César Hidalgo & Andrew McAfee – Socioeconomics of Disruptive Tech #002
Koen Smeets: Welcome back to the interview series on the socioeconomic consequences of disruptive technologies by Rethinking Economics NL. Today, we'll be focusing on how disruptive technologies are influencing not only the economy but also our society. And for that I’m honoured to be able to introduce two world-class experts that have had an enormous influence on me. Firstly with us today is Andrew McAfee, who is the co-founder and co-director of the MIT Initiative on the Digital Economy and a principal research scientist at the MIT Sloan School of management. He studies how disruptive technologies, digital technologies,
are changing the world. Related to this, he has co-authored "The Race Against the Machine", "The Second Machine Age" and "Machine | Platform | Crowd", beautiful books I have to say, with Erik Brynjolfsson. And together with Erik Brynjolfsson, he is also the only two people in the world who's both on the Thinkers50 list, for the world's top management thinkers, and the Politico 50 list, or group, sorry, of people transforming American politics. His most recent book is "More From Less:
How We Finally Learned to Prosper Using Fewer Resources - and What Happens Next", which was published by Scribner in fall of 2019. Secondly with me today is César Hidalgo, who is a Chilean-Spanish-American scholar known for his contributions in economic complexity, data visualization, and applied artificial intelligence. He currently leads the Center for Collective Learning at the Artificial and Natural Intelligence Institute, ANITI, of the University of Toulouse. He is also an Honorary Professor
at the University of Manchester and a Visiting Professor at Harvard School of Engineering and Applied Sciences. Between 2010 and 2019, professor Hidalgo led MIT's collective learning group and prior to working at MIT, he was a research fellow at Harvard's Kennedy School of Government. Professor Hidalgo is also the founder of Datawheel, an award-winning company specialising in the creation and distribution, sorry creation of data distribution and visualization systems. He's also the author of three books, "Why Information Grows", "The Atlas of Economic Complexity" and "How Humans Judge Machines", the latter available both recently as a hardcover and for free online.
The first question today is for Andy, I was hoping you could tell us more about your thinking on disruptive technologies in the economy. I was especially curious how this changed over the years. So could you tell us more about, in 2011 you could you co-authored "The Race Against The Machine" and in 2016 "The Second Machine Age". So what is your thinking in these books and how has it changed over this period? Andrew McAfee: Erik and I wrote this little book, it was a pamphlet really, in 2011. We self-published it on Amazon, it was called like you said "Race Against the Machine" and we wrote it because we had this, at the time just kind of a vague idea that we were living through one of these periods of very deep economic change, very deep societal change, brought on by a surge of technological progress. And these don't happen very often, the profound
ones don't happen very often. I think everybody knows, one of those led to the Industrial Revolution, when we finally unlocked all of the energy in the world's fossil fuels, that really did put humanity on a different trajectory. That trajectory kind of continued about a century ago, with electrification and the internal combustion engine and another set of technologies that really did change the way that we live. And Erik and I started to come to the conclusion that we were witnessing another one of those surges of technology, this time brought on by digital technology. And if the previous generations of technology let us massively increase the power of our muscles and escape the limitations of muscle power, we think what's going on now, in the second machine age, is that we're overcoming the limitations of our mental power. Which is, we're able to harness so much more computing capability to help us do tough mental things.
And I think that basic thesis has held up pretty well, and I’m not even surprised anymore by being surprised. In other words, when I see a headline about some crazy, weird breakthrough in AI, or a drone that gets smaller and smaller with a better and better camera on it and can do crazy things, I’m not, I'm super impressed by this specific thing, I’m less impressed by the general trend, because I do think that we are in an age of astonishment because of technological progress, and I think those astonishments are going to keep happening to us. Koen Smeets: And how has it changed with "Machine | Platform | Crowd", which you released two years later, and I think that there are a lot of themes that are the same but what did you do different there, and could you also expand a bit more on how this is changing the economy? Andrew McAfee: Erik and I wrote "Machine | Platform | Crowd" because we had this really interesting experience after we wrote a book called "The Second Machine Age", which came out in 2014. We found ourselves having a lot of very interesting conversations with people who run large organizations, and they kept on saying to us some version of, I believe what the story you're telling me, now what do I need to do differently, how do I think about running my company, running my organization differently? And Erik and I said, look we're business school guys this is what we're supposed to do. So "Machine | Platform | Crowd" was kind of an applied book, where we tried to bring together some concepts from computer science, some concepts from economics and say, this is the nature of the change that is coming. If you want to successfully navigate that change, here are some things that you should keep in mind.
Koen Smeets: And professor Hidalgo, how does these ideas of Andy relate to your work, and I was especially curious about that in the context of "How Human Judge Machines", your newest book. César Hidalgo: So, I think, first of all I agree with what Andy has said and has produced in his work. I agree that we're in an age of transformation, you know, but I think my work has been a little bit different in that, even though that transformation is there and I do believe that technology is an important agent of change, geographically, when we look at it, it is very uneven where it happens and there is that question therefore, that if technology is this agent of change that is transforming society, and if knowledge in principle is something that can be copied, infinitely or, why you know, is it so unevenly distributed, and why you know, are these transformative events concentrated in space and they spread so slowly? So, a lot of my work has actually focused on, like the social implications of technology as well, you know, how technology or knowledge transforms society, but also trying to understand why it happens in some places and not others, you know. And the work on economic complexity in part, is a way to kind of like acknowledge that reality by using some math that helps you take the idea of a production function in economics, and say well you know, instead of thinking that capital and labour are the factors, can we learn the factors directly from the spatial distribution of data, and if we do that, can those learned factors tell us a better story about, you know, why there is geographic unevenness in economic outcomes? And that's a little bit of what we did then. But more recently, I’ve been now focusing on work on also trying to understand how this technology is being perceived, because through the work that I did in economic complexity first, I engaged in the development of tools and I created a company, you know, that creates data distribution systems and I started to go through a different path, you know, and an academic always likes to comment on the world, you know, but building stuff it's a very different story. And as I started to build stuff,
I started also to worry about how people engage with the things that you build, and I started to realize that if people reject the things that you build, not because they're bad or not because they don't work, but sometimes because they have a different perception of reality or they're coming from a different place, you're gonna fail anyway. So I became interested in how people perceive technology, what shapes those perceptions, and that started to happen and also at a time in which AI was growing and the backlash against AI started to grow, and when that happened I saw, well, there is a huge gap here because nobody is asking themselves the question, am I even judging this technology fairly? You know, are we just kind of like bashing on it because it made a mistake, or are we bashing on it even when it makes less mistakes than human? And I thought that there was kind of like an intuition that that might be happening, but there was no data that we could use to grasp onto that, so that's why we embarked in this, like, it was almost like a four-year journey of collecting data, it's more than 80 controlled experiments that we run to write this book, and to try to understand you know, how people perceive technology, with I hope to understand, you know how they would accept it, how they would adopt it and then also what these lessons about how we perceive technology tell us about how we perceive larger systems. You know, whether it is, you know machines that are made of people, like bureaucracies, or you know machines that are made of parts, like the ones that we usually tend to think of like machines.
Andrew McAfee: And I want to just highlight that César is homing in on, I think the two most important questions that we need to get smarter about. The first one is exactly what he was just talking about, how do people judge machines, technologies, the systems, the tools that are presented to them? Because like we see that judging does not always go as smoothly as we would like. I look at vaccine-resistance in America and many countries around the world, and if we don't convince people of the factually correct things, these are incredibly important, these are incredibly safe, these are huge lifesavers, if we don't learn how to do that, we have a dire, dire problem. If we don't learn how to present powerful technologies like AI, in a way that works for the people who will be using AI, be subject to AI’s decisions, then we have a problem again. We can get to some very, very uncomfortable,
unpleasant places as societies, if we don't do a good job of presenting these tools, even if they're amazingly beneficial tools, in a way that gets them accepted, it's a deep challenge. And the other one that he talked about, which I couldn't agree more with, is that, you know technology is incredibly widely available, the places in the world that are harnessing that technology, that are creating value from it, that are creating the companies that are so profoundly important in the world today, those are not even distributed, those are very, very, very spiky. As I look at the evidence, they're getting spikier instead of more distributed over time, and there are reasons to worry about that. And so we better get insight on these two questions. Koen Smeets: César, how do you see that, and I was especially curious about that from the complexity economics perspective? César Hidalgo: Well you know, so on the one hand, you know like knowledge is the thing that is hard to copy. You know, and and we've learned that, like knowledge is heavy, you know.
And it moves as people move and they form social networks and those are very slow processes. You know I’ve just moved to Europe, I’m starting to get to know people, maybe in five or six or ten years I’m gonna have an impact here, and it might be small and just one person, you know moving to a new place, you know. So that's one thing, but on the other hand, you know what we have is that technologies, in some way have this property that is that they're labour-saving, you know that's a little bit of kind of like the basics of technology, yeah. Like why do you want technology, because you can jump faster, you can lift things that are heavier, you know and more and more it's about saving time. You know it's about kind of like being able to do more with less, as Andy puts in the title of of his book.
Now when that happens you know, because technologies are scalable, it means also that even though they're a source of wealth, they must be also a source of inequality, you know. Because they allow few people to serve many, and if we take a more anthropological definition of money, money is just like a unit of favours that are exchanged. So how many favours do I owe you or you owe me, you know that's how much money like you save or you have, and in that context some people can make many people a favour. So Google is making many people the favour of like, hey you want to find something on the internet, come to my little box you can do it, I do that favour for you and then there's tons of people that are paying to show their stuff on that box to them. And they have a super scalable business model that would have been impossible in the era of the newspapers, or that would have been impossible in other communication technologies. So that also puts us kind of like in this very awkward position in which knowledge is growing, the ability to accumulate knowledge in an applied setting, like to create products you know it's very spiky, it concentrates you know, not only because of knowledge-diffusion problems, also because of economic problems you know. It's hard to compete also, even if you can
figure out how to do it, you know it's going to be very hard to kind of like get market share and grow and keep on hiring people and convincing advertisers. Let's say you got a search engine to advertise with you know, that goal not with google that is already dominant in the market. And as that happens, you start kind of like getting you know like this forces that push you know inequality, and I think that brings us to the other questions. Which is, people start having very different outcomes, they don't see technology as this thing that is here to kind of like save time, save effort, and make the world richer. They see something that a few people get to control and that benefits them, and I think there's some reason to see that, and in that context also they start to become a sceptic about it because their place in the world is not that great you know. They're in a situation that is difficult, and in that context I think we we get all of this scepticism, you know whether it is that, you know is Bill Gates is trying to put a chip to the vaccine and all of this idiocy. I think they come
through places of discomfort, that at the same time are created by these technologies that are here to make the world better, but for which is very different, difficult to spread those benefits you know, in a way that at least would would be considered fair by many people you know. And I think that's kind of like the tension, you know that technology itself creates the inequality that limits its own adoption, you know that creates a society that wants to reject it, you know. And that's sort of like a very interesting you know vicious circle that we have here. Andrew McAfee: Now this is really interesting, because I get to disagree with César a little bit here. César: Okay. Andrew: I agree with a lot of, César, what you just said,
and there are dimensions of inequality that are clearly increasing in many parts of the world and they're clearly really, really important. You and I would both agree, wealth and income inequality is something we should keep our eyes on and it's a challenge. There are other really important dimensions of inequality that are collapsing around the world, and where the fortunate and the people at the base of the pyramid are actually coming closer together, in a way that we've never seen before. My favourite examples of this are health inequality. So if
we look at some of the things that we care about most around the world with health, if you look at child mortality, or maternal mortality, or access to drinking water, we have never ever seen a greater convergence of those outcomes than we are seeing right now around the world. I graph some of this in "More From Less", and I have trouble thinking about any kind of any global inequality that's more important than health inequality, and that inequality is collapsing around the world. Going along with that, this unease with technology that César was doing such a great job of describing, I think this is largely a rich-world phenomenon. I think it is largely a phenomenon of
people who used to be in the rich, the middle class in the rich world, who are correctly seeing some real challenges to their livelihoods. I think the elite conversation in the rich world, that César and I are both lucky enough to be part of, in general that is a sharply negative conversation. If you go ask people, I would say almost anywhere in the bottom half of the global income range, what do you think of technology? They will say, it's fantastic, please we want more as quickly as we can possibly get it. So I don't want these discussions about the negative consequences are important, I think on a global basis they're actually kind of narrow. César Hidalgo: Yeah so, let me agree and disagree a little bit too. So I agree
you know, that there are important forms of inequality that are being reduced, and the ability that we've had to generate you know growth and wealth to get people out of poverty over the last 23 years is remarkable. So I agree with with you, and with Max Roser, and with Steven Pinker, and with Hans Rosling and I’m a big believer also, like part of my research has focused so much on growth, because I also believe that it's not just about distribution, it's about that economic growth that comes from technology, that lifts people up. But at the same time, I engage a little bit with the psychology of the phenomena. And I see, you know both in rich and in poor countries you know, like just to give you an example of like a rich country example of you know people being narrow-minded. A very good friend and colleague of mine you know, a few days ago actually you know, comes with it to me when I was in the U.S. packing, with with a box that he wanted me to bring to Europe
you know for someone you know. And I [said], put it on the mail! Well this person, and mid-30s, American, very towny type, didn't know that USPS sent packages outside the US, because it was, it is USPS it's, for the US you know. So you can have kind of like that narrow-mindedness, depending on how much you've seen of the world. And I think, the what the point I’m trying to make there, is that the perception that people is going to have, depends on the opportunities that they have.
So maybe, 30-40 years ago, I would agree with you that the condition for many people were worse, you know but the ability to be aware of what was going on in different parts was also diminished. And as those conditions improved, for instance the social phenomena that we've seen in Chile, a lot you know like recently and the social profit and the rupture, it could be told as you know like people in extreme poverty, but in reality a big part of the social phenomena was people that were first generations, that they went to the university, they told them this idea, that well if you work hard, you educate yourself, you are not going to be like the previous generation, and they find themselves you know, like graduating from college, having jobs that don't pay better or in comparison and a very difficult you know life with a lot of competition of people in a similar situation, but they're much more aware of you know the world. They're much more educated, they have ability to understand things, so I do think that those two things in a way counter-balance each other because as the world gets better, you know but also as our perceptions get wider, you're able to kind of like see things in which you're gonna perceive forms of inequality that before were invisible to you, because you didn't leave the town, you didn't leave the village you didn't know that USPS, you know mailed abroad. So I agree with you on the fact you know, but I do think that perception has a component in which you know that runs in the opposite direction, as people become more aware you know, even in in the less developed countries, they also start seeing kind of you know, the systemic difference that sometimes they were invisible to them, and it's a reality that is more complex and difficult to accept sometimes.
Koen Smeets: I was also very curious for you Andy, basically on how you see that, but also in relation to some other trends and basically [inaudible]. So for instance Piketty's work, of course, basically on the increase in income-inequality, but also the decoupling that we're seeing of productivity and wage-growth since the 1980s, how are you seeing that then in the context of these technologies? Andrew McAfee: There's a huge amount of debate among the smartest economists that I follow and that I try to learn from, about exactly how pronounced those phenomena are. Everybody who I follow, agrees that inequality has been rising in most countries you know, America is a clear example. Most people agree that there has been a divergence between productivity and pay, at least in America for some kinds of workers, and it's pretty clear that more of the income is going up to the top, I forget if it's the top, you know 10 or 20 or 25 per cent of earners, but in general there's kind of a group that's doing fine and pulling away, and another group that's holding steady or in some cases even falling behind. These are absolutely things that we
need to worry about, I’m even more worried about what César talked about first, the geographic distribution of our economy is changing, it's changing in the direction of concentration, and I have trouble telling a happy story, where a bigger and bigger percentage of the GDP of America or the Netherlands comes from a smaller and smaller number of hectares. And that there's a problem with that, for the very simple reason that there are people living on those pieces of land that are getting left behind, that are not economically viable anymore. And César and I both know that as a rule, economists are not, they don't lack confidence. They think they have a very good tool kit to analyse problems, they think they have a good tool kit to address problems. When it comes to this problem of increasing geographic inequality in economic activity, most of the honest economists I know say, our toolkit, the things that we know work, our toolkit is not very good for addressing that problem. And figuring out what to do about that and
spreading back out the economic growth, there's a lot of work, there's a lot of activity, and while all that's going, on our economy just gets spikier and spikier. César Hidalgo: So let me add something to that you know, so I’ve been advocating for the last 15 years and working on trying to get economies that are less developed to go into knowledge-intense sectors, you know. The product-space is like, well what's the best way to climb that complexity ladder. Economic complexities are like, well you know what is the knowledge-intensity of your economy and how you can measure it, and so forth. But during that time also you know, I became an entrepreneur, I have a company with 26 people and so forth, and I must admit that, a lot of the lessons that I learned about you know business and economics, I learned it by being a CEO and a manager you know. And this is a lesson that I’ve been thinking a lot about recently,
which is I think very important. And this lesson is about the fact that knowledge-intense economies can be very cruel. Let me tell you what I mean by that. So let's say you have an relatively not knowledge-intense sector, like a fruit-packing plant, okay. So there's a conveyor belt, there are boxes, there's peaches, the peaches need to be put into boxes, they need to be checked and so forth, and you have many people that are putting peaches in boxes. Now, what's the marginal cost of training someone to become part of that job, to including someone? So think of it now from the point of inclusion. So someone comes in, or there is a surge in demand,
and you need to bring someone else into the line and sure, there might be some protocols of safety and hygiene that need to be you know performed, but you can probably have a person you know working on that assembly line, maybe within a week, maybe within a day, you know. Now, I run a software company, in which we have a stack of software that we've developed over eight years, that is extremely you know complex, and it has own flavour of language. When we hire people, it takes us three to six months for these people to be able to commit code, it always has to be reviewed by others, and the problem that I face as a manager, is that they're for example guys that are good developers, but that they hate teaching others, and when the new peoples come in, they say hey this guy is bothering me, I don't want to explain him how this stuff work, I’m getting frustrated because I have to correct the work, and a big source of stress and tension in the company, is that we need to hire people, but the ones that are in they don't want to spend time teaching them, and if nobody teaches them, even though these guys come with a master so with a six-year degree, and they know how to program, they're not gonna be able to be productive. So the knowledge intense sectors, they have that problem in which it's much costlier to add someone to an operation you know. And it becomes very cruel because you can have operations that are very scalable you know, Instagram you know had you know like 13-14 people at the moment that they were already like a large company serving you know, like millions and millions of users. But because of that, I find also a lot of friends of mine from college you know, that went to college with me and you know they were not as lucky as I was, is struggling to join the labour-market, even though they're highly educated and highly good, because the cut-off is really high. Because in this sector, when you're having someone let's say that you want
to be in charge of kind of like this database or the front-end of this platform or whatever, it is not as easy to integrate someone that has put in peaches in boxes on the conveyor belt. So what I think is happening right now, is that we're moving to these knowledge intense sectors, to these cultural activities you know, but this cultural activities have inherently this problem of onboarding. This extremely costly on boarding. And that also you know generates another form of you know like rejection, or displacement, or lack of inclusion that we need to worry about, because I think it's a structure. The more that we move to a knowledge intense sector, the higher the cost of onboarding, the higher the cost of onboarding the more that you know onboarding causes the stress and that people that are looking for a place might not be included you know. So
in other words you know, it seems that there's some good things about knowledge-intense sectors, but there are some things that for example we might have had in manufacturing you know, in terms of inclusion that are really hard to replicate in this more knowledge-intense economy. Koen Smeets: Andy, especially considering your work on these type of enterprises, how do you see this? Andrew: Which aspect of it well? Koen: Especially, basically, on like digital, disruptive companies that are really working at this high technology, high intensity knowledge sector. Andrew McAfee: I completely agree with what César just said. For several decades after the end of World War 2, our economies had a bottomless thirst for relative, for unskilled labour. And when I say unskilled, what I basically mean is not formally educated beyond a high level. And those jobs, because they were relatively high productivity jobs on assembly lines and things like that, they could provide an unskilled person with a medium, with a middle class income, and the hope that their kids could become more educated, get higher skills, climb up the ladder that way. Those jobs,
they still exist but they're not as plentiful anymore, because of these two huge forces of tech progress and globalization. So the middle class in the rich world is absolutely feeling squeezed and has been left out of a lot of the economic gains that have happened up the education ladder and even somewhat down the education ladder. And we should not pretend that that's not the case. Koen Smeets: And to ask you a quick follow-up on that, how do you think that we can create a more inclusive society so to say, where we can, this technology that we're describing, that it's to some extent creating this inequality, can actually help with alleviating this inequality? Andrew McAfee: The first person that won the Nobel prize in economics was a countryman of yours, his name was Jan Tinbergen. And one of his great quotes, was that inequality is a race between technology and education. Technology like we've been talking about tends to increase inequality, all other things being equal. The great leveller is to give people more skills. And over time we
have done, you know the average skill level has gone up in most countries around the world. If that is not working really well, my first solution is to re-examine how we're educating people and do better at it. One line that I like to use all the time, is that we're doing a great job of educating the kind of workers we needed 75 years ago. Okay, well let's stop doing that. It's a more knowledge intensive economy, we need very different skills, we have some idea what those skills are. In addition, we're seeing a lot of innovation in educational approaches. We, César, you know about things like 42, the school in France, like Lambda, where you share someone's income after they graduate with a with technical skills. I love
these ideas. I don't know if all of them are going to work, but by all means let's innovate on how we actually deliver skills to people. Can we do it more quickly, can we do it from a lower base, do we have to go through the classic educational path that we built up in the rich world in the post-war decades? Well no, we don't, let's go figure out what works and do more of that. Koen Smeets: César, how are you seeing these things, and I was also curious on, especially for economic students, what do you think, what are the type of skills that they should get, and especially in the context of the topics we're discussing today, like disruptive technologies? César Hidalgo: Yeah, so let me address the first part of the question you know. Which is, okay so we have like these forces that are pushing inequality, like what to do about it? And I agree with Andy that, you know education is important and should be promoted, but I do think that there are forces here at play that are extremely strong and in some way, you know invite us to think more creatively in in two directions. The first one is that, during the last you know century and especially during the last 30-40 years, we have been able to create solutions, and business, and technologies that are so scalable, that they can serve billions with teams that are relatively small, that in that world you know expecting that everyone can be included, you know uh in the production side of the economy, you know might be naïve, but at the same time there might be a lot of value of keeping people engaged on the consumption side of the economy. So if you have a homeless person in the US, which is a huge problem in many parts,
I would argue that giving that person a UBI, and having them be a consumer in the economy, is better than having that person have no consumption power you know, and be you know more of a burden on the system, be a person that kind of like gets disease, that you know it needs to be visited by social workers, you know that you know is in bad conditions and so forth. So I see more and more, that as society progress we might have to not achieve equality through some sort of distribution system, I’m very capitalistic in many ways and I do believe that giving people the ability to, you know own the fruits of what they produce and reinvest them, it's a really important freedom that needs to be protected. But I do think that you need to in some way ensure parts of society you know from the ultimate ruin you know. Because the problem is that, that's in a society that you don't have you know uh like good ways to stop that ruin from happening, that becomes an attractor. It's not a place that people bounce off from you know. Once you have like no friends, and you're in the street nobody, wants to even talk to
you you know. Like you're you're smelly, and you know and you look like someone that's scary. So you need to protect people from that ruin, so I do believe you know that things like UBI and so forth are gonna have to be part of the solution you know, and that the economy lives on the excess, it lives on top of that you know. Because nobody wants to be there, and the one that wants to be there you know, well let them be there, but let them be there with some dignity you know. And I think that's kind of like an important thing we're gonna, like look at this century thing, that is going to become more and more a discussion, and I think we're going to see that happening when we exit COVID. Because when COVID happened, you know the economy was hit but
it was hit very differentially. Like I’m sure that Andy has done great, I’ve been doing you know well you know, but there's you know I live here next to a Place, Placa De George, full of restaurants that all closed except two you know, and there's a lot of people that are having a very different reality right now, but we're not talking to them. And at some point, the COVID you know, confinement hopefully is gonna end, people are gonna start mixing again and we're gonna realize, that the wedge between society grew even further, and there's a lot of people that had a very different experience of what 2020 and 2021 meant to them. And that pressure, I think is gonna affect politics in a way that it cannot do right now, because people cannot congregate, because people cannot like you know, express themselves in the in the way that they can, in a non-confined state.
So that's a little bit how I see that, I do see it as a more, you know progressive form of Keynesianism that we're gonna need to adopt, you know I do think that there is a world in which our supply side can be so strong, that the information that people reveal through the acts of consumptions you know, it's enough to justify a minimum level of consumption that you want to give them. Because when people consume they inform you about what needs to be produced, what it's like and so forth, well you know that is something that I do think we're going to get to a point in which we're going to be so rich that the value of knowing what needs to be done or who likes what, is going to be enough to justify a very basic income. Koen Smeets: So I’d love to come back as well at the education aspect, but first, Andy how do you see this? Andrew McAfee: A bit differently than César does. And I steal here a line from Erik, from Brynjolfsson, from my co-author and he always says, does anyone think right now that there's a shortage of work that needs to be done? And to follow up on that, does anyone seriously think, that there's a shortage of work that can only be done by human beings, where we don't have robots and AI that can do the work? And when I hear Erik phrase it that way, it becomes incredibly clear to me, the problem that we face is not a shortage of useful, societally beneficial work for human beings to do. We're in a very strange period right now because of the pandemic,
where we literally cannot be close to each other, because there's this deadly disease going around. Great we're going to get past, that let's all assume that. The vaccines appear to work spectacularly well. So we're going to get back to a more normal economy. In that normal economy, is there any shortage of work that needs to be done? I think about the energy-transition that we have to get through and we have to get through very quickly, and I think about the jobs that would come if we were intelligent about that. I think of the poor job that we do in America of taking care of some of our more vulnerable populations and the work that could come out of that. So I am not
worried about not having enough work to go around and therefore needing a universal basic income. I’m worried about creating the right conditions, so that people can get matched to the jobs and the work that needs to be done, and be part of that world of work, which I believe is actually incredibly important for a person's dignity and sense of community and sense of belonging. Koen Smeets: Yeah I see a quick follow-up on that, do you think that everyone is able to do those jobs? Because I can imagine that, although we're seeing that far more people are getting university education and such, I can imagine that AI is also to some extent creating more upper level and lower level jobs, so do you think that that's an issue or that it's not an issue at all? Andrew McAfee: It's an issue, but think about jobs again that AI and that robots are nowhere near doing, nowhere near. Take caring for a room of 12 children. You think an AI is going
to do that anytime soon without terrifying the children? Absolutely not. Fixing a bridge that's in need of repair? I assure you, we have no robots on anywhere near deployment that can do that. Installing a wind farm or a solar farm, we simply don't have technologies that can do that. This is why I say that there's no shortage of work that needs to be done, and maybe in the
crazy distant sci-fi future, maybe you will live to see an economy that is incredibly productive and just doesn't need a lot of human labour as an input. I think that's plausible at some point in the future, that is not where we are now, it's not where we're going to be in five or ten years. César Hidalgo: I would agree that there's a lot of work that needs to be done you know maybe for three billion people, I don't know if there's gonna be enough for seven.
So I do think that there is kind of like, I don't disagree that there's gonna be a lot of work that needs to be done and that cannot be automated, I also see that there's a lot of people and that at the bottom you know getting hired, you know it's really hard even at an okay wage with a high level of education. What I see right now, like shuffling even like for example in the software sector that lends itself to telecommuting, it's you know like a like a big flight towards you know like middle and lower income countries, that have you know good educational systems but you can get an engineer for 2k a month instead of you know 150k a year. And even in that space as people get educated, like the salaries you know like are quite depressed and it's easy to find people you know. So like at least the problem that I have, is I have way more people that want to work with me, or that want to do something, than I can get to hire you know, given the resources that are flowing through me at the moment.
And I see kind of like that scramble you know. And so I agree that there's a lot of work to be done. I do think that there might be a differential, and in that differential you know is where frustration is going to grow, you know. It's because, it's not people that didn't try, or didn't want to, or didn't work hard. Like frustration comes from those that try,
and they didn't get the opportunities. I’m afraid that, at least my intuition is that there's going to be an important difference there, like a big enough gap to become you know a problem. Koen Smeets: And to return also to something we were discussing earlier, the education aspect. So it seems that many of the topics we today are in the economics,
we discussed today are in the economics education, but quite a few aren't. So how do you think that we should include these aspects, for instance complexity economics. César, how do you see that, should we include that in the economic education and what ways? César Hidalgo: Economics education or are you talking about education in society more in general Koen? Koen Smeets: I think it's most interesting to first focus on the education specifically for economists. César Hidalgo: Education for economists. So I do think that you know, I make some like
general distinctions to kind of like, help place different fields of thought within a context. On the one hand you know, there is a big difference to me between the social science you know and the natural sciences in the objects of study, which is the following. You can go all of your life without knowing you know how quarks work you know, and knowing nothing about nuclear physics you know. And most people can go all of their lives you know, without engaging with any of that.
And natural sciences has a lot of that you know, when you look at cell biology, genetics you know, astrophysics you know, but in the social sciences we deal with things that people that are professionally trained or not have to deal with. Everybody participates of the economy and they see that their peers that are better than worse, whether they have an understanding based on models, or in empirically verified facts, or just the stories of their grandpa, is a different story but you know, their fields you know that in some way participate in things that everybody can only has an opinion on you know, while nobody has, or very few people have an opinion on quarks, a lot of people have an opinion on the economy, for sure and the other field is ethics you know. Whether you're a professional philosopher trained on ethics or you're a normal person, you have been asking yourself questions about is it right or wrong to do this all your life you know, and you know there are intuitions that are formed around it. So in that context, the fields
that deal you know in these spaces in which the non-experts are always going to have an opinion, I think they're in an uncomfortable position of course, because they're always going to kind of like get non-expert input, but at the same time they have kind of like a responsibility to engage with these different, you know schools of thoughts and ideas of other people that have been thinking about that from a different perspective, you know. And you think about it like the social sciences, it's all built kind of like on verticals of different models of what matters about the world you know. Like economics and political sciences we think, well, in economics we try to think about the world as a collection of rational agents. In political science we try to think about the world as battles of power between individuals, but you can all have micro, meso, and macro in both of them you know. Because you kind of like subscribe to maybe different ways to think about human behaviour and collective behaviour, but the object of study has a big overlap and has you know inputs that are shared. So
that's why there's also kind of like all these this cross, there's people that go from economics to talk about institutions, and there's people that come from political science to say how like institutions are affecting the economy and whatnot. So, I think in that context, it is important to to stop the identity academics that we now suffer. In which there is this shared subject, which is society, there are all of these approaches to look at it from academia, and there's all these people also outside academia that in some way they have a valid lived experience from where there can be insight, there can be wisdom that we can draw on you know, and how can we, you know make sure that we continue advancing knowledge in the best possible way, without using our quest for the search of knowledge as a signalling mechanism, to try to tell to others to which academic community we belong or identify with. And I think that to me is kind of like the big challenge. Because in the social sciences, much more of that kind of like positioning, that within a larger topic you know people kind of like tried to sell that they belong to a community, while in the natural sciences maybe because you know we're in a topic that, hey if we're not studying it, nobody else would, you know nobody's thinking about this stuff on their day to day. It's like stuff that had to be discovered by like you know
smashing atoms together sometimes you know. Then there is kind of like a more of a fixation on the object of study no matter the perspective, than on the perspective you know itself. Koen Smeets: And for Andy, basically I was very curious on your take on this, but also especially, basically what I think it, also is that, basically the different disciplines have to combine it and such that, a lot of economists will be not that familiar with disruptive technologies. And they will have, you'll get like a basic understanding in your education but it will not focus on it, like really what will be the effects basically, what are the AI scientists saying and such, so what what should they know about that and how do you see it in the education? Andrew McAfee: I’m actually not a formerly educated economist so I’m not going to opine on the education over there. What I can say though, is the economics papers that I read I find absolutely fascinating. And the methods in the discipline have changed just in the time that I’ve been following the literature.
The empirical toolkit has gotten a lot bigger to investigate questions, the scope of the discipline has increased a lot, and I read a lot of you know fascinating stuff about the economics of crime or about algorithmic bias and fairness. These are really important questions and I think the discipline overall is doing a great, an admirable job of tackling them. Koen Smeets: César, how are you seeing this? César: Say again? Koen: César, how are you seeing this? César Hidalgo: No yeah, I do think like in economics you find like amazing work and amazing people, so that's that's no doubt you know. You do find that a lot you know, but going back to kind of like your original question you know, if we think about you know also the disciplines not in in terms of kind of like the model of human nature that they subscribe to you know, economics whether we want it or not is kind of connected a little bit to money. And money, whether you want it or not, is a technology. So in some sense it is a field
that is defined by the study of a technology, if you look at it from that foundational concept, you know. So I do think that the study of technology and economics intertwine you know from the beginning, since money is a technology, and in some sense that's I think maybe why economics has been a successful field from which to study the effects of technology in society, at least you know when it when it relates to like productive activities. Koen Smeets: And what do you think would be most important for economists to know about the currently disruptive technologies as AI? Andrew McAfee: I think they need to learn how to use some of these methods themselves, some of the coolest papers that I see have combined more traditional techniques with machine learning to go poke at the things that economists are interested in. So you know economics is a really expanding discipline, the toolkit is getting bigger. That makes César's point more relevant. Training up to become a first-rate economist, this is
actually a long, really intense discipleship to get your arms around the tools. So it's not, it's going to take time and it's going to not be open to everybody. Now we can all benefit from knowing a little bit of basic economics. But professional economist is kind of daunting these days, I think.
Koen Smeets: Yeah I think that, further in this series we'll also have an interview with a group of basically computer scientists, that are really applying the methods of economics, of reverse game theory, to really find basically how to build markets and such. So that will, I think that basically listeners that are, viewers that are interested in this, will get a bit of an idea on that. César, how do you - César Hidalgo: I agree with Andy, and the way that I see this is that math is not science and that's a good thing. And the reason that that's a good thing, is that mathematical methods usually you know tend to play the field, they go around, you know. So the idea of, you know like dynamical equations for instance you know which you know, it was very popular you know in classical physics you know, and that's how you solve systems of dynamics and kinematics. Then you know was adopted in economics and models of economic growth, a couple differential equations and so forth, then you know statistics you know, made it into economics and I think over the last 20-30 years, this empirical revolution has been a statistic revolution you know, that has been very important.
But I do think that machine learning is something that's going to contribute to a change too, because it provides a lot of tools that are very interesting to look in particularly at macro problems. Like machine learning excels when there's a lot of data you know, and you're trying to understand macro phenomena, and you're trying to predict and look at trends and so forth. The work that I’ve done, one way that I communicated for example, if I’m talking with computer scientists, is like hey what I’ve been doing is applying machine learning method to understand the economy of the macroscale. The product space is a recommended system and economic complexity is dimensionality reduction applied you know to specialization matrices. These are
tools from machine learning, that when applied to economics they also work. And they should, because that's the nice thing about math you know, dynamical systems they work in physics, they work in chemistry, they work in ecology, and they work in economics. And by the same token, we should find machine learning tools to be useful in all of these domains. So I do think that
there is an opportunity now to expand you know, the way that we look at economics, or to answer some of the same problems from a different perspective, and maybe even discover some new problems by using this new toolbox, that machine learning and AI are bringing to the table. Koen Smeets: So considering time, we have about five minutes left, I think we have to move towards the closing question. And we do then have quite a few minutes for the answers there. And we always have the same one in this series, and that is, if there's one thing you could say to students economics watching today, what would that be? And I think it makes most sense to do it in the order of the first questions as well, so first for Andy? Andrew McAfee: Alright, go investigate real-world problems, right. We we have,
the discipline of economics has a really important way of looking at the world, a way of thinking through questions. The, you know, the mathematical, the statistical toolkit, the technological toolkit for investigating those questions is getting better, the data available are getting better and richer all the time. Go work on an actual question that somebody out there cares about, and spin up on all the latest methods, all the latest and greatest papers that have been written, to serve as a model for how you should go about doing that work.
Koen Smeets: And do you have any idea, any recommendations how to best find the real world problems to tackle? Andrew McAfee: No, I have no idea how to do that. Just, I’ve stumbled across mine by luck, it feels to me all the time. Koen Smeets: I think that's the same for me. César, what is for you basically the one thing, if you could say one thing to students in economics watching today, what would that be? César Hidalgo: I would say, be nice. And the reason for that is that,
if you want to engage in a creative environment, you have to let bad ideas grow, mix, mutate to the point that, sometimes bad ideas become good ideas. So, one thing that I did learn you know, like at a place like the Media Lab, that works in this intersection between technology and the arts you know, was that some of the most successful entrepreneurs that I met you know, and these were people that you know made it big, they made you know like video games that sold you know millions of copies and they were loaded. They were you know some of the least rational people that I met, and some of the most emotional people that I met. And they were kind of like, I want to do something that made people feel like this. And they would come up
with like a lot of weird, bad ideas but once in a while ,when they would get a good one, it would resonate with people and they would be able to execute it with that finesse you know, that artists have you know, that allows them to do things that enjoy wide adoption, acceptance, and they develop fans you know. At the end you know, like the consumers of a product, if they're fans you know it's fantastic you know. So that's usually what you want. And I realised I do kind of like think a little bit differently. I come from the natural sciences,
which are very analytical. I think that makes me similar to economists, we tend to be very critical you know, and try to sort of like find the holes on the problems and argue them you know. But I realised that a lot of what we like in the world and what we study when we study the economy, comes from a very different breed of people you know, that are creative because they live in this environment you know, that is emotional, they're nice to each other, and they let kind of like bad ideas have sex until they have good children, you know. And that's something that, I think is important for creativity and when you manage teams in particular. Because a lot of the times
you know, like you may cut the wings you know too early if you are too critical when things are not yet fully formed. And they can surprise you, if you let them grow to their next stage, so be nice. Koen Smeets: I think that basically the combination of to pursue your crazy ideas but to be kind to people, is a beautiful way to end today's interview. I want to thank both of you profoundly for your time, it was a beautiful interview and we look forward to having all the viewers again here next week. Thank you! Both: Thank you! César: And thanks Andrew, always a pleasure to chat with you. Andrew: Always good my friend, I’ve got to run, thanks for including me. Koen: Thank you, thank you! Andrew: Bye César, have fun there. Koen: Bye-bye!
2021-04-18 13:24
