Sean Carroll's Mindscape: Science, Society, Philosophy, Culture, Arts, and Ideas - 72 | César Hidalgo on Information in Societies, Economies, and the Universe

Episode Date: November 11, 2019

Maxwell's Demon is a famous thought experiment in which a mischievous imp uses knowledge of the velocities of gas molecules in a box to decrease the entropy of the gas, which could then be used to do useful work such as pushing a piston. This is a classic example of converting information (what the gas molecules are doing) into work. But of course that kind of phenomenon is much more widespread -- it happens any time a company or organization hires someone in order to take advantage of their know-how. César Hidalgo has become an expert in this relationship between information and work, both at the level of physics and how it bubbles up into economies and societies. Looking at the world through the lens of information brings new insights into how we learn things, how economies are structured, and how novel uses of data will transform how we live. Support Mindscape on Patreon. César Hidalgo received his Ph.D. in physics from the University of Notre Dame. He currently holds an ANITI Chair at the University of Toulouse, an Honorary Professorship at the University of Manchester, and a Visiting Professorship at Harvard's School of Engineering and Applied Sciences. From 2010 to 2019, he led MIT's Collective Learning group. He is the author of Why Information Grows and co-author of The Atlas of Economic Complexity. He is a co-founder of Datawheel, a data visualization company whose products include the Observatory of Economic Complexity. Web site MIT web page Google Scholar page Wikipedia Talk on replacing politicians In My Shoes (documentary film) Datawheel Amazon author page Twitter

Agents who are Realtors do more than open doors. They analyze market trends, interest rates, comps. They can tell you about flood zones, mixed use zones, and decode acronyms like HOA, APR, MLS. They connect you to lawyers, contractors, even Phil, the Seward scope guy. They negotiate, coordinate, advocate for you, close the deal with you, and hand the keys to you. They bring you home. Realtor's are members of the National Association of Realtors, right by you.

PNC Bank brings you Call of the Wild Money Moves. You hear that? YOLO! That's an internet troll telling you to put all your money into a single investment. YOLO! YOLO!

He wants you to liquidate your emergency fund. YOLO! And buy a digital racehorse named Silicon Steve. YOLO! Stay vigilant. He's very persistent. YOLO! Guard against Wild Money Moves with PNC Bank. Briantly boring since 1865.

Hello everyone and welcome to the Mindscape podcast. I'm your host, Sean Carroll. And by now in this society we live in, everyone understands that information is something very important, whether it's the kind of information we get over the internet in terms of news and what's happening, what the weather's going to be like the next day, but also information in a more technical sense. If you have a company needs information about what sales are going on, where the customers are, what products are available, and what you should be selling. If you're a scientist, information is the data that you have about the universe. Information is another way of thinking about what we know about the world. So it's an extremely general concept. But there's so much information around us right now that it becomes a subject in its own right to understand what information is and how best to harness it. And there's really no better person to talk to than today's guest, Cesar Hidalgo. Cesar was trained as a physicist, but he quickly got

into the idea of statistical mechanics of information, which led him, believe it or not, into economics. And he started studying not just economics in its own right, but how data flows through economic channels and how it becomes actual physical products. So now, after spending a long time as the lead of MIT's collective learning group, Cesar is newly a chair at the University of Toulouse in France, but he's also the head of a startup company called data wheel. He's been very involved in data visualization and how that can help us understand what's happening in different places around the world. I can really recommend his book called Why Information Grows, that starts much like this conversation you're about to listen to

does from the basics of what we mean by information at the level of physics or even philosophy into how information moves around, whether it's through a biological organism, through a society, through a culture, through an economy. It's a different lens. It's a different lens. It's a different It's a different way of thinking about what's going on all around us. And Cesar is a very charismatic proponent of thinking about information in interesting ways. So I think you're going to like this. Let's go. Ask yourself, what are your best people spending their time on right now?

Expense reports, receipt chasing, month in close that takes weeks. You become what you spend on. And that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts. AI agents that handle the manual stuff automatically. So your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF.

All right, quick quiz for the hiring managers out there. What's worse? Being understaffed or being poorly staffed? Well, that's a trick question, because both are recipes for chaos. Either way, just say to yourself, this is a job for indeed sponsored jobs. You'll get matched with candidates that meet the skills, certifications and everything else you're looking for.

Or go a different way and get no traction. Seriously, sponsored jobs posted directly on Indeed are 95% more likely to report a higher than non-sponsored jobs. It really is a no-brainer. Spend less time searching and more time actually interviewing candidates who check all your boxes. Less stress, less time, more results. When you need the right person to cut through the chaos, this is a job for Indeed sponsored jobs.

And listeners of this show will get a $75-sponsored job. credit to help your job get the premium status it deserves at Indeed.com slash podcast. Just go to Indeed.com slash podcast right now. Indeed.com slash podcast. Terms and conditions apply. Need to hire? This is a job for Indeed's sponsored jobs. These are Herdago. Welcome to the Mindscape Podcast. Thank you. So you do, you started as a physicist. Is that right? That's right. And you moved through at any point we were officially an economist? No, I never got a degree in economics, but I've been working on topics related to economics, I would say now for 14, 15 years.

Okay, so you know the lingo. You're pretty... Exactly, and I have a lot of colleagues and enemies in... Colleagues and enemies, good way to put it. But now you're also very interested in data visualization and things like that. So to people who are not an expert in any of these areas, what is the 30,000-foot view of all this? Like, how do you think of your project of putting these things together? So the way that I define my work is I say that I focus on collective learning.

So what I try to do is to understand how teams, cities, nations, and countries learn. How do they acquire new knowledge and how they put that new knowledge to use? And to do that, I do a lot of things. I study the creation, diffusion, and valuation of knowledge. And I've contributed a lot to that literature and economic geography and innovation. But I also have created lots of platforms to integrate and distribute large volumes of public and private data as a way to improve the way that we see our world. Okay.

And does the physics background help you here? I think so because at the end of the day, what you rescue from an education like that of physics is that to understand the world, you need to always have an interplay between theories and experience, you know? So that duality is useful in physics, but it's useful in economics

and it's useful in most other fields. So I do think that my work still is always in the boundary between what the data is telling us and how we interpret it. Okay. But even more than that, you have, so I read your wonderful book, why information grows?

Yes. And the word information is obviously playing a large role here, right? And information has different definitions in different contexts. It's closely related to things like entropy that physicists care about. So how do you think about information?

What is your idea as soon as you say that word? Yeah. And of course it depends with whom I'm talking to, but in a more technical sense, I like to think of information as like this sort of like third thing that is very basic and important to understand. So in the universe we have sort of things, you know, we can matter, things that we can, you know, we can touch or that have some sort of embodiment, you know, but we also have the movement of those things. We can think of in terms of energy and momentum and so forth. But there's a third, you know, quantity that we need to consider, which is not things nor how they're moving, but how they're arranged or ordered.

And to me, that's the basic idea of information. You know, it's like, you know, the sequence of things, the way in which you stack a deck of cards. If you shuffle a deck of cards, you don't change the mass, you don't change the energy. But you're changing something. That order is information. I guess probably people, when you say the word information, they think that that information is about something, right? That it contains meaning, not just data. So when you shuffle a deck of cards, the meaning doesn't really change.

Or maybe it does. I'm not quite sure how you think about it. Yeah, so maybe a better analogy than a deck of cards is to think of DNA. So if I change the sequence of, you know, nucleic acids on DNA, I can transform, you know, one piece of DNA from encoding one protein to encoding a different protein. Like that protein and what it does and, you know, in the context in which it's being used, you can think of that as the meaning. But the little piece of DNA that is a certain sequence doesn't know really about that meaning.

That meaning is beyond it. it's part of the environment and the way that that sequence of order interacts with the rest of the environment. So I try to think of information when I think about it in fundamental terms as those sequences, but of course when I'm talking with someone, for example, from the field of communication or media studies, understand that information there is much more related to meaning and you can have concepts like misinformation, which in the DNA example, you know, would be a little bit, you know, harder to build. I guess that's what I'm trying to get at. So for the DNA molecule, or just for a set of letters on a page, is any arrangement equally contain the same amount

of information, or do they contain more information if the context they're in cares about what they say in some way? Well, you know, it depends on which definition we're using again. So if we're thinking kind of from a pure Shannon perspective, you know, basically a random sequence is going to be the one that contains more information because it's the hardest to predict. Right. You know. Even though it means nothing.

Exactly. But let's say now we're in the context of communication, you know, and I'm trying to communicate something to you. Well, the words that are going to contain more information are the ones that reduce your uncertainty more about what I'm trying to say. So maybe a more useful way to think about information in that context is not simply how many bits do I need to encode something, but how much do I reduce your uncertainty with each bit that I provide to you? Okay, good. This is sounding nice in physics-y, and I like this. We're going to get to the role of information in economies and firms and networks and things like that. But let's stick with the physics angle here. I mean, how do you think about the origin of information, you know, all the way back to sort of the evolution of the universe or the evolution of life or something like that? That's a good question because in some way, information and order and complexities conspicuous in our planet. So we're marveled at it every time we go and see a landscape or, you know, like walk around the city. But at the same time, you know, if we were to take a spaceship and travel, you know, across our solar system or beyond, we would see a universe that is quite barren, you know. Complexity is not everywhere. Complexity actually concentrates in places like our planet. And it leads us to ask why. Why is our planet so rich in complexity where other places, you know, are so barren? Like the moon is not...

Exactly. The moon is not complex. Or even like if you go to that Takama Desert here in our planet, you're going to find places that actually don't have too much structure beyond the geological formations that you can find there. And some telescopes these days, by the way. Yep. Yeah, indeed. So to understand the origin of this complexity and disorder, I think the best solutions that I found there is work like that of Ilya pre-gosing.

So Ilya, you know, as you know very well, is a very famous, you know, statistical physicists and chemist, you know, that he started to study physical systems that were out of equilibrium, you know, near equilibrium, but out of it. And he found that those systems that were out of equilibrium tended to self-organize into what he called dissipative structures, you know. So think of the little whirlpool that forms when you take the plunger out of a bathtub, you know, that's a structure that is not haphazard. You know, there are correlations there, there are certain order, their structure, you know, and that structure is something that emerges, you know, when that system is going through a state in which

it's flowing, you know, in the system in which is like going from one energy state to another. So he says when the systems are out of equilibrium, when they're moving from one state to another, they're kind of organized. And that is an important clue because if we think about it, a lot of the structure that we observe in nature, you know, it's in life. And life is an out-of-equilibrium system that has to be sustained by energy flows. We have to eat many times a day. We're breathing and we do get energy also out of the oxygen that we breathe. So there's a lot of energy consumption that we need to stay out of equilibrium because

the only way to maintain our level of organization is to stay out of that equilibrium. So I think pre-gogion is the one that gives us a first good clue of why they're structure in the universe. Yeah. So in other words, from the perspective of someone like me, if you just let the system go all by itself, it would go to a maximum entropy state, right? It would be boring equilibrium. It would, you know, the bathtub would just become flat, right? The water on top of it. But you're saying following pregogene that in the right circumstances, if you feed it some energy,

in fact, some energy in a low entropy form, then it obtains this orderly configuration, I guess. Exactly. And that's conspicuous not only in biology. Now, our lives live around electronics that, you know, they become useless the moment that the battery runs out, you know. So they're able to process information, they're able to, you know, show information on their screens, you know, encoded as pixels or whatever you're using as a display type of technology because they're consuming energy. So that energy consumption, you know, is essential to kind of like keep order because otherwise, you know, entropy does what it does. So what should we say that differentiates really the Earth from the Moon in this case? I mean,

we're at the same distance from the sun, right? Yeah, but I think here there's a lot of things that have happened that allow us to preserve information effectively, you know, and there's also like a chemical complexity that maybe might be missing in places like the moon. So on the one hand, we do have an atmosphere, you know, that makes a big difference, you know, and we do have, you know, oceans, you know,

which, you know, also make a big difference. And those structures together with others have been able to create path dependencies, you know, for replicators like, you know, DNA and RNA, you know, to create order, sustain it and reproduce it. So in some way, the complexity that we've observed today is not the result of like an instantaneous event or a condition that is present today on Earth and not present in the moon, but also of a long path-dependent process in which this complexity has grown over time because we're

able to generate more of it per unit of time than we lose, you know? Right. Yeah, I mean, I guess if I haven't actually talked. thought about this, so I'm going to say things that could be disastrously wrong, but probably Venus, even though it has an interesting looking atmosphere, the surface of it from the few pictures that we have, doesn't look that different from the moon, right? I mean, it's kind of like there are some rocks lying around and nothing more organized than that. And probably there are conditions specifically to Earth related to the existence of both solids and liquids and things like that that

allow for these channels to open up and complexity to develop. Is that the right way to think about it? I think so, yeah. And of course, this is a tough question. If we would have a succinct answer, we would have had, you know, like the sound bite that solves the problem of the origins of life. I know, yeah. But I do think that even though we might not have that,

we do have clues of some of the conditions, you know, that lead, you know, to the creation of complex structures such as life. And one of those is the need, you know, to have energy flows. But the need also of not losing them that quickly, you know, and to be able to preserve the structure. And here on planet Earth, we do that by, you know, having certain solids and crystals that support that structure. For instance, you know, DNA is a very stable molecule that is able to preserve a lot of complexity and information over, you know, long periods of time. And, you know, with the ability to replicate, it allows, you know, as to have the conditions that we need not only to preserve information, but then to make it grow.

And it's not even perfectly stable, right? I mean, if DNA were absolutely stable, never changed, wouldn't do the job. You need this kind of this flexibility, I guess. Exactly. You need to be able to explore, you know, those spaces of configurations so that you can actually then, you know, grow in complexity because complexity requires diversity. So there's some simple version of information, which is just that, you know, when entropy is low, there's secret, in some sense, there's a lot of information because you know a lot

about the system. But you're making the point that it's this complexity that gives us the ability to really make use of that information. Is that the right way to say it? Yeah, I do think that complexity in many ways is kind of like a better term, you know, maybe because it's a little bit more loosely defined for most people, you know, but I do think that, you know, when you talk about the planet Earth as a planet that has a lot of information, maybe people, you know, think about like the media and the libraries.

When you think about it as a place that is very complex, maybe people get a better idea that we're talking about like that, that complexity that is involved in ecosystems and our society that is absent from the moon and it's also absent from libraries. Ask yourself, what are your best people spending their time on right now? Expense reports, receipt chasing, month in close that takes weeks. You become what you spend on, and that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts. AI agents that handle the manual stuff automatically.

So your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF. All right, quick quiz for the hiring managers out there. What's worse? Being understaffed or being poorly staffed? Well, that's a trick question, because both are recipes for chaos. Either way, just say to yourself, this is a job for Indeed's sponsored jobs. You'll get matched with candidates that meet the skills, certifications, and everything else you're looking for.

Or go a different way and get no traction. Seriously, sponsored jobs posted directly on Indeed are 95% more likely to report a hire than non-sponsored jobs. It really is a no-brainer. Spend less time searching and more time actually interviewing candidates who check all your boxes. Less stress, less time, more results. When you need the right person to cut through the chaos, this is a job for Indeed's sponsored jobs.

And listeners of this show will get a $75-sponsored job credit to help your job get the premium status it deserves at Indeed.com slash podcast. Just go to Indeed.com slash podcast right now. Indeed.com slash podcast. Terms and conditions apply. Need to hire? This is a job for indeed sponsored jobs. Yeah.

Yeah, so there's this interplay. I wish I understood it better, maybe because no one does, between complexity and information, right? In some sense, you're making the point that complexity makes use of information and information makes complexity possible. So I'm not sure they're the same thing, but they're at least symbiotic. Yeah, I do think that they're related. And a part of kind of like a set of symbiotic relationships.

The other aspect that I do think is important on that relationship, is what I call the capacity to compute, you know. So if we go back to the DNA analogy, you can think of the DNA and you can think of the cell. And DNA by itself is quite useless. It cannot reproduce by itself. It requires all of this machinery, you know. And the same is true for a lot of the information that we have.

It's like a recipe without a kitchen, you know, and without a cook, you know, cannot transform itself into a dish. So we do have also that ability to then grab a piece of encoded information as a set of instructions and transform it into something. That ability to transform information into new information or to reproduce it or to recombine it

is that computational capacity that we've observing biology, we've observing society and that I think is the true mystery. So I call that like knowledge. And my separation between knowledge and information is information is what is encoded and knowledge is this ability to make. And you can make things by, you know,

making a car. A car is information, you know, it's an organized structure. just like DNA, you know, or, you know, you can have knowledge when you are making a new cell type, you know, as cells differentiate, you know, and that knowledge, that ability to make is ultimately what is hard to accumulate both at the biological level and at the social level. And you said the word before compute, the ability to compute, now you're saying the ability to make. Are those the same thing?

Like in a lax language, you know, when we're talking. You can be lax, don't worry. Exactly, at 30,000 feet away from or more, you know, Yes, I would say, you know, you have this order structures and you have the ability to make those order structures. You know, that ability to make, we can call it the ability to compute, the ability to transform a string of bits into another string of bits, where those bits are encoded on a magnetic tape or on a piece of DNA. You know, from this perspective, it would be relevant. Of course, you know, there are in other situations in which you want to make those distinctions. Right. And presumably there's also phase transitions or at least transformations along the way, where

where the system becomes better and better at accumulating and using information. Yep, yep. So life would be one multicellular life. I'm thinking of all these things. I just had a podcast a little while ago with Kate Jeffrey, who was a neuroscientist. And I had given a talk saying how complexity can evolve, and she wanted to say, yes, but also it goes away sometimes because they're disastrous events. Exactly.

It's not at all guaranteed that it comes and just grows monotonically. Yeah, indeed. and I think that's true for, you know, economy, societies, and ecosystems. You know, like we've seen the collapse of ecosystems. We might be risking, you know, a big ecosystem collapse now with climate change, you know. And we do see it in social process. It's like process of social unrest, you know, there are countries that sometimes everybody thinks that they're going fine and, you know, things, you know, turn very quickly, you know,

like what's happening in Chile this week, what, you know, has happened in many places in the past. Yeah, no, worldwide this is definitely going on. So let's make that transition. So, you know, let's presume that in the last 15 minutes, everyone understands the origin of life and how it takes in low entropy energy. At what point does life become economics? At what point do we talk about trading information back and forth like that?

So what I try to communicate in why information grows. I don't know if I succeed, but what I try to communicate is at the end of the day, you have these systems that have a finite ability to accumulate knowledge, to accumulate that capacity to make. And the only way that those systems can transcend that limited capacity is by developing, you know, collective phenomena or collective systems that include multiple units. So you go from single cellular organisms to multicellular organisms because you could never

achieve the level of complexity of a multicellular organism with a single cell organism. But multicellular organisms, you know, they peak at the human, let's say, you know. So far as we know. As far as we know, you know. And humans are also of limited capacity. The older you get, the more that you realize that you know very, very, very, very little, you know, about everything that could eventually be known. So humans transcend that capacity by forming teams. Teams transcend those capacities by forming organizations.

Organization transcend those capacities by belonging to industries, you know, and also by having cities. And you have this different, like, Russian doll structure of organizations that then grow all the way to our planet in which we are able to accumulate more complexity and more knowledge, more of an ability to make by always, renormalizing ourselves into, you know, groups of the units that bumped into a ceiling. So division of labor in some sense is really what's getting us all this mile. But division of knowledge, which I would say is different than the division of labor, because like I can have a division of labor in which we're all doing the same thing and we have more of us doing the same thing, you know. But the division of knowledge is quite different because we're doing different things and we're passing on, you know, those inputs to each other.

as a way to create things that would be impossible for each one of us to do. So when you're making an aircraft, it's not that you have 100,000 people company in which everybody is making an aircraft by themselves, you know, with a hammer and with some metal, you know, it's everybody's doing something different and that allows them, you know, to create a few aircraft a year, you know, that are the product. Now, if you are in a lower complexity, let's say, activity, like the production of T-shirts, in that case, you might have more of a division of labor

and less of a division of knowledge, you might have, you know, warehouses full of seamstress, that they're all doing the same task, doing the same shirt, and in that case, you have more of a division of labor, you have economies of scale, but you don't have that division of knowledge that you would have served in a complex industry like pharmaceuticals or aircraft manufacturing. I presume that for a complicated aircraft, there's literally nobody who has all the knowledge requires to do it. Of course, yeah, that's the whole point, and that's the idea that, you know, basically, you know, you can think of the complexity of an industry as the number of what I call person bites of knowledge.

Person bytes. Yeah. B-Y-T-E-S. Exactly. Of knowledge that you would need, you know, to be able to create a product. Yeah. Okay. But was that, did we skip ahead?

When I think of economics in general, I mean, I think of trade and barter and, you know, maybe currency and value. And those can precede the sort of division of knowledge. or not, I'm not sure. Yeah, so in some way, you know, my way to enter economics has been through a gap in the literature, you know, that was there in which products were very much considered like some sort of epiphenomena

that was not very differentiated in the economics literature. So, you know, economics is a field that comes more from tradition of like bankers and merchants, you know. So it's kind of more like about trade and interest rate and prices, you know, and money and money. and the cost of employment and those things. But whether you're producing ladders or you are producing apples

or you are producing cars or you are producing t-shirts are things that are differentiated not too much in the models and the empirical work that was traditional in the literature. And I think I was lucky to enter the field at a moment in which more fine-grained data was becoming available and that allowed us to start characterizing products and industries in a more fine-grained matter, not just by talking about how much labor or capital they need,

but by actually looking at the patterns of production to be able to infer how much knowledge they need to be produced or other properties that differentiate them in ways that the traditional literature had not maybe paid attention to. So do you think that kind of differentiation was there from the start from the first primitive economies or is it something that allowed economies to take off later in the game? No, the differentiation is actually very hard to achieve.

So one thing that is quite conspicuous when you look at international trade data or you look at data on the geography of industrial, economic, or innovative activities in a country, is that as you move from the simpler economic units, the towns to the more complex, the cities, or you move from countries that are relatively down in the development, ladder to the ones that are at the top, you have like this subset structure

in which diversity grows, you know, with the level of development and complexity so that the places that do few things, they not only do few things, they do things that are common to the places that do many things, you know? So you have kind of like this world in which creating that diversity is difficult

and those that don't have it are stacked not only with a primitive and limited offer but one that is very redundant with everyone else. So it's very uncompetitive too. So these are also mechanisms that would actually help us explain, you know, part of the inequality because, you know, if what I can do is something that everybody can do and it's a few set of things, it's very hard for me to be competitive and have a decent income. If I can do a lot of things that nobody else can do and people want, I think I have it made.

I want to pause to talk about Peloton home exercise bikes. This is really one of my favorite ways to work out. mostly because it's just so flexible. The idea of having an exercise bike is a great one. It's a really good workout. Cardiovascular fitness is important. We all know that. But the idea that you can do it whenever you want

and still enjoy the features of a class with an instructor sort of egging you on. So I don't know how your schedule works. I don't know when I'm going to have time to work out and so forth. And also my motivation comes and goes. So I like having someone there telling me what to do. It's a virtual person if you're on Peloton. So you can either sign up for a live class or a large library of pre-taped classes.

Now Peloton has a new 30-day home trial. So you buy the service and try it out, worry-free, for 30 days. If you decide it's not for you, you can return the bike for a full refund. Peloton will even come and pick it up at no cost. So this holiday, give the gift to Peloton, use a promo code, Mindscape, at OnePeloton.com. That's O-N-E-P-E-P-E-L-O-T-O-O-O-N.com. and get $100 off accessories when you purchase the Peloton bike.

But, okay, but so when I think about division of labor, and maybe division of knowledge goes the same way, I think about the industrial revolution. I think about Henry Ford building the Model T and an assembly line. I have too many questions to ask. That's why I'm hesitating here. How did that get started? Who invented that?

Was that a sort of side product of other technological information innovations like the printing press? Is there a story to tell about how information bred upon itself? So now we have more of like a big history question here. So in that big history perspective, I do think that there's kind of like a long line of events, but there are also some transitions that I think are very important. The first one that someone like Yuval Harari

emphasizes quite a bit on sapiens is like the cognitive revolution. It's like the development of language. And there's some particular properties of language that we have. because human languages are not like animal communication systems. We have the ability to talk about hypothetical things, you know. So an ape can have the ability to maybe tell another ape that there's danger or that there's a lion or there's a snake, but they kind of tell them about their idea for their new blockbuster film.

That idea to talk about fiction was a big, you know, cognitive revolution that happened, you know, 80,000 years ago, 100,000 years ago, you know, and that is expressed in things like the development of, you know, like more advanced tools and an acceleration of our ability to, you know, accumulate knowledge and so forth. Then we have the agricultural revolution, you know, which people know about. That's, you know, much more reason about like 10,000 years ago, you know, with that, we finally start accumulating slowly people into cities, you know.

That agglomeration is really important because at the end of the day, you know, just like biology, you know, accumulates knowledge, you know, in DNA and that knowledge has to be preserved and pass on to the next generation. we humans accumulate cultural knowledge, and if we're all dispersed in small groups, the ability to accumulate knowledge and ideas through generations is limited. But once we start forming cities, we start living together, we're going to accumulate more knowledge. We're going to have maybe also people that are going to be dedicated to those aspects of life.

In the beginning, maybe religious leaders and political leaders and so forth. And the next revolution, I would say in the sequence of events, at least the way that I understand them is the writing revolution, you know, that is very important in ancient Greece, you know, like even though writing is older than ancient Greeks, you know, it was not that prevalent and it was not that advanced, you know, it was like much more based on like, you know, accounting systems or even in administration and religious events. But in ancient Greece, you know, writing kind of like explodes as a form of expression that is used to communicate, you know, and document like complex ideas, and they go through something that could be considered similar to the Renaissance,

you know, that happened later. You know, they actually get close, you know, to maybe, you know, like what could have been an industrial revolution, maybe, you know, if history would have worked differently, you know, and that writing revolution produces a lot of knowledge. And then after the writing revolution, which I would date, you know, loosely around 700 BC, to me, the next big revolution and what is the starting point of everything that, that is modern is the printing press, you know? Because when Gutenberg, you know, adapts the printing press, you know, that existed in Asia

and develops this removable type printing press, what happens are a lot of things. First, you know, and this is all from Elizabeth Anchenstein's work, you know, the printing press as an agent of change. But first, you know, for the first time, scientists and scholars can have access to multiple books. like before printing, like only kings had little libraries, books were transcribed by hand, and everybody has written a book,

quickly realizes that no matter how cool your life is, you don't have enough stories. You have to read and share other people's stories to fill up your books. Second, you know, what happens also is that printing is the first economic activity that is urban and scalable.

Think before printing. What was the way in which you could make money that you could make it big. You know, well, you know, you had to have access to resources where it is a lot of land and a lot of serves, you know, that would, like, farm that land, where you might have access to maybe mineral resources or you, like, looted next door town or something like that.

But printing was something that you could produce in a city, you know, with a relatively small team of people. It was kind of like, you know, sort of IP intensive, very IP intensive. But if you have a book that sold and you could sell copies, you could, you know, become rich really quickly. And evidence of that, is that the number of printers per capita in Europe stabilized after only 50 years.

So it was a huge... Zero to 60 very quickly. Yes. Yeah. It was a huge economic boom. Like in 50 years, you know, you stabilized that number because you had an activity that basically was very profitable, was very urban. So it's the first time that kind of like someone in the city, you know, can really start

making it big by manufacturing something at scale, you know? It's really the first type of like, you know, mass production is printing, you know. And then we It was content, as we would call it today, right? Yes. That was the point. You had to monetize your content for the first time ever. Yeah.

So that I think is huge. And it gives rise to like a new cognitive revolution, you know, that we call the enlightenment, you know, eventually, you know, that involves a lot of the most important discoveries of, you know, science in the middle of last millennia. And after that, you know, as a society chugs along and, you know, the 16th 17th century comes along, you know, printing accelerates, you know, like it takes 200 years for people to discover that you could print short formats, okay? Okay.

Like magazines and pamphlets and so forth. Eventually those lead to a change of institutions, you know? So, you know, it's hard to understand the transition from, you know, monarchies to democracy without printing. And, you know, after those changes of institutions, you know, and with the acceleration of science and technology, we develop new forms of communication. like, you know, film and radio, then television, now the internet. And I like to think of the history of our planet or of our last, you know, thousands of years,

as a history of changes in communication technologies that have reconfigured the way that we create and process information and that we generate and produce knowledge. And those are the eras that, to me, have contributed to this big history. You're confused about your credit score. One site has one number and another site, Something completely... What? That can't be right. It's okay.

Forget everything except MyFICO. These free scores from other apps can differ by as much as 100 points from your FICO score that 90% of top lenders actually use when you apply for a credit card, personal loan, car loan, or mortgage. For the moments that matter, get the score that matters, your FICO score. Visit MyFICO.com and get started for free today. Ask yourself, what are your best people spending their time on right now? now. Expense reports, receipt chasing, month-end close that takes weeks. You become what you spend on,

and that's not what you're building toward. Brex is the intelligent finance platform that eliminates that work before it starts, AI agents that handle the manual stuff automatically. So your team can spend their time on what actually compounds. It's time to get Brex AF. Learn more at brex.com slash AF. Yeah, no, that's very helpful because I guess we glossed over a little bit, but in your book, you certainly emphasize the fact that not only there's this thing called information, but it's embodied, it's crystallized, right? There's some stuff that carries the information. And all of these changes in communication technology, which sounds sort of mundane when you put it that way,

they're new information flow technologies, right? And if you make it cheaper for information to flow, you enable a lot of new things. They're huge, and they always laughed upon when they first merge, you know? Like, nobody took Twitter seriously. and probably now is the first thing that politicians take seriously. You still complain about it, but yeah, it's important. It's quick.

For instance, think of like the music industry. A lot of people are sort of surprised that musicians kind of make money anymore, but if you think about it from the perspective of communication technologies, what it's curious is that there was a short period of time in which they could make money. Yeah. Because there were musicians

at the time of the ancient Greek. their musicians is still today, but for most history, musicians only could perform live, and live performance were hard to monetize, they couldn't spread, but as, you know, technology evolved and you had radio, and then more than radio, you had then records. You know, those records, you know, allowed to constrain the diffusion of music in such a way that you could monetize it because music was trapped on those, you know, discs, it was trapped, you know, on those magnetic tapes. And you could make a killing because, you know, the marginal cost of producing a new tape or a desk was very small.

You know, you could sell them for $10, $15, you know. And all of the musicians that were big in the 60s, 70s and 80s, they're loaded. Nowadays, you know, it's impossible to make money that way because music cannot be trapped in a physical media. And what was curious is that musicians could make money for a while. So that's interesting way you put it. It had to be shareable but not too shareable. Exactly. Yeah, yeah.

There's that sweet spot. I wonder what lessons there are for other things. I wonder if it's the same way. I mean, certainly as someone who writes books, you probably feel the same way. The first thing that happens when you write a book is pirated versions appear on the internet. But they're not that better than the physical books. So I don't think it's doing a huge amount of damage. That's interesting. Okay. So that flow of information through different media obviously affects what people can do. But then how, where do we get to things like Henry Ford? And you have this wonderful example in the book of this giant plant. I've never heard of it before,

but there was, you know, what was it called? The River Rouge. River Rouge. Yeah, I didn't know about this before. But to this day, that was the largest, you know, integrated facility. I don't know because I've been to China a lot and like, like they have amazing stuff there. But it didn't grow monotonically. There was a peak at that moment. Yeah, exactly. So the River Rouge, you know, was this, you know, fantastic plant, you know, that Ford created that, that literally took like soybeans and metal, you know, on one end and produced cars out of the other because it was a complete vertical integration. I would say today we don't manufacture anything with that level of vertical integration, like value chains are very distributed and international. That's, for instance,

why, like, Trump tariffs are kind of like stupid when it involves Mexico because, you know, the U.S. value chain in the automotive sector, like, you know, which is the number of one country in the exports of beer. It's also Mexico because a lot of like the US beers manufactured. So like the value chains are integrated and disaggregated between different countries. And we don't have those monolithic models. But at the time, you know, of Henry Ford, you know, where like transportation was not that good, that's why, you know, those slow cars were such a good business, you know.

You needed maybe to produce that level of vertical integration to be able to get the economies of scale that he needed, you know. to produce affordable vehicles. Nowadays, things are different, but it's impressive, you know, what they were able to do at some point. I mean, again, maybe it's a version of a sweet spot in that they had the differentiation of knowledge enough to do the economy of scale,

but it was still expensive to do things all over the place. So bringing it together in the same geographic location was a useful thing. But I think in the book you had this example, I'm going to mangle it, but, you know, in Chile, you can mine copper and other raw materials, and you would like to have a battery and you can in principle make them,

but it's actually cheapest to just send your raw materials to Korea, have them make the batteries, and then buy it back. Yes, so that's a good point because at the end of the day, you know, like for us to produce things, you need to combine, you know, like, you know, materials, technology and knowledge. And the way that the world works is that the world is kind of like lazy, yeah? Like basically we try to minimize costs. That's a more economist way of putting it.

But because of that, you have to answer. of the question, what is easier to move? Is it easier to move the knowledge? Is it easier to move the materials? Is it to move the technology? What are the factors that are easy to move? And the hardest factor to move is knowledge. Which seems weird. I mean, knowledge is not very heavy,

but it's hard to move it from brain to brain, I suppose. It's super heavy because people tend to confuse knowledge with kind of like encoded pieces of information. So, like, would you trust a brain surgeon that the only thing that has done is read Wikipedia pages about

brain surgery and has never been there. So, like, knowledge is very experiential. Or even read a textbook. Exactly. Yeah. Someone that had just read a textbook on brain surgery, I wouldn't allow them to operate, you know, and they don't, you know, like, that's why you have residences and you have all of this, you know, practice systems, you know.

But when it comes to, like, complex industries, you know, the knowledge is embedded on large teams and that makes it very hard to move. So knowledge has this temporary monopolies. So, for example, when Ford figures out how to build a car. and he's able to put all of that together on the River Rouge and he's producing cars. And there's other people producing cars also not that far from him. They're not producing cars in San Diego. They're producing cars, you know, like they're in the Midwest.

Exactly. And Detroit, yeah. And it spans kind of like from there because knowledge is hard to move. So it's easier to bring the steel there. It's bristling to build the, you know, the coal and all of the other materials. And nowadays, I think it's kind of like the same, you know, like Silicon Valley has monopolies over markets that they discover. China is now getting there

because it's a country that is very technologically advanced and sophisticated and the products are easier to move that the knowledge that you need to make them. So if you figure out something that people want, you're going to have that monopoly because the product is going to diffuse very quickly.

But the ability to make it is going to diffuse very slowly. And until it does, there's going to be only a few people that are going to be able to supply that and therefore, you know, they're going to have their day. And this is related to that distinction you draw between knowledge and know-how or explicit knowledge, I guess, and know-how. I mean, in some sense, there are things you can put in a book, but there are things that are just in the

human brain or just in the individual people. And that's why it's heavy, because moving people is really hard. Exactly. So in the literature on knowledge that is used traditionally on business schools and knowledge economics, innovation economics, people make a strong distinction between tacit and explicit knowledge. Explicit knowledge is all of the knowledge that I can communicate through an act of communication. I can codify that knowledge. Like the recipe that I can put on a cookbook

is something that I could communicate through a page and therefore is explicit knowledge. The experience of having cooked with, you know, the chefs from El Bully or any famous restaurant is something, you know, that is much harder to communicate so it would be considered tacit knowledge. So the best examples of tacit knowledge are think of sports.

So imagine like Michael Jordan. He's extremely talented. He really knows how to handle a ball on a basketball court. But imagine you have a seminar and he speaks and tells you about basketball for like three days, you know, how much better of a player are you going to be after the third day? Epsilon, yeah. Yeah, not that much, you know, because, you know, that knowledge is tacit and it's acquired through practice and the world is full of tacit knowledge and is the one that is hardest to see because it's not as obvious as the knowledge.

is that we can codify. And closer to our experience, I don't know if you've had a professional basketball in your career in your past, but graduate school is the same way, right? You know, students come in, they know a lot of equations or whatever,

but they don't know what it means to be a scientist. They haven't seen it in action. They haven't done it. I agree. Yeah. So I came to the U.S. for grad school, and I had a very good advisor. You know, Laszlo Barabas is a very famous,

you know, physicist that works on networks and so forth. And I remember, like, soon I learned that it was a waste of time to go to him to kind of like talk to him about like technical details. That you should figure out by yourself. But what I wanted to get out of him is like to understand, okay, when a problem is relevant, you know, how should I communicate to people, you know? How is different people going to interpret this?

All of those type of things, you know, how to think about, you know, like a career and to connect the different things that you're doing. And those are things that are hard to learn if you don't have a model of someone that knows how to do them and you're kind of like learning as an apprentice. And I think a lot of people sometimes don't understand that. I think the 20s, you know, you better get someone, you know, that is where you want to be and be a very humble apprentice of them. So if we take this on board and we appreciate the importance of tacit knowledge and know-how and how difficult it is to move around, what are the lessons that we get from that for either building an organization or organizing

economy or trade barriers and things like that? So there's a lot of lessons because there's a big literature on this, you know. And so there's a literature that looks at the geographic diffusion of knowledge, you know. And as you can imagine, because knowledge is sticky, there's a lot of barriers to the diffusion. So it's hard for knowledge to travel long distances. Now, the reason why it's hard for knowledge to travel long distances is because it's socially embedded. So, like, it's actually that social networks are geographically circumscribed, and that limits the diffusion of knowledge. So then the question is, what are the things that are?

limit the ability of people to create links, you know, you have language barriers, you have cultural barriers. All of those have been shown extensively to limit knowledge diffusion, to limit trade, to limit other things. So the question then is like, now, if you're a country or a city or a region

that wants to develop their economy, what you're trying to do is to accumulate knowledge, and the question is, how do you do it by following the laws of knowledge diffusion? Okay. And here are examples of people that did it wrong, and examples of people that did things better. You know, examples of doing it wrong.

I don't know if you ever heard of the University of Yachai in Ecuador. No. So earlier this decade, Rafael Correa, who was the president of Ecuador, decided to put a billion dollars, you know, on the creation of a new city of science and technology that he hoped would compete with knowledge production centers across the world. very idealistic, you know, plan, you know, a billion dollars in Ecuador is 1% of GDP, okay? So that's a lot of cake, you know.

And basically, you know, what they did is they grabbed a piece of agricultural land, like two hours north of Quito. And they tried to start building kind of like this university and city and industrial park and so forth. But if you walk around, you know, place like Kendall's care of Manhattan, you know that you don't build a lot for a billion dollars, you know. And if you have to bring every brick and every person that lay bricks to the place, you build even less. So quickly, like, the plan, you know, like unraveled, you know, that university has gone through like six or seven, you know, university presidents by now. They were able to track a few scientists to move there. You know, that didn't go well. You know, there was an article in science, you know, talking about their experience and how they, some of them were leaving.

Some of them got fired. They got a few students there, like at the beginning, like I think it was just a thousand, you know, that became very radicalized because they believed on the dream. But, you know, a thousand dollars, you know, sorry, a thousand students at a billion dollars, you know, that's a million dollars per student. So you could have put all of these kids in Stanford, you know, for like life, you know.

And, you know, so they were like, you know, and basically it was this idea that if you have enough money, you could create knowledge anywhere. Yeah. And that's not true because their process of diffusion that constrain the creation of knowledge. In the case of Ecuador, they had two chances. One was called Guayaquilda. One was called Quito, which are the like two centers, you know, where they have accumulated knowledge.

So what are the channels that can promote knowledge diffusion? So one of them which is really important and it's also very well documented is migration. And migration is one channel that is also very biased towards the most talented people in the world. So there's a book by Bill Kerr, which is a professor at HBS, that is called The Gift of Global Talent. And in that book, you know, you find a lot of interesting facts. One of them is, like, people, you know, without a college degree, about 1% of them migrate. People with a college degree is about like 5%. Inventors that have filed a patent, about 10%.

Nobel Prize winners about 31%. Nobel Prize winners in the US since the 70s is like 60 something percent. Yeah, we get them off everywhere. Exactly. So you do have kind of like this thing that there is a tale of talent that is extremely global and that is the one that helps create innovation, the ones that help create jobs. You know, it's similar. You get similar numbers if you focus not only on formal education and academic credentials.

If you look at people that have formed, you know, Fortune 500 companies or unicorns, They're super biased towards foreigners, you know. So then, you know, one of the lessons that you need to do is, you know, well, how do you attract that global talent, you know, because at the end of the day, it's a game of global talent because also this talented, famous individuals are the ones that are going to help you attract other ones. The U.S. for a long time has had a huge advantage on that space. It receives about 50% of all of the PhDs that migrate in the world come only to one destination, which is the U.S. And now that... The falling percentage, I presume, now. May it change.

Yeah. Yeah. This June, the world comes to Los Angeles. Kick off FIFA World Cup 2026 at the FIFA Fan Festival at the iconic Los Angeles Memorial Coliseum. Watch matches live on Giant Scream. Feel every goal with thousands of fans. And celebrate with music, culture, and flavors from around the world. Join us June 11th through 14th opening weekend as the tournament kicks off in Los Angeles.

Tickets are just 10. $10 in kids under 12 or free. Get yours now at Los Angeles FWC26.com. Ask yourself, what are your best people

spending their time on right now? Expense reports, receipt chasing, month in clothes that takes weeks. You become what you spend on, and that's not what you're building toward. Brex is the intelligent finance platform

that eliminates that work before it starts, AI agents that handle the manual stuff automatically. So your team can spend their time on what actually comes compounds. It's time to get Brex AF. Learn more at brex.com slash AF. Well, that's a very believable story. I had two previous podcasts, which made similar points, one with Jeffrey West,

about how things scale and, you know, the ingenuity and creativity certainly scales with population density in some way. And then another one with Willowson, where we talked about the political aspects of these and the openness to new ideas that was associated with with cities versus more conservative rural divide. But then, so that raises the question that you brought up the word inequality earlier. You know, it's great to attract the best talent and, you know, they want to be in these concentrations of brilliance and productivity. But then how do you spread the rewards of all that productivity and brilliance widely to

everyone? So that's tough, but I do think that we have found out a few things, you know, quite recently, actually. So without going into the details, we might go into this later, but there are ways of measuring the knowledge intensity of economies. Like how much knowledge is there in New York, vis-a-vis San Diego, Tokyo, or whatever. And when you use those measures to look at knowledge concentration in cities or in countries,

you do find relationships with inequality as well. So, for instance, economies that are more complex tend to be less unequal, economies that are more extractive and less complex. sorry, economies that are more complex are less unequal. More equal. More equal, exactly. And economies that are less complex are more unequal. Okay.

You know, so it's not obvious to me. It's not obvious, but like it's really hard to have the level of inequality of Switzerland with the industrial structure of Peru, you know. So if your exports are based on like three, four sectors mainly, you know, when those sectors are about, you know, mineral resources structure that, you know, are industries that are very regimented and hierarchical, you know, because they're about, like, you know, safety, about production. They're not about creativity.

You don't want a minor to start digging anywhere and get creative, you know. You don't want creativity back down there, yeah. Exactly. So you have productive structures that are not geared towards, you know, innovation and, you know, gears, like towards structures that are more hierarchical and unequal. So at the international scale, you know, as economies become more complex, they become more equal. And as economies become less complex, they become more unequal. So for countries like, you know, like Chile and, you know, Peru and Angola, all of these countries that depend a lot on mineral resource destruction, whether these are fossil fuels or other forms of minerals, one thing that they need to do to reduce their inequality is they need to sophisticate their practice structure so they can generate those middle income jobs.

because those middle income jobs are not going to happen, you know, just through the redistributed policies. You do need policies and social safety nets, but you do need to change the practice structure. Now, if you look inside countries, the relationship flips, which is kind of cool, you know, from a statistical perspective. Why that might be our hypothesis right now is that because within countries, you do have more spatial equilibrium. Okay. So a lot of the effects that Jeffrey West talks about, you know, are effects that are not just about the size of cities, but also about the migrable. into cities. So Peter Hedstrom has a paper on science advances

that finds that a lot of those super lean and scaling relationships described by Westamberancourt and actually quite explained by migrants coming to cities and those migrants being the most talented people. So the most talented people from the Rural-A-Relis are the ones that migrate and help provide that extra punch of productivity that cities have. So a city like New York is very unequal,

not because it's only based on New Yorkers. It's because a lot of poor people migrate to those places Similar to San Francisco, you know, they're like big magnets and attractors of people. And at the country scale, the more complex a region or a city is, the more unequalities. The least complex, the more equal it is. So you have a Simpsons paradox, you know, not from Homer Simpson, but the statistical symptoms of paradox that the correlation that you serve at one scale, you know, inverse when you disaggregate that data into the next scale.

Right. Okay. Well, so then what's the, is there an immediate policy implication for this? So what are we going to do with this knowledge? Yeah, so I do think that there are a few things. One thing that we've found also related to this is that the more complex economic activities concentrate in space much more than the least complex economic activities. That paper is coming out on nature of human behavior in a few weeks.

And what this tells us is that, well, if at the end of the day, wealth is going to be increasingly generated in cities, we do need to have better ways of including more people into cities. Because I think one of the pain points of the United States right now is that the major cities of the U.S. are failing to accommodate more people because of problems with infrastructure, you know, problems in the ability to build and so forth. So, like, if you think about it, like San Francisco.

San Francisco is the worst, obviously. That city in China would be a 20 million people city, given, you know, like what is able to make. But for that, you know, it would have, you know, 27 lines of subway, okay, that would be, like, really fast and modern and autonomous, you know. it would need to have another type of infrastructure. Like, I don't know if you've been to Shenzhen. I've not been there, but I know about it.

Yeah, Chen Jens. It just came into existence from nothing. Yeah, exactly. It's 23 million people. Like, so the U.S. hasn't been able to produce 21st century megacities. It produced 20th century megacities. But nowadays, you know, I think you're going to need to include more people in the centers of production.

And if you don't get together, you know, good infrastructure for transportation, you don't get together good ways of, like, densifying neighborhoods. in a way that they make them livable, but at the same time are able to include maybe twice the number of people that they had before, maybe three times. You are going to start having more and more social pressure because it's unsustainable to commute for three hours, you know. People are getting excluded because cities, you know, have the need to include more people than the one that they're able to write. Yeah. I mean, just to play devil's advocate here, I'm probably mostly on your side, but these are subtle questions. We're having this conversation in the middle of a charmed. neighborhood of Cambridge, Massachusetts, with a lot of small individual-sized houses, or at least

few unit houses, the character of the place would utterly change if the density were that of a large Chinese megalopolis, right? And is that entirely good? Well, you know, in some ways, I live two blocks away from here. Yeah. That's a huge privilege. Yeah, you know? That's a huge privilege. People don't want to give it up. Yeah. Many people in my company have to commute much longer, and that commute is part of kind of like their unhappiness. And if you think about it, a lot of the social unrest that is happening in many countries, you know, is happening, you know, is being triggered when the price of gas goes up, when the price of transportation goes up because commuters are unhappy, you know, commuters are the ones that

are kind of like excluded and they have to kind of like make a journey every day to the places where people like me that have the privilege to be able to afford a home here, you know, live. So I do think that if they could make Cambridge more, you know, affordable to more people, you know, I think it probably would be healthier. You know, maybe we'll be able to hang out at night more and get better ideas for our businesses and for creative activities that we want to do. So I do think that if you get along with your neighbors, it's good to have more of them. It is. I was impressed to learn, not impressed, but interested to learn that the average commute time goes up. up and up, the more you're in a big city.

Yes. So despite the fact that things are more dense and nearby, it takes longer to get there. Yes, yeah, yeah. New York has huge commute times. Yeah, in L.A. Yeah, so we have all of that data. You know, we have data USA.

We have the main tool to distribute and visualize U.S. public data that we launched in 2016. And there you can create maps of like commute times and you immediately see that when you create the average travel time map, boom, all of the cities light up. When you create the commuting alone map, then, you know, it's the opposite. You know, so commuting alone, you know,

is people driving cars in more rural areas, you know, average travel time, you know, goes up, you know, in places like D.C. and, you know, New York and so forth. Well, that's another side of what you do. So if information flowing around in different forms and in different media is driving the economy in various ways, we're living in an era now where there's so much information that just dealing with it all is a big issue. So you've been involved in data visualization and projects that try to sort of make sense of all this information we're getting? What drives you to do that?

So it was a little bit of a, let's say, lucky path, you know? So when I started doing this work on economic complexity and relatedness, there were like two big papers that we produced, one in 2007 that came out in science and another one on PNAS in 2009. And those papers became, you know, very popular. But one of the things that those papers had

was that they had some non-traditional techniques that we had invented, that you could use to predict the products that a country was going to export in the future, that you could use to predict the future economic growth potential of countries based on their economic structures. And I started to get a lot of demand, you know,

of people to generate, you know, reports on that type of topic, you know? And as a scientist, you know, like you always want to be working on the next thing, you know? So, like, they say, oh, yeah, we want to do like, you know, like a relatedness and complexity analysis for this region of Brazil or for, you know, this or that. And that was kind of like boring, repetitive work.

Sure, you've done that already. Yeah, you've done it. So when I started my lab at MIT, the first grad student that I hired was Alex, Simoese, and the job that I gave him is, we're going to do like a self-service tool for this demand. I had done something similar before for a paper in which we look at, you know, correlations between diseases using hospitalization records. And Alex started building, you know, this tool, which became the Observatory of Economic Complexity, is now the number one tool to distribute international trade data in the world. And then we found out that, you know, maybe like what people were more interested was on the platform and the tool, rather than, you know, the analysis that the tool made. And that was redeployable.

So then we created a tool, you know, together with the government of Minas Gerais in Brazil that, you integrated data for more than 50 million workers, all of the basically formal sector economy of Brazil, data on trade, on industries, you know, and employment and also education for all of Brazil. And then, together with a colleague from Deloitte, we started working on the creation of a similar platform for the US. But one of the things that we did there

is that we realized that, you know, we needed to go beyond economic data, you know, so it includes data on demographics, on health, on insurance, on, you know, commute times, you name it. And that's Data USA that was launched in 2016. That project became very popular, and it became some sort of like the dream thing that people in statistics departments of many governments in the world wanted to have. And that started to generate a demand to create more projects like that.

So, you know, there's Data Korea now, there is Data Chile. We're releasing Data Mexico on January. And we've been creating these tools that what they do is something that is very, simple but it's not easy to do you know it's easy to use but it's hard to do is to integrate you know 15 20 30 different data sets but not just provide files but integrate them into narratives so what we do is we transform data into stories you know that's the main form of integration that we provide that allows people to find the data on the web you know the stories have you know text that is generated semi-algotimically using the data visualizations also you know so imagine for the

The U.S., the U.S. has like 70,000 census designated places. You know, each one of them has, you know, a complete profile, you know, with more than 70 visualizations and a lot of text and information. So even if you're a little town, you have all of your census data and your BIA data and your BLS data USA. And then we have more advanced tools, like ways to integrate data and download it, you know, but integrating data from multiple sources, ways of creating custom visualizations. and that has been something that has done very well.

Also, we've done similar solutions for private sector companies in which we integrate the data from their marketing departments and logistics and so forth to create platforms that people can use in a strategic decision-making. And are these also useful for academics doing studies of demography or whatever? They use it a lot. Just to give an idea, right now on our online properties, we get over a million people a month. Okay.

So we run service, you know, to try to figure out, you know, who they are and how we can serve them better. And like a platform like Data USA, you know, 35 to 40 percent of the people that visited, it's like academic of some form, whether it is like a high school student doing a homework. Yeah. Or whether it is a university professor, you know, using it in some report. Like if you go to Google Scholar and you search for the URL, it would be a relatively well-sighted paper. Like, unfortunately, you don't get crazy for those ones, you know. Maybe one day, you know, you're going to be able to put your websites on Google Scholar as well, but that's another story. So we do get people from academia.

We do get a lot of people from local governments. Okay. Can people just go to the website and use it? Is it a fee for service? No, it's totally free right now. We're thinking maybe in the future to add some premium features, you know, for like, you know, people that are using this. Some people use it, for example, to do market analysis, you know, and we could provide like premium features on that case. But so far it's completely free, open source, you know.

So it's a very open project. And is that similar to you also mentioned offhandedly the urban perception, the idea of using actual photographs of different places in the city to sort of, but not just show them in a slideshow, but learn about them from the computer. So it's related because if you see what I'm interested in, I'm interested in kind of like these applications of science and technology to society. You know, and one of the things that I tried to do over the last decade is to find alternative ways to collect data and hopefully data about aspects of society had been hard to quantify before. So in 2010, I got the idea that we could use Google Street View Images to quantify evaluative aspects of cities, you know, which place looks safe, which place look lively, which place, you know, look depressing, beautiful and so forth.

At that time, a lot of people were like, oh, this is crazy. These are all subjective things. Never going to be able to do it. It's all meaningless. But I discovered that on the one hand, there was a literature on urban planning of people that have attended and had done that many times, but with very, very small sample sizes of both images and people. So what I did is I did a crowdsourcing study and that quickly became the largest data set ever, you know, of visual perception service that was called place pools. and that allows us to classify 4,000 images and we got over 100,000 people rating them.

And we discovered a few things. First, that people's preference were very transitive. So it's not that all over the place. If I show you like a picture of a really nice, you know, like beautiful, well-kempt neighborhood, and I show you like a picture of a, you know, like a very, you know, a sketchy, you know, favel, industrial, you know,

like people tend to, let's say, answer that the first one is safer than the second one, and that tends to be quite a universal preference. So the difference between images is so large that it overwhelms the differences between people. And we could use those scores to like measure, you know, like the segregation. You don't know if it's actually safe. All you know, everyone perceives it to be safe. Yeah, and we don't expect it to be because what we wanted to measure was the perception. because perception can also have an effect that is independent of whether that location is actually safe or not.

So there are different things. They don't have to be the same. I mean, we've all been in cities in neighborhoods where we've said, oh, this looks safe or, oh, this doesn't look safe. And articulating why we're saying that might be difficult. Exactly. So what we discover, though, is that crowdsourcing was very limited for the amount of data that we needed to collect. So a city like New York, Manhattan alone has about 80,000 street segments.

So that's a lot of a street segment. So let's say you want to get one image per street segment, you know, and you want to evaluate that image. And let's say that you want to compare that image with only 10 others, you know, to be able to get a decent score, you know. It's imagine it's like college football, but with, you know, like 80,000 teams and 10 games, you know. Exactly. You know, so it's super underdetermined matrix.

You know, still you need a lot of traffic just to evaluate all of those images for one city and one dimension. So what we decided to do is to say, let's grab the data that we have and let's train computer vision algorithms to do the clicking for us. Okay? And we find that actually

those computer algorithms work very well and they allowed us to scale to create urban perception maps with hundreds of thousands of images. And then we could use that to study how perception affected behavior. Okay, and what did we find? So we teamed up

with a team of Italian and colleagues that had mobile phone data, and they could see the activity of people in cities as a function of the time of the day. So then we could see if people tended to avoid and safe-looking places, controlling for distance to the subway, for distance to the Central Business District,

for other things like the density of jobs, the density of population and so forth. And we did find very interesting effects. So first, we find that people tend to avoid and safe looking places. Like, after controlling for all those things, you find less people in those places than you would expect. And those effects are modulated by demographics. So it's stronger for women and for the elderly,

but it's reversed for, you know, people below 30. So, like, young people tend to like... They hang out in those areas. Hang out on their unsafe looking places, you know? young male especially, you know, but like elderly women tend to avoid those places. And it kind of like makes sense, but it helped formalize something that I think will have an intuition for, but now, you know, we actually have like, you know, hardcore data to be able to show it.

I mean, it's another example of what we were talking about, a different way of thinking about conceptualizing, different kind of information in some sense, or at least a different way of sharing it and thinking about it. And that's the other, the last thing I wanted to ask you about was how. this relates to your interest in collective memory. I did have one podcast guest, Lynn Kelly, who studies memory palaces, ancient memory palaces. So the idea of remembering things by associating them with physical geographical location, she thinks that Stonehenge, for example, was used as a way of remembering what you would call know-how, right, tacit knowledge that

was, because they didn't have writing, they couldn't pass it down that easily. But nowadays, we have the internet, we have books, we have TV, it's a very different kind of thing. is our collective memory of who we are and what kinds of things we pay attention to been affected by these technological changes? Yeah, so I've studied how collective memory is affected by technology, language, and time. So we did a paper with Steve Pinker and other people in which we looked at the network of global languages, okay? So that's a network in which each node is a language, a language are connected if they're likely to be translated or spoken by the same So let's say English might be connected to German if a lot of books get translated from one language to another and a lot of people that speak German also speak English and so forth.

And we map that network using three datasets, a dataset with over two million book translations, Twitter. So we detected the language of tweets and then if you tweet in English and you tweet in Spanish, you know, I can connect those two languages because you are expressing knowledge of both and Wikipedia edits. So it's not reading Wikipedia, but if you edited the page, you know, of Einstein in German and you know, and you edited the page of Einstein in English, you probably know how to write in German, so you probably know both languages. And interestingly enough, when we compare that network, you know, with a dataset that we created of globally famous people,

we found that the centrality of a language in that network explained the number of famous people produced by the language better than the population of that language, better than the wealth of that language. So that tells us, hey, you know, like fame, or global fame, it's something that is very much dependent on this network, because if you think about it, the network of languages is like the most aggregate version of the global social network that you can have.

You cannot have a social relationship if you don't speak a language, you know? You cannot make a friend just by nodding, you know? So in that context, you know, we learned that, you know, much of what the world knows is going to be modulated by this network, you know, and there's a lot of implications about that, you know. So just to be clear, is the network take the form of saying things like, if you know English, you're more likely to know French than Chinese, whereas if you know Chinese, you're more likely to know Korean than to know English,

things like that? Exactly. Yeah, so the network gathers all those links. It's, you know, there are paths between every language and every language. But a language like English, for instance, is a global hub, you know? Yeah. So someone from Portugal and someone from Vietnam, probably they're going to speak something. The most likely thing is they're going to speak English.

Yeah, it's a bad example. But there's also, like, regional hubs. And also, like, for example, French connects a lot with African languages, you know. Also, you know, Spanish connects with, like, you know, Kichua and, you know, Mapudungun and other, you know, languages from, you know, native South America. Then you have also Arabic as being as an original hub, Chinese, you know. And then you have language like Russian is a very important regional have on all, you know, like Eastern Europe and Northeast Asia, you know.

And then you have peripheral languages. So it's a very hierarchical network with like English sitting at the center, then a ring of, you know, regional hubs. And then, you know, like smaller languages that are kind of like on the periphery, you know, connected to some of those regional hubs. I would love to see a picture of this. Is there an image out there on the internet? Yeah, language.comedia, dot MIT. Do you.

Very good. Thanks. And then we also studied how communication technologies affected our collective memory. And we use this data set that we've created over 70,000 famous biographies to look at how changes in technology change the number of globally famous people that were born each year and the occupations associated to those people. So, for instance, you know, when you think of, you know, Einstein, he would be a physicist, when you think of Michelangelo, he would be, you know, a painter. And what you find is that when new communication technologies were introduced, the composition of our collective memory change, not only the size. So before printing, you look at the matrix of our collective memory, and it's mostly political and religious leaders.

Okay. After printing... Bosses of hierarchies one way or the other. Exactly. After printing, you get famous artists and famous scientists for the first time. You get a lot of painters, you get composers. You know, you start getting astronomers. Writers, you know, that are also there.

but then there is the second printing era, which we talked before, when people start doing like shorter formats and so forth, and there you have an explosion on the sciences, then now you start differentiating, you know, natural philosophy into more sciences. You get a lot of famous writers and so forth and composers. And then you have the invention of film and radio,

and that generates a huge shift on the arts. Because before that, it was the painter and the composer. Then it's about the performer. So it's the musician, is the singer. Composers disappear. Musicians and singers are the ones. ones that now become famous, actors become famous, not play writers anymore. You know, so there's kind of like that shift.

And then you have the introduction of television. With television, you create the fame of sportsmen, you know, because sportsmen were not that famous before television. And it's a live performance, you know? Like, it has to be at the right time when you watch the game. And visible in 3D and, yeah, you can experience it there. Yeah, okay. So you do have kind of like that change in the composition of our collective memory with

introduction of communication technologies and it's very clear. So to me, now when I think about like the history, I don't think about like the modern times and the Renaissance and the things that I learned in school. I think of it in terms of like, okay, what was the dominant communication technology at the time? And that's the era that I set myself in. And what's going to come next? Yeah. So the next thing that I is. Instagram influencers. Ah, next on that. Like, well, what is, do you know TikTok? I know of it. I've never used it yet. Oh, dude. Far behind. It's. It's. You don't need an account to watch.

Okay. That already tells you something. So, you know, it's massive, I think, in different ways. First, it's like the quality of the performance that you've observed there, it's very good. You know, the amount of attention also that is on TikTok, it's amazing. Like, there's, you know, like all of these videos getting like millions and millions and millions of likes. So it's not, you know, like a marginal traffic. It's huge, you know.

And it's the first global social media because it's a Chinese. company. So it's the only social media that is popular in both China and the West, which makes it quite interesting as a global phenomenon, you know, and it's not based on peer communication. It's actually broadcasting. So in TikTok, you have like two channels, the for you, which is like a TV feed in which you just scroll, and the following, which is the people that you're following, and you have a feed of who you follow. And that's it. You know, you can search. The search is horrible. It doesn't work.

It doesn't provide good results. And it's very passive like television. So they reinvented television for the phone era, you know, in a zapping world in which like content is 15, 30 seconds long. And, you know, it's those 15 and 30 seconds, there's a lot of good quality content at that range of size. I mean, Twitter and podcasts are as cutting edges I'm going to get, I think, technologically. But they keep coming along very quickly.

So, yeah. And so I guess the point is that how we think about ourselves changes when these media change, right? How we remember ourselves. Indeed. What I want to say about TikTok that I like, which is interesting, I do think that it's more social and more family. So when I use Twitter, I use it by myself, when I use Facebook, it by myself. When I watch TikTok, I'm in the couch with my wife and my daughter. And the three of us are looking at the same screen at the same time and talking about what we're watching.

And I do think that that's important because we do have a lot of interactions right now in which people are interacting with screens independently. And I grew up, probably you grew up with people sitting in front of a television and having that more of as a collective experience, you know, in which you have to negotiate what you watch. You have to talk about what you watch. And I do find that, you know, that's a good thing compared to like the social media that has been more peer to peer but isolating. Well, I like to end on optimistic notes. Is that your optimistic note? What do you think about the future? What should our optimism be? Where should it be located? I'm an optimist, you know, and I don't know if that's a good thing or a bad thing. And I think maybe genetics.

But I tend to be optimistic even though, you know, like today, you know, Chile is going through a very difficult moment with all of the things that have happened there during the last four or five days. But I do think that at the end of the day, you know, the positive things in life tend to add up, you know, and build on other better than the negative things in life. So going back to the beginning, we talked about where information and order grows, and that's because our ability to create order and complexity is larger than the rate at which is getting destroyed. And I think that's also true for many of the positive things, you know. So I am an optimist because I do think that at the end of the day, you know, things add up, you know, better, you know, as we move along. All right, Cesar Hidalgo. thanks for some good information there.

Thanks for being on the podcast. Hey, my pleasure.