Tyler Cowen: Identifying Talent and Measuring Organizational Capital

Welcome to World of Das, a show for data enthusiasts. I'm your host, Oren Hoffman, CEO of Safegraph. For more conversations, videos, and transcripts, visit safegraph.com/podcast. Hello, fellow data nerds. My guest today is Tyler Cowen. Tyler is professor of economics at George Mason University, uh, host of Conversations with Tyler podcast. He's a blogger at Marginal Revolution, and he's an author of a lot of books including one of my personal favorites, The Great Stagnation. Tyler, welcome to World of Das.

Oren, thank you for having me on, a real honor.

Oh, well, thank you. Now, in your Great Stagnation book, you, you, you kind of make a case that innovation has been decreasing over the last 50 years. What do you think is kind of like the core thing that's responsible for the slowdown in that innovation?

First, a qualifier. I think it's possible in the last 2 years we've emerged from the Great Stagnation, The vaccines, for instance, are very valuable. But put that aside. I think the core problem was that we had a new technology, which I call fossil fuels plus powerful machines, and we did everything possible with it. We built cars, we built planes, we electrified the world, right? That was incredible. But at some point, we're simply making the seat in your car a little more comfortable. And while that's nice, it's not a major breakthrough. So by the time 1973 rolls around, we're having a lot of marginal breakthroughs other than the tech sector, which you live in. But a lot of the other sectors were pretty stagnant for about 50 years. I would also say we have regulated business too much. And as a society, we've become more complacent and more risk-averse for cultural reasons.

And we have to innovate in the last years. We had to get the vaccine out quickly. We had to do some of these other things that were— we may have been coming out of this and seeing a little faster innovation cycle because we've been under threat before in the last 50 years for other kind of threats, but we didn't necessarily see that core innovation increase.

Well, I think threat is the most plausible hypothesis. As you know, a lot of innovation came out of World War II, the actual development of the computer and even the airplane and also nuclear power. So if you look at 9/11, that was a kind of threat, uh, developing the quality of drones, which later had commercial uses, partly came out of that. So it's too early to say, but I think both in the areas of green energy and biomedical innovation, not just the anti-COVID vaccines, but you look at malaria, you look at sickle cell anemia, we seem to be seeing a coordinated set of advances. Too soon to say for sure they're all going to work, but to my eye, it looks pretty promising.

Now, one theory that, you know, could be why innovation is slowing down is just we've just become way too comfortable. Certainly we're a lot more comfortable today than we were 50 years ago. Do you think there's something to do with that, that maybe we don't have this desire to innovate enough because we're just fat and happy?

I definitely agree with that, but I think we need to add qualifications. So people in the year 800 were not that comfortable, right? I'm not sure how much they innovated over the next 100 years. So it's the unquiet combined with opportunity good mix of talent, problems to be solved, crisis and emergency. You need a lot of things to come together to innovate. The natural state of affairs for mankind is poverty and stasis, right?

Now, the— I mean, even if you look at things like other innovations, let's say music or fashion, like, you know, in some ways it's very hard to distinguish today from some of— some music and fashion was in the 1980s. Like, why hasn't there been as much innovation even in some of those areas? You would think like, okay, we're not going to focus on new technology things, but at least we would innovate in the arts over time.

Well, if you look at music, say a 5-year period from 1963 to 1968, every year there's incredible innovation. You can listen to music and almost know what year it's from, not what decade. And since then, you've had sort of two new genres, electronica and rap, which in turn have relied on innovation. But that initial burst, it comes, I think, from the electrification of the guitar, and the Beatles and some other groups developing the recording studio as a thing you can use. And then you have people very quickly sort of filling out that space with all these permutations from those ideas. And then you have a period of stasis again with rap being its own thing, electronica being its own thing, the latter being driven by digitalization of music. And now we're in this period, there's a lot of good pop music, but you listen to it, it could be decades old, it would sound the same. Maybe they use fewer guitars. What kind of innovation is that, right? To me, it's sad. I think maybe more great songs in a year than 30 years ago, but way less innovation, and most of it's just not that interesting.

Yeah, when I listen to a group like Queen or something, it's like almost every song was innovating. Yeah, where'd they get this from?

Yeah, Bohemian Rhapsody, think of all the different influences in there, right? Like the culture of Zanzibar, Italian opera, British musical tradition, rock and roll, The Beatles, everything. Beach Boys, barbershop choir harmonies, all packed together in this thing that's its own.

The last few decades, we've seen a real acceleration in data in the economics field. Do you think this is kind of an overrated thing or an underrated thing?

Well, you might not like my answer, but I think it's become overrated. It's what people do, and they're doing it for a good reason, but most things that people do, they end up overrated. So my fear is we're producing these perfect quality papers,, but they end up siloed and people are losing somewhat the ability to construct and evaluate synthetic arguments of the kind, say, that might be required by Nate Silver or Zeynep or Alex Tabarrok, like to try to figure out what we should actually do with the policy. So the data work is great, phenomenally well-executed, but to my eye, it's become a little overrated.

Is that because to get a PhD and to get tenure, you have to get this super micro thing and get published in a super micro way and you just don't have the ability ability to think bigger until you're in your late 30s or something? Or, or what's the constraint?

You have to learn programming at a high level. You have to track down the dataset and get the rights to it. That's the very hardest thing to do. That itself is a full-time job. So in your spare time, are you pondering the more philosophical questions of microeconomics or economic policy? Typically not.

Are there datasets you wish were more accessible for economists so we can, so we can learn about the world a little bit more?

Well, here's two things I wish for in particular. If we had individual-specific markers of different kinds of talent above and beyond normal five-factor personality theory, and we could see how well they correlate with the success of those individuals, right? I would love to have that data set.

Oh, well, give me an example of like the— like what are the— what could we potentially learn from that?

What predicts an individual being good at a job, which is something you're very interested in. You have a wonderful tweetstorm on this. From some number of years ago. So the Five Factor Personality Theory is the best we can do, nothing against that. But the 5 factors like disagreeableness, you're either high in it or you're low in it. But the really successful people, I would say, are selectively disagreeable, right? Like they know when to turn it on and off. Or conscientiousness. Well, what is it you're conscientious about? It can be a negative if you feel, oh, I've got to stop working on this project for Oren to get home to walk my dog. Do you want to hire that person? They're highly conscientious to their dog, but you're like, oh, wait a minute. So if we had better data there, sector-specific, job-specific, coded by individual, I think we would learn enormous amounts. And when it comes to, say, finding new talent from Nigeria, South India, Pakistan, wherever, we could do much better, much more quickly. Another thing I would like to see us have data on is what Arnold Kling has called organizational capital. So my sense, and this is not confirmed, but during the pandemic, output has done pretty well. Companies have focused on getting the product out the door, keeping revenue coming, like making the next Zoom call, just not exploding. They've had to do that, right? Not a complaint per se, but long-term tasks like figuring out who is the 23-year-old in your organization who will someday become a leader, you don't have much face-to-face contact. You don't have the temperament or the time or the interaction to be able to figure that out. And we are eating into our corporate seed capital, I think, by focusing on getting jobs done now, making the next Zoom call, not enough on developing long-term capabilities and vision and talent. We don't know that. If we had a dataset on what I call, Arnold Tinkoff's organizational capital, I think we'd learn a lot about how businesses operate.

And how, how would you even be able to measure organizational capital?

It is hard, right? That's why we don't have it. But even a bad data set, if you had time logs of what people do and which interactions are considered to be productive, and you could sort of track this through time and see what ended up being important— was it company picnics, was it one-to-one lunches, you know, was it in fact the Zoom call— that's what I think we really need to know now. It didn't seem important 3 years ago, But you're asking for a data set now that's going to matter. I say it's that.

Well, in some ways, like, we're probably more likely to have that today because if we're doing everything online, we at least can track it better.

Correct.

Um, so we could track all the Zoom calls, we could track the email, we can track our Slack things that we're doing, etc. So, so at least we could— we have the ability to gather the data today that we probably didn't. It was probably much harder to do when we're like in-person interactions.

And we'll have some sense of like, well, those offsite, like, group interactions every 3 months, do those actually substitute for the office? I don't know. I think with time we will know.

I have this belief that every successful person has like a core superpower. And my belief is that your superpower is, is that you're really good at asking questions. And I've probably been in like 50 conversations with you over the years, and you just have like one of the highest hit ratios of asking the right questions to make the people around you think. And I think we'd have a better world if everyone got better at asking questions. So are there some tips you can bestow upon us mere mortals about how to ask good questions?

Your questions should be highly specific. The question itself should be detailed and intelligent and set a kind of standard for the answer, right? So the person feels they better match up to the question.

Uh, okay.

It's good to put before the question a kind of anecdote or confession to raise the stakes, like this really matters. And then make it clear to the person you are actually listening. And I think another input into asking good questions, if you're a high-status person, with the same question, you'll get better answers than if you're a low-status person. So try to be a high-status person. But that said, I don't think that's my superpower.

I mean, do you think when you're saying you're listening, do you think just a lot of people are asking questions that either they already know the answer to, or they don't really care what the answer? I mean, what do you mean by you have to be listening?

If the person you're asking senses you're not listening, you're asking as a formality. A lot of people doing job interviews, it's their job, they go through it, they're tired, they're not listening, they get automatic answers that maybe are fine. I really care about the benefits at any company I work for, right? Fine, but not informative. If you show you're really truly listening, engaged, caring, interested, you will get much better answers.

Tyler Cowen: Interesting. I think you're also really good about asking questions in areas like you don't have expertise in. Obviously, you know a lot about economics and you can ask great questions there, but I've been in lots of different conversations with you. I remember one about— we were talking about neuroscience and you were able to get into these questions where you didn't even— you probably had some cursory knowledge, but you weren't nearly anywhere at the expert level of most of the other people in the room? It's hard enough to ask good questions in a field you know a lot about, like economics. How do you ask good questions in a field you're more of a spectator in?

I would say, first of all, my true superpower, in my view, is that I can read 5 to 10 times faster than a person of comparable intelligence. I just know more things.

Because you're more well-read, essentially.

Right. It doesn't fully address your question. But I would say this, the answer to most questions in my view is training, just train. So I run a podcast, Conversations with Tyler. I do one every 2 weeks. Everyone I speak to is super smart, mostly way smarter than I am, more accomplished, wealthier, whatever, some combination of all those. So just train, train, train, train, train. I'm 59 now. I started in on all this stuff like when I was 13, 14. So that's like 45 years of training, and I never let up. It's totally relentless.

And you think it's just compounding at 10%, 15% a year every year, and you're just—

Exactly.

Okay.

And so many of my peers, they reach age 47. I'm not saying they only shirk, but they stop trying to improve. They just cruise at a level of working pretty hard. And I think I just don't stop trying to compound returns.

I see that too. A lot of people reach a certain kind of level and then they're, they're kind of plateau. Um, I'm not sure it's like on purpose that they want— that they just plateau at that level, but they, they stop putting in maybe like the core effort to keep improving. Um, is there anything people should be thinking about on that, on that either like career track or just kind of a core improvement track that they, they— because they may not understand how, like, what the difference is going to be 20 years hence.

Well, those people are probably the sane ones, to be clear. So maybe you should all be advising me how to become more like them. I suspect it's something you're born with or not. You can cultivate it at the margin, but there's some kind of relentlessness, durability, persistence that characterizes a lot of very top performers, but it's fairly rare. I think usually it's worth more than IQ, provided you're smart enough. I don't know that it's better for people. It's probably better for the world. Elon Musk could retire and not spend all the money he has, but he's driven to do whatever he's going to do. Great.

Is there a way of measuring ambition? Because I can tell you of the 22-year-olds I know that I thought were super ambitious, a lot of them maybe reached a certain level of success and then the ambition didn't kind of keep propelling them forward. And then there are other ones that maybe I didn't think were so ambitious, but they turned out to be maybe way more ambitious. Like, I don't know that as a 22-year-old I could have measured my peers in ambition. Do you think there's some— like, if you think of those 5-factor things, is there some way of measuring ambition, um, before you could know if someone was super successful?

This gets back to your earlier question: what's the data set you want, right? This is a big part of it. But I'm partly a fan of Peter Thiel's question, which I think is surprisingly effective, though it may get spoiled as it becomes better known. Just ask people, how ambitious are you? People who are not that ambitious actually have a hard time faking it. So that question is somewhat—

Did you say like from 1 to 10, how ambitious are you?

No, not 1 to 10, because no one's going to say like 7, and you won't know what their scale is. Make them use actual words. How ambitious are you?

Got it.

Okay. And the bullshitters, they're going to be some people who say, like, I want to rule the universe. And I think often, but maybe not always, you can tell they're bullshitting. And the people who are quite specific as to what they want to do, how they want to do it, even if they end up being wrong in the specifics, you can tell they've thought about their ambition and thought it through. And they say it with conviction and something in them, like, leaps forward. That almost kind of scares you a bit, right? Then I think, aha, we've got something here.

But if you think, if I went back to like the 22-year-old Tyler, do you think I would learn something from that? Like, what do you think you would have said back then?

Again, this is the data set I want. But I think I would have said, I want to be one of the most influential thinkers, and then fill that in with detail. But I believe that's what I would have said and indeed did say when asked.

Another thing that you think a lot about is time. Um, and, and you've kind of interesting take on the value of time where you kind of— you, you believe that we shouldn't discount the future. And this kind of goes a lot against what I've learned all my life and maybe mainstream economic thinking. Like, what's your reasoning behind that?

Well, I think the discount rate should be zero for happiness or well-being, right? So there's no intrinsic reason to put off going to the dentist. But the economist's view that you discount dollar flows because there's an opportunity cost, because you can invest at positive returns, I fully agree with that. That's just common sense, and I think that's what you mean. But when it comes to well-being, should we sacrifice millions of lives in the future rather than one life today just because the future comes later? I would say no, that's immoral. We should treat the present and future on a par in that sense. And that's going to mean we should invest more, take a longer-term view, actually work harder, and be more creative. So that's consistent with what I call my common sense morality.

So a life 100 years from now is worth the same as a life today or something?

Correct. So we should do things now to make those lives better. Most of all, that means leave them with healthy, well-functioning institutions. We can't microplan their lives, right? A lot of investment, a lot of science, a lot of innovation. It's what's worked in the past. So I think we have a moral obligation to do those things in a way that is a stronger obligation than I think what a lot of other people believe. You read philosophy, we have all these other moral obligations, race, gender, all that. I don't mean to dismiss that, but what I want to put on the table is our obligation to help the future as much as possible by doing good work and good trust and good institutions right now and bequeathing them.

And how would you think about it on a personal thing? So obviously, there's short-term happiness, short-term fulfillment, long-term happiness, long-term fulfillment. Sometimes there's these trade-offs that one has to make between the two. Maybe I don't eat the chocolate cake for more longer-term fulfillment, longer-term happiness or something. How do you think we should be thinking about those trade-offs between our self today and our future self?

I think in most cases, there's a reasonable degree of harmony of interest, but there's not in every case. If you face a grave decision, should I get chemotherapy for cancer? You're definitely going to be more miserable when you get it, I think. It might save your life later on. I'm not sure I have any particular insight into those kinds of choices. But I would just say treat your future self as being as valuable as your current self.

We talked a little bit about the last 2 years and how maybe innovation has increased. What are some other big, you think, positive surprises in the last couple of years since COVID And what do you think are maybe some of the big negative surprises?

Cost of producing effective batteries seems to be declining much faster than we had expected.

And is that something you think is related to COVID, or was that inevitably going to happen whether COVID happened or not.

It seems it was inevitable and it preceded COVID, right? But how consistent that trend has been, the cost curve falling for solar and wind has done better than I think most people were expecting 15 years ago. Again, unrelated to COVID, that's a big deal, right? So on the green energy front, just that electric cars have come so far, the share price of Tesla, remarkably high. Again, if you go back only a few years and people are like, I'm going to get rich selling this short. I'm not saying no one says that now, now, but it's like, whoa, you better watch out. We see Elon is in fact this model of the truly ambitious person, and he's going to kick your butt. And that's great news. And I would say certainly the market didn't expect it.

What are some of the maybe the negative surprises?

Well, how poorly policymakers and public health authorities have in most but not all countries responded to COVID. That's been just terrible. That's well catalogued. But as we're speaking, the FDA gave final approval to the Pfizer vaccine, what, 2 days ago? That is insane. And they're insisting for such a long time we have to take it, and they won't even approve it themselves and say it's fine, although it saved many hundreds of thousands of lives— not many hundreds, a few hundred thousand lives— and it's the best studied vaccine in history.

Well, talk about fast grants, because I know there's something you pioneered, and basically, you know, where you're quickly giving out these researchers money to help them further their research. And I kind of— I think the key term in these fast grants is they're fast, right? Um, and so you have limited data you could take in to make a decision about what— how you grant, how much you grant, etc. What are the core heuristics that you think are the most important to use to make these grants faster?

I think the heuristics were very easy. It was actually supporting work that was going to matter rather than supporting nice research for its own sake. The real contribution is not that we picked better projects, but that simply no one else saw it as their job to get money to these people quickly. So a simple example, when the new variants first came along, the US was doing almost nothing to track them. And the potential tracking stations, we took 7, 8 of them actually, and sent them money within 2 or 3 days, just sent the money and we said, track. And then suddenly within days, the US had a lot more tracking capacity. Now, there's thousands of people who, in a sense, could have thought of that, but there was no one ex ante whose job was, if something like a new variant comes along, it's your job to get the money right away. And it was our job to do that. And that's what was scary.

That actually sounds really hard because maybe the new variant one, it's more obvious, but to be thinking of all the core important problems is actually— and then going out there and finding the people that are working on them, seems like a really, really— it seems like hard to do that fast. Like, how are you able to streamline that?

Well, that's why it's useful to read 5 or 10 times more quickly. But look, we had a team of 20 referees who were superb, and they had an incredible base of knowledge, and they were, you know, tenured at well-known universities and laboratories, and they did the refereeing. They deserve most of the credit. But I had the attitude This doesn't have to command unanimity or consensus. If a subgroup are enthused, we're going to do it because this is like venture capital. If some people don't like it, screw them. Our goal here is to solve the problem, not to look good in the eyes of everyone. That was important too. It's not knowledge or data per se, but it's temperament.

Is there using that FAST grant model, can we use it in other places in society? Maybe you could say like Y Combinator is like that in venture capital, but are there other places you think we could mimic FAST Grants in?

Well, in part, FAST Grants was modeled after venture capital. Y Combinator, but many other venture capitalists, yourself included, you could say, right? You're an angel investor and you do VC. I bet a lot of your best decisions have come pretty quickly. You felt it was right. And you rely on your best decisions paying off a lot, and you're willing to accept a lot of losses because not everything's going to work out, right?

Yeah.

So that attitude is not what you find in academia. It's not what you find on the government research funding side. There are slow bureaucracies based on consensus, and even it's like emergency funding. National Institutes of Health typically took 4 to 5 months. We were doing things often in 2 to 3 days. Our longest would be like 2 weeks. So in a pandemic, speed really matters. And we funded many different things. I assumed final responsibility for all the grants. I'm not like a presence in the biomedical community. And just, I never gave a damn if anyone was happy with my decisions. The view was we're going to look for vampire killers and, uh, full steam ahead.

Well, and there's this— if you mentioned like, like whether it's a National Science Foundation or all these other like granting bodies that are out there, even like some of these big foundations that are granting, they, they seem to be more and more granting to established people and established bodies. And, you know, the, the average age of people who are getting the grant seems to be getting older. Um, is, is, is it just a kind of a risk tolerance thing, and, and maybe we should make it— make— give them a sense that they can be a little bit more risky, or they don't want to fail, or why do you think this is happening?

I think there's more specialization, more bureaucratization in societies.— you know, running a major lab is more important. That tends to be older people. All those things may be fine on their own terms, and the NIH has funded a lot of important projects, including early research on the mRNA vaccines 20 years ago, but that doesn't mean they're good at responding quickly. They are not, and they didn't. So for there, you need other institutions. I would like to see the creation of a new branch of something like the NIH, but completely geared toward acting now when emergencies arrive. And we all saw COVID as an emergency. But keep in mind, if you're someone with a disease, it's an emergency to you. Malaria, dengue, and all the way down the list. Those are emergencies to very large numbers of people right now. They don't feel like it to us in the current media cycle, but that's an illusion.

Now, you've also been the case that like big corporations are underappreciated. Um, and, and at least from my perspective, they certainly seem like better places to work than they were 20 years ago. But how, how do you think people are just not appreciating these larger businesses?

I think they're starting to, especially with the pandemic. So we're doing this over Zoom, right?

Yep.

Zoom is now a big company. It wasn't at the beginning of the pandemic.

That's true.

Yeah. But think how well and how quickly they scaled up. That was pretty amazing. Look at Amazon. How quickly Amazon shifted into delivering so much food to so many American households. Truly astonishing, one of the great miracles of our time. Probably only Amazon could have done that. Obviously, they're a big company, all the bigger now. People have seen who has stepped up to the plate, and the performance of big business in the pandemic, I think, has really been astonishingly good. And many people have seen that.

And, and you think that will continue? Do you think there's like more, uh, uh, essentially rents that will go to the bigger players, or how do you think that happens over time?

I think big businesses have done a good job attracting talent because they pay more, offer bigger benefits, and it's just easier to find friends at work in a big company, right? Yeah, if there's diversity So, uh, if talent has been underrated for a while and big businesses invested first in talent, and then there are network effects, I think they will continue to do well for the foreseeable future.

How do you think startups can compete with that?

I think in the Bay Area, we're seeing a lot of very rapid evolution where startups are distributed by nature. And like, maybe someday they'll all get together in an office,. And so far, it seems to me that has worked. And in general, the rate of business formation during the pandemic has been much higher than average. They may not all be startups in the Oren Hoffman sense, but in the literal sense, they're startups, they're new businesses. So it seems to be working, and people doing this out of their homes and over Zoom. At what point that will stop scaling, that gets back to my questions about organizational capital. I would say we don't know yet. But so far, so good, right? Full steam ahead. Keep on trying it till it don't work no more.

One of the things I think you do pretty well is like, is to question your core beliefs. Do you have some sort of systematic way of doing it? Like, do you like have a meeting with yourself once a year and you go through your core beliefs and see if you still believe them? Or do you have any, any way of actually thinking them through? Or, or does it just kind of evolve naturally over time?

I don't know if I'm good at questioning my core beliefs. Maybe I'm good at questioning my peripheral beliefs.

Okay.

And that comes from interacting with a lot of different kinds of people, having ties to the East Coast and the West Coast, having grown up in what was originally a working-class family, but now being connected to elites, having lived in a bunch of different countries, having visited 100 different countries, having learned as an adult two other languages, spending a lot of time with culture and the arts. And you just put all that stuff together, you get good at cracking cultural codes. But my core beliefs, I guess I think I'm dogmatic about them.

What would be an example of some core beliefs that you think you're dogmatic about?

People should work and innovate, not complain too much, not be whiners, have a relatively positive attitude. Life is wonderful. I don't think I do a great job questioning those. I think they're true. So I'm not looking to throw them away. I'm just like, I'm going to run with these. They might be somewhat wrong, but at this point in the game, what other hand do I have to play?

Is there some way of— some advice you have of maybe keeping your identity small enough so that you can feel comfortable changing one's beliefs?

Maybe my strategy, which I'm not sure is good for everyone, but just to make your identity so large that your dogmatism gets soaked up in your core beliefs, and you're so committed to those, you feel you have the freedom to change your peripheral beliefs.

All right, I like that. That's pretty interesting.

Like, I'm just not going to stop having a positive attitude. I think it's not going to be like, oh, I'm 68 years old and I've decided to reinvent myself as a whiner. Yeah, yeah, maybe it will happen, but somehow I don't think so. I'm too dogmatic to do that.

Right, right. You're too— you're kind of wrapped up in an optimistic view or something.

Yeah, like, let's do something. Like, even if things look bad, like, my goodness, let's get out there and do something.

Or you're a little bit obsessed with asking your podcast guests around their, you know, production functions. You've asked a lot of people now, both on your podcast and off, about their production functions. What have you learned from their collective responses?

I mean, probably not much at all. So they all say different things. I'm not sure I should take any of the answers literally. Sometimes I wonder if I'm not making fun of them by asking them about their production functions. I think the audience enjoys it. It's a way to get people to open up. I enjoy the answers, but it's maybe the answers I take least seriously. And a lot of people are just not very self-aware about how they've succeeded, maybe especially if they're highly successful. I would interpret all that in strategy and fashion.

I would say that's the main thing I've taken away from some of these, from some of those questions as well, is that the answers are, are, you know, I don't think are true. And I don't think they're, they're, they're actually trying to lie to you, but I don't think they necessarily have ever thought it through as to why they've been so productive, why they're able to produce so much stuff. You ask this author that's, you know, written 30 books what your production function is, and they're like, well, I like to have a cup of tea in the morning or something. Like, that's almost certainly not the true answer.

And in my asking, there's a bit of, "Haha, look at what this guy's going to say." Right?

What do you think are some of the most important skills that a person could have to lead to success that maybe were true in the past that aren't true in the future?

There's some NBER working papers on this, but soft skills and social skills seem to be bringing higher returns than they used to.

Today?

Today.

So like empathy or something?

But not empathy taken alone. So everyone obsesses over STEM, but I would say people who have expertise in something concrete— could be STEM, but doesn't have to be, could be like geology— and then have great soft skills, they seem to be earning much more. And that's in the data, maybe not proven, but the papers on it to me seem pretty convincing. And it accords with my intuition and my anecdotal observations.

And by soft skills, you mean like They're not jerks, or what do you mean by soft skills?

They can communicate with other people, synthesize information, outline a vision to other people, set expectations, accept feedback, give other people feedback, and then wrap the whole thing up into some synthetic package that they can then move forward with in the days, months, years to come. It's really a pretty tall order. It involves a lot of subcomponents. And if big businesses are doing better, that means teamwork is more valuable. People who have this, they're like— we used to say glue guys, but that's sexist. We've got to give up the alliteration and say the worst-sounding glue people, but like with some intellect. And you've got to know something about some area. You can't just show up like, hey, I'm the empathy guy. I love you all. I feel your pain. Right? That's like, come on, take a hike.

Got it. So you need a little bit of both. There's like a Laffer curve of being somewhere in the middle or something, or?

Synthetic synthesizers is another way to put it.

Yep.

Hard and soft skills. So the person who is only like a super nerd, incredible calculator, you can outsource them with work from home. Like there's going to be someone in India smarter than they are. Like they're going to do well, but they're not actually the top earners. The people who can do both, they're the real prizes.

If we're moving to a world that is going to be more virtual, in some ways it means that one could interact with more people. Right? And do you think that's gonna expedite innovation, or do you think that will quell innovation because you don't have as many repeated interactions with the same person?

So far it has radically accelerated intervention— sorry, innovation. There is this question, are we eating into our organizational capital by doing so many things at a distance? But I'm very bullish on the trend. It could be like only 20% of previous work becomes work from a distance. But if you can bring to bear on your team, the smartest people from Pakistan, Nigeria, South Korea, wherever, I think that's going to be amazing. And yet you wish you could see them more. You fly them in once or twice a year. That's inefficient. Like that's probably all still true. But the value of getting some key people from everywhere, I think will be very, very high. And especially for America and especially for Silicon Valley.

Tyler Cowen: Now, how do you think the reward should work? Because there are companies like Google have made a public statement that they've chosen to pay people differently based on the location where they live. So if you move from New York to Indiana, you get paid less. If you move from Indiana to New York, you get paid more. Other companies saying, okay, wherever you live, if you're doing the same work, you get paid the same. What do you think is the right approach? And do you think there's a certain type of model that will win long term?

I think we'll be forced to move to an even higher degree of pay by performance and bonuses and measurable output. As people are at a distance and this goes on for years, like in part, you don't know, you've got to measure it. So everything will be more bonus-driven, more metrics-driven. And the person in Nigeria, they're going to accept a lot of pay risk, but like even the downside, they'll be able to live in Nigeria think they'll be willing to do that. And it won't be like an upfront salary is the main part of the deal. You'll get too much shirking with that.

If you think of organization, if you think of like sales and engineering, sales is weirdly, when it deals with their employees, is extremely data-driven. And engineering, when they're dealing with like employees and compensation, is almost the opposite. It's almost like very little data-driven when it comes there and somewhat hard to measure things like performance and stuff with engineers. You kind of know it when you see it. Do you think companies are just going to have to get smarter at measuring productivity and measuring output?

We'll have to measure it more. AI might do some of that for us, right? So maybe the company won't ever be that smart, but it will ask its AI, like, "Hey, who did the real work on this?" And there'll be an answer. It won't even be transparent to you. Like, how does GPT-3 work? Oh, you know, complicated, right? But you'll go by the answer when setting yearly bonuses.

I've known you for a while, so I have a couple personal questions I've always wanted to ask you, and this is as good time as any to—

I think no limits here.

All right, the first one is, I think I remember you once saying that you proposed to your wife Natasha after knowing her for like 3 weeks.

Correct. And 1 week of that I was out of town.

Okay, well, okay, now it took me a lot more— I needed to collect a lot more data to propose to my wife. So what did you— is there heuristic that you used to shortcut? I mean, this is a massively big decision in your life. Um, if you made the wrong decision decision, it could be, it could be really bad for you. Like, how did you shortcut that decision? I know it's turned out really well for you, but, um, you could see scenarios where it didn't. So like, what did you do to like feel confident in that, in the heuristics that you used?

Well, to think of it as like data gathering on Natasha has it backwards. My way of thinking about the strategy was when could I get her to say yes? So if there's a moment when I could get her to say yes and kind trick and trapper into this. I'm like, "I got to take this." So the notion I'm going to spend a year thinking about it, like, no, you have your yes, you've got to do it. Right?

But I mean, you have some sort of sense that it needs to work out for both parties, right? This is a marriage, both parties have to work out and ideally both parties have to continue to be happy with this decision for a very long time. There must be some sort of heuristic you use to come up with this. This decision, right or no?

I don't know. It seemed like the right thing to do. It wasn't the result of an elaborate plan and it's worked out. So again, I'm not recommending it to the audience necessarily, but I suspect if you had the data, very long stretching relationships and then a proposal, I think they would do worse. I'm not saying it's causal.

Yeah, yeah.

I think it reflects a difficulty of of committing among at least some subset of people. Yeah.

Or maybe reflect the fact that maybe deep down they knew it wasn't right and that's why they never did it or so.

Okay.

Interesting.

And maybe they settled or, gee, you know, became too inconvenient not to.

Yeah.

So I don't think it's such a great signal to wait too long.

I checked my inbox before this thing and I realized that we, you and I had met because you responded to a cold email that I sent you in 2007. Uh, and then I looked through and you're just like incredibly good at answering email. And I assume based on who you are that you get a ton of email. So how do you— do you have like a strategy doing it? Do you have a philosophy about that? Like, how do you think through, like, the just thinking through all this random strangers that send you email?

I sometimes say that answering my email is my business model, and I mean that very seriously. So, I don't respond to abusive emails or spam pitches, but I do at least try to respond to every email I get, though don't tell anyone. If you're listening to this podcast, forget I ever said it. And that's how you meet really important people, is to take them seriously. So, you got to yourself send out the right bat signal to get the right people emailing you, but you emailed me. Patrick Collison emailed me once. This was before he was so well known. I thought, "Hey, this guy sounds smart. I'll write him back." He and I ended up doing Fast Grants together. So very seriously, answering email is my business model. By the way, how good was my response to you?

It was great.

How good was your email to me?

It was probably pretty good, yeah.

Yeah. Well, forward me the exchange if you're willing to. I'm curious to read what we each said.

Yeah, yeah, yeah, I will. I mean, I invited you to something and you basically responded very, very quickly saying, "I'll be there." And was I there? Yes. And you showed up. In person, in real life, which is pretty amazing. This has been incredibly interesting. The last question we ask all of our guests on World of Das, if you can go back in time, what advice do you wish you could have told your younger self?

That when I get to be older, don't give me any advice. Because look, for me, things have gone pretty well. So there's always the risk with advice, even if you make a local improvement, you'll screw up the global path. So if things have gone well, yeah, of course you could have done better. But again, type 1 and type 2 error. And it's like, let's let that one sit. So no advice.

We're kind of saying, going back to what you said earlier, it's just, just work on compounding the good stuff.

Yeah, but I already knew that. And advice is dangerous. Like, how well can you predict other people's paths? Advice maybe is overrated. I think a lot of advice is a placebo. The person asks for advice, because they want the feeling they've done everything possible before doing what they're going to do anyway. Now, I don't mind that, but once you realize that, it's like advice. Like, you know, I don't necessarily think advice is advice. It's about helping the person process their own mental and emotional state.

All right. Well, this is great. This has been awesome.

Okay.

I highly encourage everyone to follow you on Twitter, to subscribe to Conversations with Tyler Cowen, which is one of my favorite podcasts. Any other thing people should be doing to find you on the internet or or anything else you'd want the audience to do?

My blog is Marginal Revolution, and there's books by me on Amazon and plenty more, all findable through Google.

Perfect. All right, well, thank you, Tyler. Thank you for joining us.

Look forward to seeing you next door, and take care.

Thanks for joining us. If you enjoyed the show, consider leaving a review. For more World of DaaS, that's D-A-A-S, DaaS, you could subscribe right here on our YouTube channel, or you could find me on Twitter @auren, A-U-R-E-N, auren. I'd love to hear from you.