketball, whereas baseball, I think, I don't want to say it's like requires more skill, but I mean, do you remember Fernando Valenzuela?
Seth: The heavyset guy?
Yeah. But he was like a great pitcher. But I mean, it's, I'm just saying in baseball, like some of these guys, you're just like, okay-
Seth: They look like plumbers or whateve?
I'm not gonna lie. It's like, because like, sometimes, like, I have, like, non American friends. When I was in college, like, we watch sports, and it's just like, okay, like, I know, you guys say soccer players are like, small and shrimpy. But like, that guy looks like he could barely like, move, you know, like some baseball players. I mean, they used to be I don't know, because I don't watch anymore. Like maybe they're all fit now
They’re still the, the Mets have a player, that's…
Some of these guys. It's like me as you could probably outrun them, you know what I'm trying to say? So it's like
Baseball, baseball also has a huge nurture component. So you could also, you know, I mean, one of the things that's interesting is baseball is actually more dominated by brothers than basketball is just generally brothers, the percent of players who are brothers, the odds that a brother reaches, you know, the advantage that a brother has, compared to an average person in baseball is higher than in basketball, but much less dominated by identical twins. And I think what's going on there is that the nurture effect is enormous. And, you know, that, you know, I talked about in basketball, a lot of Hall of Famers, you know, don't start playing the game till they're 13, 14, 15 years old, and you can still reach a Hall of Fame in baseball. That's not true. You know, you look at the greatest baseball players of all time, a lot of them their father played baseball, a lot of them their father, just trained them in baseball from a very young age, getting your swing exactly right. Getting the form
Of it right at a young age, practicing over and over, you know, as the ball gets faster and faster, the pitch gets faster and faster with proper form is very important. I don't think you can just, you know, start playing at 15 and become a great baseball player. So nurture, I think is extremely important in baseball in a way that’s not true in basketball. What is true in basketball, nurture is important for shooting I think so I talked about that in the book. Where if you look of sons of NBA players, they're extraordinary shooters, extraordinary. You know, particularly free throw shooters, you know, the best free throw shooter of all time. Steph Curry, his father was in the NBA. And I think what's going on there is if your dad was in the NBA, your dad learned kind of proper form at some point, and he teaches you proper form at a very young age, and you practice it over and over again, and you become a great shooter.
And this gets to another, I knew this, because well, this is anecdotally, I will say, anecdotally, if you just observe the NBA, so there is a stereotype. You know, only, you know, there's this whole issue, my friend Lee Johnson was talking about this, a lot of stereotypes are true, actually, you know, but for a while there was a stereotype of, you know, basketball players are like, you know, off the street, you know, literally off the street. They're from the inner city. And, you know, all this stuff of the you know, maybe Allen Iverson from the early 2000s. Again, I'm dating myself, partly because I just don't follow sports anymore. But was, you know, kind of like a more quote, urban, like, these are euphemisms, but basically, a lower middle to like underclass, you know, black Americans are kind of like the stereotype, you know, urban, you know, you know, ghetto, whatever, all these words, right. But that's not the reality. The reality is somebody like Grant Hill is, I think, is I think his father was a football player, not a basketball player. But, you know, but more like middle class Grant Hill, Kobe Bryant, whose father was a basketball player, they're more from well off backgrounds, often from middle class backgrounds. And, yeah, and that's just more statistically true than that they're from, you know, you know, from like, deprived inner city backgrounds. Can you talk a little bit about that part of the book?
Yeah you know, that, yeah, this stereotype, there's this idea that, you know, if you came from a rough background, you're more driven to reach the NBA. And, you know, you had no other options. So if you're poor from the ghetto, basketball is your only chance of having a decent life. Whereas if you're, you know, the son of doctors, or lawyers or professional athletes, or, you know, other people and middle class, upper middle class, wealthy people, you're not going to, you're gonna have other options, and you're not going to be as driven. You're not going to play basketball day or night to try to escape your situation. And this has never been true. I think Adams and Dubro are two academics who studied this, I studied it anyway, you look at the data coming from more privileged backgrounds, is a big advantage towards reaching the NBA. In some sense the stereotype never made sense. Because if you think about what are the advantages of an advantaged backgrounds, you know, higher socioeconomic status, one of the biggest advantages we've long known is non cognitive skills. So interpersonal skills, trust, discipline, you know, ability to to get along with others. This is why crime rates are much lower among people from higher socio economic status. And, you know, those are so important in reaching the NBA, your ability to get along with other people to, you know, to have discipline to, you know, and there, there are so many stories of NBA players from tougher backgrounds, or potential NBA players, great basketball talents from tougher backgrounds, who just started fighting with their coaches getting in trouble with the law, you know, never develop their potential and never reached the NBA. Whereas players from middle class backgrounds, Michael Jordan, you know, he talks about how much his parents helped him in avoiding the temptations of stardom and avoiding, you know, all the pitfalls that a young successful black male can fall for. And you know, Chris Paul, and other example, his family joined him on an episode of Family Feud. These backgrounds are super helpful for developing a career in team sports. So, you know, this stereotype never fully made sense. I don't know exactly where it came from. I guess the reason for the stereotype is just, it is true that African Americans in the United States are disproportionately from poor backgrounds, more likely to have single parents, teenage mothers. So because the NBA is dominant, you know, it there are so many African Americans in the NBA, it is true that even though there's a disadvantage to coming from a disadvantaged backgrounds, there are a lot of NBA players from disadvantaged backgrounds. You know, Allen Iverson, you mentioned LeBron James, grew up poor in Akron, Ohio to a single teenage mother. There are plenty of you know, Kawhi Leonard grew up in a really difficult background in Compton, LA, there are plenty of NBA players who did come from difficult backgrounds, but just you know, weighted from, you know, to the African American population at large, there are much many fewer of them.
Yeah, well, I mean, I think part of it is, you know, what, and who has salient as well. You know, if you're if you if you fit the part, you know, because a lot of Professional sports is the narrative and the story. You know, there's a lot of stuff. Look, there's a lot of stuff on the court that happens to pay attention to. This I think more evident in the Olympics where it's like, you know, a lot of people aren't following these sports, and you got to create a narrative around the players and, or not the player but you know, like the competitors. But the, you know, the same thing applies to professional sports, like, you create a narrative around people. That makes them you know, kind of like, interesting, you know, I think LeBron James is, I don't know, his background is a little bit. I'm not saying it's like, horrible or anything, but he's, I don't think he's from a little middle class background, my understanding?
No, no, no, his mom was, I think 16 single mother in poverty in Akron, Ohio.
Yeah. So I’m just going to say, it makes a great story, you know, from like, you know, nowhere to be the greatest basketball player on the face of the earth. And so I, you know, I think I think it's salient to people, I think that's where it’s coming from, but what you're talking about non cognitive skills, you know, also, like, I think the term non cognitive skills, I don't know, I don't really like it, because like, a lot of these are also cognitive, but like, let's set that aside. I know what you mean, a lot of what you're talking about is the exact same issues that people talk about people from lower socioeconomic status backgrounds, in terms of just a regular job, you know, like, and, look, you know, we've talked a lot about genetics being tall and all these things, but you still got to show up, you know, do not I mean, like Latrell Sprewell, which, again, people can like who are listening to this? Who are basketball players know that I definitely stopped watching in the mid 2000s. Because all my references, yeah, but
I liked them. Because I consider that a glory period of basketball.
But now Latrell Sprewell like that guy left 10s of millions of dollars on the table, because of his inability to control his rage at his coaches. That's basically what it was. Right? So like, can you just talk about what happened with Latrell Sprewell? It wasn't at the book,
Is he the one who choked P. J. Carlesimo?
Yes. You’re choking your coach, you're choking your coach. And so he was like, suspended for years. Like he couldn't come back. I think he did eventually come back. But Latrell Sprewell, was actually like a second round pick from what I remember. And he was one of those people that kind of flourished in the NBA, and really, really overperformed he was on the road to stardom. And then he just flipped out attacked his coach at look, they don't want you on the team, because like, you're also like a bad role model for the other players. You don't want to be assaulting the management.
I actually did a study, which I didn't include in the book. It's pretty clear that the lower your socio economic status, the more technical fouls, you commit, you know, so NBA players from lower socio economic status are more likely to commit technical fouls, commit more technical fouls per minute, which again, goes to it would be surprising if it were otherwise, you know, it's the same. If you look at incarceration rates, you know, the Raj Chetty has done research, you know, the relationship between socioeconomic status status and incarceration rates is just insane. You know, from, you know, five times higher if you're, if you grew up in the bottom percentile than if you grew up in the top percentile, socio economic status. So I think you see the same thing among, you know, kind of the crimes in the NBA of technical fouls, let alone what you're talking about, which is more extreme acts of assault.
Yeah, so let's talk about - loopback a little bit. Why there are so many Lithuanian basketball players, and Yugoslavia basketball players, you mentioned that you you know, India and China. I mean, there are there any Chinese NBA players right now?
Zero now, there have been obviously,
Yeah, so the all that whole, like, you know, predicted boom didn't happen. So like talk- , let's talk a little bit about that. So that's interesting.
Yeah. So you kind of predict what, you know, we I measured how much per birth how many NBA players are produced in kind of every country in the world. One of the things that’s interesting is the United States is not the highest. So the highest is US Virgin Islands. So that's where Tim Duncan, for example, is from. Montenegro, Bahamas, there are basically three regions of the world that produce NBA players at an order of magnitude higher than other regions are four regions, the United States, the Caribbean, the Baltic states, Latvia, Lithuania, Estonia, and then former Yugoslavia. And a lot of so a lot of the reason it is not - so one of the things that predicts how many NBA players an area a country produces, not surprisingly, because we talked about the importance this variable is height, so average height differs enormously in different parts. So the world so Netherlands is the country, I think right now with the highest average male height at six feet tall. You know, India, I think it's 5’6” 5’7” something like that. United States, it's about 5’9 5’10”. So this is going to have a huge impact on how many NBA players you produce. Particularly, you would know this, that the way the normal distribution works is even two inches, you know, a two inch increase in the mean is going to lead to something like 20 times more seven footers. I forget the exact number I had in my book. But that an increase in the mean, a small increase in the mean, leads to an enormous increase in the tails. So average heights really important. But another thing that's really important, also, not surprisingly, is the popularity of basketball. And the I was kind of why is basketball so popular in Latvia, Lithuania, former Yugoslavia. And it seems to be just random things from history, which is really fun and interesting. And kind of, you see the randomness of history. Lithuania in the Euro Cup after World War One. They were trying to build camaraderie among Lithuanians scattered around the globe. So they, they decided to have Lithuanians from around the world compete in the Euro Games competition. And this meant that a lot of Lithuanian Americans were competing at basketball. And because basketball was really only popular in the United States, that was a huge advantage. So they had they were led to victory in the European tournament, from a player who a Lithuanian American who had played on UCLA and was just way better than all the Europeans who didn't really play basketball. And this led to a huge surge of popularity in basketball in Lithuania, which still is there to this day, and Yugoslavia, the communist when they came to power after World War Two, they wanted people to play various team sports to learn the principles of communism, selflessness, you know, caring about the team at the expense of the individual. And this and basketball was one of the sports identified as basketball is a sport identified as one that would be useful to, you know, to teach these principles. So this random decision of the communist government after World War Two, is still playing out in the enormous popularity of basketball in the former Yugoslavian states.
Yeah, although I do want to say, my understanding is, you know, aside, you know, if you're looking at heights in Europe, the highlands of the Western Balkans are a secondary mode, aside from say, like the Netherlands and Scandinavia, right?
Yeah, they're also very tall. But you know, why are they why are there more NBA players from Latvia and Lithuania than from the Netherlands?
Yeah, so that's a - Because , like, they're not? I mean, I've always wondered, they're not like, super tall. I mean, they're tall, but they're not super tall, you know?
Yeah. I mean, sometimes it's also just, it varies within region. So apparently, there's a part of Croatia, where the people are really tall where most of the NBA players are coming from.
Like, that's why Montenegro is, so I think it's the highland areas of the Dinaric Alps, which like, there's, there's all sorts of hypotheses why that is. But anyway -
Yeah, and you know, South Sudan, produces more NBA players than other parts of Africa. And that's a tribe Manute Bol was a member of it, Luol Deng was a member of it a tribe or the average height, I think is about 5’11. So that, so you know, that if the average is 5’11, as mentioned, you're gonna have an unusual number of extremely tall people.
Well I mean, you see this, you know, we're talking like, you know, the nation state is not always a good for these, like athletic things, you know, because like nature, you know, human variation does not necessarily always follow the nation state line. There's internal heterogeneity. So, you know, it's like, let's, let's go to like, say, China, or India, there are parts of China, and India, where the average height is pretty close to the American, you know, some of them, I put up parts of northern China, probably a little taller, because America is now heterogeneous with shorter populations. But let's just compare it to like, you know, Europeans, right, Southern Europeans look to be genetically somewhat shorter than northern Europeans. It's not just nutrition, which kind of makes sense, like Italians eat fine, whatever. And then northern Chinese are definitely genetically taller than southern Chinese who are quite short, actually, most of the Chinese that you encounter outside of China are actually southern Chinese. So the perception that you have of Chinese people and height is actually under estimating what you would find in China. Like, you know, I have a friend and he's dating this Chinese woman who's like 5’10” and people are like, woah , you're 5’10” She's like Okay, whatever but her family's from basically Manchuria. And yeah, she's a little on the tall side, but it's not crazy tall, you know where she's from, but they're very few people from that part of China in the United States. Most people in the United States are from, you know, they're from Fujian and Guangdong, you know, more southeastern provinces in India, you know, the northern part of the country, especially the Northwest, you know, some parts of the South as well actually Kerala, for example, the people are not that short. Where my family's from in the east, I'm 5’10”, or 5’8”. We're from Bengal, I'm from Bangladesh, but that used to be part of Bengal anyway, basically, the eastern part of the subcontinent is filled with short people. And polygenic scores show that we are genetically shorter, like so I'm like a giant among Bengalis, you know, some of that is East Asia, Southeast Asian origins. So some of our genome is Southeast Asia, they're a little shorter. In some of the statistics that you report, that's also true. And, you know, people have run polygenic scores on Bengalis and it’s in 1000 genomes, and they're genetically shorter than the other South Asian samples. And that's almost, I think, mostly attributed to East Asian, Southeast Asian ancestry. Anyway, I'm gonna like, just like, the long story short is like I have read. There's a caste in India called Jats. There's, like, you know, Nikki Haley's a Jat there’s like, there's like 10 million, or like, 20 million, whatever, out of like, 1.4 billion people, like 90 some percent of the wrestlers in India that are competitive are Jat. Now, some of it is like, maybe the culture, there's certain stereotypes are not gonna get into, but also the people the Jats, they’re like, pretty tall. And they're also robustly built, you know?
So yeah, I think these things also become, you know, snowballs running down a hill, if you have a little advantage, or, you know, even or let alone a big advantage genetically, that it kind of builds into the culture. And it can become, you know, and then, you know, I think, probably I think, you know, one of the things I didn't get into the book is, why are African Americans so disproportionately represented among basketball players? And I think there's definitely a huge cultural element to it, you know, whatever the original reason that, you know, there were all these stars, Michael Jordan, Magic Johnson, Bill Russell. Now basketball, if you look at the data, I think 35% of African Americans are defined themselves as enormous basketball fans, I think among Caucasians, it's about 15%, or 10%, it's much lower. So you know, twice the popularity is going to lead to, you know, a lot, that's going to have a big effect on people reaching the top of the sport being the best shooters in the world, things like that. So, you know, so I think, definitely, yeah, advantages can kind of compound you know, probably, similarly, you know, Jewish success in various fields, I think, you know, whatever the initial advantage, once you kind of know, I think of this in my own story, you know, once you kind of know that you're, you're part of a group that has success in certain academic fields, you're going to be more drawn to that field. And-
Although basketball used to be a Jewish sport, there was a Jewish, Jewish Golden Age.
Boxing too had a very high Jewish, there are lots of interesting sports. So swimming, I don't know why that is, maybe it's has to do with the socio economics of it. But you know, a lot of the great swimmers. So I forget the guy who had the record before Michael Phelps, the famous swimmer was Jewish, and a couple, a couple of the people on - Spitz, Spitz I think, and a couple of the people on the if you look at the American team, a large surprisingly high percent of American swimmers are Jewish.
Interesting yeah. Now that you mentioned that it makes sense, or like, you know, I, yeah, it's not just - I'm not gonna get into it. But like now, I'm like, going through my head of like, 80s and 90s Olympics, and I'm starting to see the pattern, but you have to point it out for me to like, notice that I guess. So that makes sense. Okay, so I want to you know, we've been talking about the book. It's already out. It's on Kindle, hardcover, paperback. You wrote it like, you wrote it in 30 days, right?
Yeah. I'm toning that down now. So initially, I should probably mention- you're supposed to - I'm trying to get better at marketing my books are supposed to say the name a lot of times so it's “Who makes the NBA” and it can can be purchased on Amazon. I'm toning down the it was created 30 days. Initially, I'm like, this is kind of an interesting hook. You know, people will be intrigued that you wrote this thing in 30 days and a bunch of people told me that they had the opposite response. They're like they don't want to read something that it seems like a vanity project. Hey, look at what I could do in 30 days or like I challenge myself to write something in 30 days, you know, you want to see that - you want to read someone's best work.
That you crafted it. Crafted it over and over again.
Yeah, so I I'm toning down the 30 days thing. I think that's unfair, because I am extremely proud of this book, I feel like it may be my the best book I've written, even though it was written in 30 days. And I think that speaks to the power of AI that I think, three years ago, if I had said, I wrote a book in 30 days, people would be right to be skeptical and think, you know, that that's going to be a piece of crap. Like, you can't write a good book in 30 days. But I think because of AI, you know, you can write a treaties on basketball, you know, good, interesting treaties on basketball with a lot of new findings in 30 days, but I'm toning that down, I changed the cover, I'm like - A bunch of people are like, they don't they didn't want to read it, because it was written in 30 days. I'm like, alright, I'll just follow what I'm learning from the marketing. But yeah, it was, it was these AI tools are so insane. Like, I can't even tell you as a data analyst, how much they are game changers for my creative process. To be clear, I didn't write, use chat GPT to write, but I've always been a very fast writer, so that's never been the time consuming part of my process, I use AI for the data analysis, which takes forever, you know, merging data, cleaning data, you know,
Oh I know, I know all about that
It freaking takes forever. And, you know, even if you're using an RA or an intern, you know, they, you know, going back and forth. You know, it's it's, you know, it saves some time, but it's still time consuming. But with AI, it's just friggin instantaneous. It's insane.
Alright, so I'm gonna, I'm gonna actually do a podcast with my co founders at my company, we do you know, biological data, you know, mostly exact same, you're what you're talking about, you could do like a search and replace exact same issues that bioinformatics people talk about. And, you know, we do believe, like, part of our, you know, pitch deck, whatever your elevator pitch is, you have to have, there's no way that we're going to proceed into the future with like, you know, you know, it's not big data, it's like Mega data, you know, it's like, you know, petadata. You need to use artificial intelligence, to actually aid humans in manipulating and operating on these data, there's just like, no way that we're going to be able to do it with our, you know, late 20th century mentality, of just going forward with just our brains alone, using I mean, what we're trying to do is like Archimedes lever, right, like, we're human beings, but we have these abilities enabled by these computers. But now the data is so big, that you can't just like address it, you can't just it's like, you know, people don't write code in assembly anymore. You know, you have all of these higher level languages, you have multiple layers above the machine, you know, language, whatnot. So I think this is the same thing that's going on with data with data science, you know, for a long time. I mean, a lot of us, like, a little older than you, but a lot of us came up with Excel, and Access, you know, and now you know then, everyone moved to SQL, you know, relational databases. But now, the amount of data you have, you know, R can't scale, we all have these tools that everyone's used to where there's one layer between you and the data, I don't think that that's going to, like cut it now. You need to have the helper, you need to have, you know, another kind of like -
Well part of it is just for big datasets, you know, you need a helper. But I think part of it is just the speed, you know, it takes people a while to write code and, and to write code instantaneously just totally changed the process. I think one of the reasons, I found so many things that I found very interesting. And it sounds like you found very interesting, and I hope readers find very interesting, in such a short period of time was just you can test them so quickly, because the code, you know, just anytime you have an idea within an hour, you can test your hypothesis. And that's, like, so different than the way it used to be. And so you're just kind of firing through all these different hypotheses until you find something that's right and interesting and new. So it's kind of wild it was, it was definitely the most fun. I would say it was the best month of my life, probably, in part because the topic basketball is such a passion of mine, but also it felt like a) I was doing all the unpleasant parts of my of my work process. So I've never particularly liked coding. I've definitely never really liked debugging code. But I've always liked coming up with ideas, testing them. And I felt like I could do that. And then the AI would just do all the annoying stuff that I didn't want it to do. It was so fun.
Yeah, I mean, you know, stuff like GitHub co pilot is also really changing. Changing the game, I think, in terms of like how people are working on code. So? You know, yeah, so you did this really quickly, it helped with the data analysis. And, you know, I don't want to get into it, I think a lot of the listeners, I know the demographics of my listeners, a substantial minority are data scientists or in technical fields, Its just the whole process of testing and iterating, is, you know, I'm not gonna lie, I kind of like it actually, sometimes when I'm working on data, like, there's, there's a zen when you're in the mode of doing that, but it is time consuming. Just figuring out the questions to ask, and how much quality control you need, you know, and all this stuff, and, you know, is your result robust, and, you know, maybe you got to replicate it, like, do some subsampling, you know, there's all these things you got to do, okay. And you just got to do it. But you don't have to do them, you know, maybe a machine could do it, because they're not like, cognitively intensive idea. They're not cognitively intensive processes. Let's put it that way. Like, they could be like, you know, operationalized, and like, you know, artificially intelligent sized, like, let me just use that word, I just made it up. But you know, so I can see that. So you have a plan. Now to be the content -
The plan was the plan. Also, the plan I came up with yesterday in like, a possibly manic moments. So yeah, like, Yeah, I do -
You have a plan to be a content machine, like talk a little bit about that?
Well, I just think I could write a ton of books. And, you know, so I, yesterday, I was having a bit of it felt a little bit like a manic moment, but I'm like, you know, I can have a baseball book out by opening day, you know, the end of March, then I could have an Olympics book out by the summer. And I think, you know, people love to read the Olympics book in the, you know, when the when the Olympics are going on in Paris. And then I could have a football book by the start of the football season. And then I'm like, you know, I could just keep pumping these things out. And then I get really into, like, the business of trying to sell the books, because like, I, you know, one of the things that's interesting, I'm self publishing these books. And for the math on self publishing, you really want a series to make ads make sense. Because, like the ROI of an ad, you know, if you just sell one of your books, you know, if you're paying, let's say, 30 cents for a click, and you sell, you know, it's hard to kind of break even, you're getting five bucks for a book, you know, a high percent of people have to buy your book, to break even. But if you add that they buy that book, and then they have, let's say, a 50% chance of buying another book of yours, and a 10% chance of buying all your books, you know, so the baseball book, The basketball book, the football book, The Olympics book, then you can really start crushing it. And then you just like turning the dial up on the ads, you just start flooding social media with your ads for your book. So I've kind of started having fun with it. I, it's interesting, I usually, like my earlier books, I hated self promotion and marketing. I found it kind of skeezy. And it wasn't about that, for me, it was about the content. And you know, I feel like I let let the publisher do all the marketing. But now I'm kind of leaning into the self promotion and marketing, because it's kind of fun, you know, you know, this, you're you're running a business, it'd be really fun to think through the numbers and try to see if you could turn it into a successful business. And I think, you know, my instinct is if I could, yeah, if I could keep popping out these books that are, you know, I think really good in very little time and have a series of them. And then, you know, really understand the advertising market. I think the book business, it's interesting, I was thinking about this, you know, everyone says the book business is terrible. Like when I quit my job at Google to be a writer, my boss, who I loved, and love, was very skeptical of my plan. He's like, you know, you're not going to make any money writing books. And, you know, this, the, you know, and I was, you know, making money as a data scientist at Google. But I think, you know, one of the things interesting about the book business is that a lot of authors, like don't want to try are just terrible at making money because they think it's like, against the spirit of writing books and art. And James Patterson, the writer made a ton of money, because he realized that he could just market - that books were not that different from toothpaste, he had been a marketer, and he realized that he could market his books the same way he had learned to market Colgate, or Crest. And that's and you know, he was competing against other writers who are terrible marketers. So it's kind of interesting to think that maybe you know, it's not as bad a business as some people think. If you're like, if you treat it more like a business I think a lot of authors don't treat it like a business don't try to calculate their ROI and you know, and figure out the proper way to advertise and market and everything. So, I kind of got excited by you know, maybe this could be a fun challenge to you know, build a business around writing books that I think you can write a Great book now in 30 days, which, you know, and this is my first attempt at it, so presumably I'll get better over time as I kind of learn even more tricks. Some I don't know, I'm like this could be a really fun way to - a fun business or something.
You should, you should pitch yourself as the Brandon Sanderson of data science books.
I don't know who Brandon Sanderson is.
Google it. A lot of my listeners will know who Brandon Sanderson is. Brandon Sanderson is - So he is a fantasy writer. He's the guy. Do you know what the Wheel of Time is?
Seth: Nah
So Wheel of Time is basically a, I'm gonna call it a Tolkien knockoff. It's not really totally fair. But it's, uh, it was written in the 90s. It was kind of written like, what if Tolkien like lived like 50 years later. So it's a you know, it's more appropriate for, you know, audiences in the late 1990s, early 20th century, the author, I think, developed a heart condition. And he died in 2007, before his series was finished. So I think he was at book seven. And so Brandon Sanderson is a fantasy writer, a young fantasy writer, Mormon guy, you know, not that being Mormon,… but he's from Brigham Young is one of like that crew. And he finished the series, I think, the two last two books. And then he writes a lot he's working on like a decology, he has a bunch of other books. At during the pandemic, he wrote, like three or four extra books. And so he's publishing them under his own imprint or whatever. And he's just making a massive amount of money because he's basically a content machine. And he's actually recently written about AI, and how that's going to transform like for example, fantasy illustration. I don't - science fiction and fantasy. illustrators are just not, I mean, you got to be really good or have a brand, because AI is so good at making those illustrations. But also, some of the writing is pretty formulaic. So he was just straight up that, you know, AI is gonna be an issue, although he believes having an author or the connection to the author still matters. But the reason that I just bring up that analogy is, you know, he has like, figured out how to market his ability to generate content, at a very fast - he's like, the anti George RR Martin, who cannot finish a series. You know, like, George RR Martin is just like, I don't - you know, there's just so many memes about it, you know, like, he's too busy, like at the pizza eating contests in New Jersey and all these things. So, you know, but he's George RR Martin still has a lot of money, he still makes a lot of money on royalties. But you know, when you're talking about series. My understanding is genre fiction. That is really yeah, that is really where - Because what happens is, you have like a, an audience. Maybe they're like in their teens. But like, there's, they could still be following you. There's your series, like, when they're in their 40s, or 50s. You know, like me, like some of this. I don't read much fiction anymore. But like, if I had time, you know, there's things that I started reading probably in my teens, the authors are still kind of going, you know, and so my understanding is, it's cumulative. Because yeah, you do lose a certain number every X number of years, but you're also gaining new people. So if you have the base to work with, and you're producing the series of people that are, you know, still hooked on this, you end up having this like, massive, massive number of people that will automatically buy what you produce, because they know what you're going to produce. And they've already read other stuff. And, you know, you can say it’s a sunk cost fallacy you can make, you know, or spin in a positive way. But people are going to consume what they already know that what they don't know. So if you wrote a book, so I think your idea is like, Okay, you have you already figured out, like you prototype the system, you produce the book on the NBA, and that you're gonna do like theme, right, like thematically, kind of like shift, but it's not going to be like a massive deviation. So people who get the next one, they'll be like, Okay, I know what I'm gonna get, you know?
Yeah. And also, yeah, I think that's right. I don't know if it's gonna work. But I'm like, I think this one so far feels to me pretty successful in that, you know, the comments on the book are, are very positive. And people, you know, and, you know, I was able to produce it in very little time. And I'm kind of learning a little bit. I'm working with a guy who's helping me on the marketing element, and we're kind of working social media ads and trying to get the ROI on them positive. So I'm kind of like figuring out the system. But I think yeah, it's, I mean, it's a new idea. It's, it's not I didn't like, like, when I started this, it was really just, I just want to learn these tools and I was having to, I'm like, this is the best gonna be the best month of my life. Even if I didn't make a penny on it. It would still be a useful use of my like a great use of my time. But now I'm kind of thinking you know, maybe there is a business here so I don't know it's interesting to think about it would be kind of unique, you know, Most, it'd be a little bit different. I don't know if they you know, cuz I think the series we really think fiction, you know, you can write, I think it's well known that you can write, you know, tons of fantasy novels or tons of sci fi novels. It's not really as as usually we think it's a nonfiction. You know, even though the ones who do have anything resembling a series like the Freakonomics team, here, they built, they put a book out once every five years. But I think because of AI, you could actually put these out way more frequently.
Well, man, you’re living the dream sounds like you're living the dream, I can hear the excitement in your voice. I'm excited to see what you come up with. You know, and, you know, I'm excited about the idea of the series, I hope a lot of my listeners are too, I hope you get, you know, some I hope this is a good marketing exercise for you, as well. You know, it's, I'm interested as a content creator, I'm gonna use that word, pretend that I'm a Zoomer, but myself about what the possibilities, you know, I use chat GPT actually, I mean, it's replaced Google a lot, obviously, hallucinations and other things, you got to be careful. But if it's a field, you know, and you kind of alluded to this, I mean, you're not using chatGPT chatGPT you’re using the code creator, but, you know, hallucinations, I think are a little overdone. Especially if it's in kind of a field, if you're like, querying, a landscape that you're already broadly familiar with, hallucinations are very, very obvious. The problem is, you know, if you're writing about something you don't know about, which you probably shouldn't, you know, you're not going to pick up the hallucination, because you don't know anything about this. So I think this is why it's so easy to pick up some of these plagiarisms of undergrads, because, you know, they're taking a class in something they don't care about. And, you know, they don't know that it's a hallucination. Obviously, the professor would, because this is their field, you know, and so if it's like something you know, about, it can be very, very useful, because it just shortens the time to just like create - also, I don't know, you know, I don't know, like, for sure, but like, you know, if this is gonna be fixed, but like, a lot of times the ChatGPTs output they’re relatively, like unstructured, in terms of the way they're telling you the facts. So that gets given you the fact that you as a human being, still have to rearrange and configure it as a content creator, for other human beings. You know, I mean, if people will pay for just lists then we’re all out of business, you know, but I don't think that's what it is, you know, what I'm saying? Like, you created chapters, like your -
I mean, I think we're eventually going to be at we’re all out of business, I think it's eventually, you know,but I don't know how long it's going to be. So eventually, like, right now we're in a period where, yeah, we have a lot to add, and the AI is a tool. But I think eventually, that I don't think the AI is like, going to be incapable of being a better writer, stop hallucinating, or, you know, writing things more in a more structured way. So I think, yeah, I do think we're all out of business eventually. But hopefully, there'll be a decent period. Between now and then.
You know, I hear this so much, you know, it's just like, This is a weird time. Just reflecting, it’s a weird time to be alive, man. Just because AI has been, you know, we're of a certain age AI has been, you know, it's gonna be like, in five years, you know, and it's been, it's gonna be in five years, from my whole life until like, the last couple of years. And now it's like, oh, okay, it’s kind of here. You know,
It's so wild. Yeah, it's like,
It's, you know, it's a paradigm shift.
A lot of people say, like, as hyped as it is, it still feels almost under hyped. And I kind of agree with that. I just, it's so it's just,
Well you know, I don't know when it's gonna show up at the productivity numbers, because so many people over the last year, because, you know, this, like, you know, people tech and stuff like that, you know, these, these AI tools have really changed programming, and they have changed the productivity of a lot of people in technical fields. Because, you know, the people can use them, they know how to use them. And the fields are amenable to you know, a lot of this automation.
Yeah, I think yeah, it's wild. It's like it's already changing things. Like, you know, it's already it already works it already. Yeah, it totally transforms coding.
The future is here. The future is here.
Yeah and that's what this book is all about. It's like, it's like, you know, you can write a good book in 30 days now, which again seems completely insane. It seemed completely insane if you know, and it's not just that it's a good book. Like in the in a field of data science were you are doing original research, which is a field that notoriously would take a long time to write a book would take years and now it takes you know days. It's so insane.
Yeah, I will say you're just talking to civilians, you know, normies my life you know, being Tech Tech adjacent. The last the basically like, you know, GPT, and all this stuff has been around for a while, but okay, I don't want to get into detail because like, I'm gonna do another podcast with the artificial intelligence guy. But, you know, I think part of this issue is User Experience matters. The opening it up to the public, with the chat interface transformed, open AI, at change the game. So that shows that it's not just technology. It's also, you know, just how the technology interfaces with humans. Now that it interfaces with humans, it became real and accelerated, you know, open AI has transformed itself. Okay, so now we're at this stage, and everybody in the field, who could use AI is now talking about AI. Conversations come up constantly about these tools about how they're going to transform things and what you're doing and how exactly, you're using it. And then you talk to regular people in regular jobs and for them It's like, oh, it's like a headline here or there. They don't actually know what is happening, because they don't see it directly yet. But like, underneath the hood of our economy, there is some sort of revolution that is starting in my opinion. You know, I like their stuff like in CNC, like manufacturing, there's a lot of places where it's starting to, like, just explode is my understanding, you know,
Yeah, that's it. Yeah, I think we're gonna start seeing like, Yeah, we're gonna start seeing weird things play out in the economy, like their, I bet you it just, you know, the I now beside the book business, I bet you you're gonna see like, some, like coloring book franchise go viral, where they're really good at marketing. And they're just like, there - or a recipe book, and they're really good at the AI stuff. They know the AI stuff well enough, and they're really good at marketing. And you're gonna see all these, like new authors, just like blowing up out of nowhere, using AI, that would be my guess. Like, you know, that's just -
Well Sports Illustrated, the Sports Illustrated, got in trouble for using AI journalists, right?
Seth: Oh, I didn't see that.
Yeah, Google it, there's a whole thing were Sports Illustrated was trying to use AI generated articles by non real journalists. Obviously, this is just a way, you know, I, you know, I'm a entrepreneur, I am now part of, you know, capital I guess, technically, you know, you think about like, you know, we have a fixed pot of money, and you want to hire the fewest number of people take care of the like, you know, my philosophy, I think philosophy a lot of people is, you want to pay the people that you hire as much as possible to get the most productivity, alignment and investment out of them. But you need to keep your headcount as small as you possibly can. Because you know, you have a fixed amount of money and you want to grow that money, you want to be cashflow positive, all these other things. So I get why Sports Illustrated did what it did. But it's like, you can't, you can't just like do this on the sly, like, you have to be transparent about it. Because you know, people, they think that, you know, they think they're consuming like human created product and they're not. I mean, this is gonna be more and more of an issue, as we go forward, you know, where it's like AI and artificial generated stuff is gonna be able to pass off, as you know, organic. This sounds really weird that I'm saying it this way. But this is the only way I can explain it. We don't, we don't have the words to describe, like what I'm talking about really I think.
I don't know that, that's necessarily like I felt in writing this like, even though it was AI assisted. But it still felt like an expression of my own human creativity. I think you could say the same thing with a Sports Illustrated article that was written with the help of AI, you know, that some human being had to come up with the idea they had to go over the AI, as you said, the main exercise, they have to prompt it, they have to run it a bunch of times to find, you know, the version that works best. So I don't know that, that I don't, that doesn't bother me, that doesn't feel like we're not, you know, getting an expression of human creativity. You know, like, I think one of the things we didn't talk about in this in “Who Makes the NBA” is that I have a lot of art in the book. And I have no artistic talent. I can't draw or anything I've never had art and anything I've written before. And that felt like I could use mid journey at DALL E to express creative ideas I have that. Otherwise, I wouldn't be able to express you know, the section on genetics. I'm like, it'd be cool to have a piece of DNA dribbling a basketball, you know, which is a creative idea, I guess. But, you know, previously that creative idea would just never go anywhere, because I had no way to express it. But now thanks to you know, this was, I used DALL E I can express that idea. So I don't know that, you know, yes, the AI drew it, but it was still my idea. And I had to determine, you know, which was the best of all the options I was given which really was expressing what I was trying to express. So I feel like AI can enhance the human creative - human creativity.
Well, I mean, I think, you know, if you don't have anything else you want to end on, I think that's, that's a good, optimistic, positive way to end. You know, mostly we talked about your book, but I mean, I'm glad we talked about AI, because that's big thing as 2024 I mean, we're recording this earlier, 2024, although, you know, I'm going to be posting this soon. So not gonna be much latency. But, you know, I will be talking more about AI, with people using AI. Scientists, and also practitioners. And I think it's fascinating, because something like the NBA, sports, this seems to be a very all American, you know, just normal topic to talk about sports statistics, you know, ESPN heads, like people like that. But you know, you did it with AI, we talked about AI. And it just shows how kind of like, the present the past or the future could just kind of come together to this mishmash. And this is America. This is the world in 2024. And it's nice to be in the future. Let's hope we have a few more years. I don't know. That's the dark aspect, where it's like, you know, there's a lot of uncertainty there. But you know, what, there's a lot of uncertainty in the past. And, you know, we live in interesting times. And, you know, I hope I hope we all make the best of it. And yeah, the book is, actually let me get sorry, I What was it? Give me the title again? Give me the title.
Seth: “Who Makes the NBA”
Okay, yeah, “Who Makes the NBA” it's on Amazon and stuff like that. There's, like you have it, you have it in print for the the boomers you know, and Kindle. That's, that's how I actually, well, you gave me a PDF, but I also got it on the Kindle version. But anyway. Yeah. So so you could get it, like, whatever version that you want to it's out there. And so I guess we'll be we'll be seeing more in the near future in the summer, hopefully, you know, a boomlet of your books out there. So I'm excited about
In spring.
Okay. See, I was like, Oh, this is like acceleration like accelerationism in book writing, you should you should create a movement you should be you should be the founder of it, you know? Yeah, literary accelerationism. Like lit/acc lit forward slash acc. Like you are the like, that would be your like your brand. And I'm just trying to think of like, how you want to market yourself, like, lit/acc I just like gave you something.
Yeah. Yeah. Again, I think the thing I'm struggling with is that a lot of people don't want to read the books unless you suffered for them. So that's why I'm struggling. Like they don't want to. They're skeptical of
Razib: Oh, sure. Sure. Sure. Sure.
So I'm trying to figure out that in the market.
Okay so but this is a yeah, this is a whole thing. And like, you know, cognitive psychologists like Paul Bloom have talked about it in terms of like, when you consume something. It's not just the output that humans want to consume. They want to consume the story behind it. Okay. So what you need to do is, like, I'm just giving you whatever, because I would never want to talk about my process, you know, says like, it's kind of weird, but you need to, like, make your process interesting and real to people. And cool, which I think it's an easy - I’m an easy audience, because I'm a nerd, you know, maybe that is your audience. I mean, that's not a trivial audience, you know, but just, you know, your story of using AI to generate narrative, right, like, as long as you can, like, frame it in a particular way, you know, because look, like, I'm sure, like 200 years, 150 years ago, or whatever, the typewriter became a thing. Some people were like, oh, that's kind of like artificial. You didn't like, you know what I'm saying? Yeah, I was like, You didn’t write it in longhand. But now it's like, now typewriter is like being a hipster. You know, or George RR Martin uses WordStar. From the 80s. So my point is, I think you could make it you could spin it in a way that makes you seem like artists. Like they want to, they want it to be artisanal. Right. And so like, if you like, so basically, somehow figured out that people think your whole wheat instead of white bread. You know. Because like, I think about this all the time, like, you know, TV dinners, white bread in the 1950s and 60s, those were futuristic, those were cutting edge. And people were proud that they had like these refined breads, or they could just like, you know, pop, pop the TV dinner into the oven and would come out today that's kind of like low class. Now it's got a totally different connotation. So you just you just got to figure out like, what the brand is, you know, do the product market fit all that sort of stuff. Because you are now an entrepreneur, like you know, your freelance, whatever, I probably should be recording this. I'm just like giving you advice, but I'm just trying to say because like, I do this myself, I have to do this myself. I'm not - I don't work for anyone else. I just worked for myself. So you’re your own brand, you gotta like, position yourself and figure out like, what you're selling who you're selling to, and then how to optimize it. But, you know, it's not necessarily Goodharts rule where it's like, you know, if you have like a test that measures something, the test starts to become useless. If you over optimize yourself, you seem kind of, like artificial, people don't want to buy, you know, so you need to maintain a level of authenticity, that's still efficient in terms of delivering the product, so you maximize your utility, right? And so it's like, it's like one of these things where you have to, like mix authenticity with pragmatism and practicality. This is advice for everybody out there, who's a content creator. Also, like, even if you're an employee, you know, I tell people, you know, all this crap about, we're family, that's BS, right? On the other hand, you know, you have to be careful of how you behave. Because this is like an iterated game, you're gonna have different jobs, you know, so, okay, this is kind of, like off topic, I'm gonna close soon, but I'm trying to say is like, okay, like, let's say you go in as an employee, because that's a more understandable thing for most people. You go into it, you work for a company, it's a corporation, you know, it's a corporation, you know, you're a cog, you know, that they would fire you like, in a literal instant, even if they say you're family, right? But what's your what's your rapport with your manager, you still have to be a human, even though you understand that it's kind of this artificial situation where they might have to fire you, you know, not for any cause. But just because you're not efficient, you're not useful anymore, because you're just an output to them. And what are they to you? Well, they're just way you could get a salary, right? So you could go in there totally bloodless, and just try to be super efficient, and be like, okay, like, I got what I want out of it I’m quitting with, like, the minimum 20, you know, two weeks notice and whatever like that. But you know, your managers a human being, they might not want to give you a good recommendation for your next job then. Right. And so, you know, and like the manager themselves, they paid, they put up a kind of a fake front, that you’re family, you know, they take you out to lunch on Friday, and all that stuff. So my point is, you're still a human being, you still have to maintain your humanity within a system that's somewhat rationalized. So as you go forward, pivoting it back, but we're talking about using artificial intelligence, it's obviously a rationalized system, we're making it more efficient, where it keeps the utility, and yet we're still human beings. And, you know, we've used tools for, you know, basically the whole history of our species, perhaps? Perhaps we are a tool using species, we've used tools for the whole history of our species. And so you still have to maintain like, some semblance of your humanity underneath that to brand and market yourself to other human beings. You know, so that's my, that's my final thought. I know like some people are gonna complain that Razib was editorializing at the end. But you know what, like, if you're listening to this long, but that's not you. Alright, thanks man.
Seth: Alright, thanks Razib.
Whole genome sequencing is used for adults and children every day to assess risks for 1000s of diseases. Orchid, a genetics company led by scientists from Stanford is able to do this for IVF embryos. Now instead of waiting for a diagnosis parents can assess if their embryos have genetic variants known to cause severe conditions before their child's even born. No other tests can detect these issues so thoroughly or so early. So check them out at orchidhealth.com.
