Matt Clancy on Innovation Studies, Progress, and Remote Work (unedited transcript)
Ben Yeoh (00:01): Hello, and welcome to Ben Yeoh chats. My personal podcast. If you're curious about the world, this show is for you. What's the future of remote work? How best do we spark and develop innovation? In this episode, I speak to Matt Clancy. We discuss how Matt is thinking about innovation, the censorship of work he's doing on what nurtures innovation, and the future of remote work. If you liked the show, please like, and subscribe as it helps others find the podcast enjoy, Hey everyone. I am super excited to chat with Matt Clancy. Matt is a progress fellow at emergent ventures. He teaches at Iowa state university and writes on sub stack a newsletter called new things under the sun, which you should subscribe to if you're interested in anything innovation-related. Matt has also synthesised many of the emerging studies on remote working. So, Matt, welcome.
Matt Clancy (01:00): Thanks for having me on, it’s refreshing to be here.
Ben Yeoh (01:04): Great. There has been plenty of debate as to whether innovation has been stagnating and whether it might now be rejuvenating to some extent, and there's this observation and several studies that showing or suggesting that it's costing more R and D dollars to achieve similar progress across several domains, such as pharmaceutical drugs, crop yield, semiconductor productivity, and that seems to hold, but there are counter-arguments around whether we're measuring the correct things. And I think Dietrich's Vollrath has attributed apparent progress a lot to human capital factors. What's your reading of the arguments? And do you have a sense about where we stand and how obvious this observation might be or not?
Matt Clancy (01:51): This is a huge topic. But so, there's kind of two waves of stagnation that people talk about. And there was the big stagnation, which was kind of starting in the seventies or total factor productivity is one way to measure. It was one of the people trying to measure the technological progress of the entire society. It's roughly a statistical way to pull out. I would say the extra oomph you get from the same number of inputs. And that was something that grew at a rate of 2.8% between 50 years around 1920 to 1970, maybe, and then it fell to 1.6% since then. And so that's sort of the big one because that compounds over time and it makes a big difference, but then there's been this sort of smaller mini stagnation since the two thousand. And 1.6 fell again. And where I stand is that I think that the significance of the big stagnation is real, but I think it's a little bit overstated for some reasons. I think the significance of sort of the litter stagnation since the two thousand is not really about technological progress slowing down. It's more about other things. So Vollrath has this really great book Fully Grown, and he really digs into a number of the second stagnation, a little stagnation, and it shows that you can really account for this without having to go to technology. It's kind of that old quip about you don't mention God in your explanations. I wasn't the car or something, and he's God, have no need for that hypothesis.
…does the same thing with this little stagnation. … if you look at declines in geographic mobility, the decline in the start and stop rate of firms entering exiting, you can account for most of this total factor productivity slowdown since the two thousands and a couple other things. And you don't really need to say anything, it's due to technology. And also, a big one is the transition of consumers from purchasing easily, products that are manufactured and are really easy to make them better and make total factor productivity better for making a TV and consumers are just buying less of that stuff and buying more and more stuff like services, healthcare and education and stuff, which is harder to advance. It's harder to improve the total factor productivity of those things. Think of daycare, my kids go to daycare and they're cared for by like one person for every four kids. And they're babies. And you don't necessarily want to have one person caring for 60 kids or something like that. Right? There's something about that person. So it's hard to improve the efficiency of labor in that sense. Anyway, we're spending more money on that stuff, because basically the other stuff is so cheap and we've gotten so good at making it. It can be the case that there was no slowdown and technological progress and making TVs or making computers, but we kind of satisfied our demands, we’re ok, these are great. And I'm going to use the money I saved instead of buying an even bigger TV, I'll buy only a little bit bigger TV and use the extra money on this other stuff.
And that means more spending is allocated to the slower-growing sectors of the economy. There may not have been a slowdown. It's just that we're spending more on the slower-growing sections. And so, it looks like a slowdown in TFP. That's a long-winded thing about the second one, but the first one you can't necessarily write it off using the same kind of techniques. And Noah Smith has this kind of interesting theory that it's all about the energy shocks in the 1970s. And that technology was on this trajectory up through most of the 20th century until the 1970s that was premised on using more and more energy to do more and more cool stuff to fly to the stars or robots and all sorts of stuff. And in the 1970s, the oil shocks made oil really expensive. And I think that there's this my working hypothesis and I haven't dug into it. This is the kind of thing where, it kind of qualitatively sounds like it could work, but I haven't dug into the numbers to see if it really can account for everything. But my sort of pet theory is that similar to Noah Smith is that oil shocks meant people felt we can't count on. They're just being abundant energy. And instead of diverting resources towards R and D, that's going to depend on there being abundant energy. We're going to divert R and D towards energy efficiency. And I think this created this big expectation. Scientists were just as productive as ever at figuring stuff out, but the kinds of things they were figuring out what the things that people had been conditioned to start to expect through movies and visions and extrapolating from trends.
And so instead of getting supersonic planes and space flight and stuff, we got energy efficient Peters or stuff that sounds lame, or there's this better installation, but there's this clip that the speed of travel across the United States hasn't sped up at all, but it's gotten a lot more energy efficient over that time. It's got a lot safer.
Ben Yeoh (07:39): Planes become safer.
Matt Clancy (07:40): It's has and so it's the future doesn't look like we thought it would because we were extrapolating and we sort of changed tracks. And so that's why I think that it's real, but also that it is a little bit over I don't know, like exaggerated, because I think the progress is real, but it's just the world feels qualitatively different when you have new kinds of products that you've never seen before doing a new kit with new capabilities, versus when you have your existing products being more efficient and safe it just doesn't feel necessarily things are advancing at the same rate because you're not being surprised and wow, but I don't know if that necessarily, I wouldn't necessarily think the second kind of innovation is any less important. My take is that the energy shacks grant set us on a trajectory that was different. And so it doesn't look like we expected it to, and then on top of that, there's this period of probably inefficient research or where R and D wasn't producing technologies as quickly as it had before, because they had to sort of retrain, rebuild up the knowledge set to do this new type of on R and D that was promised on different things than abundant energy. And that meant people weren't as productive until they sort of rebuilt the knowledge base to where it was for energy efficient. So there's this famous paper by Daron Acemoglu about climate change and how we need to subsidize technological progress in addition to doing a carbon tax.
And he sort of talks about how R and D for fossil fuel-based transportation have continued all along internal combustion engines get better and better and better. And as they get better and better and better, it gets harder and harder to switch to energy-efficient stuff, because there's going to be if the technology is not at that level, then there's this big step down in living standards as we transitioned to this other thing. And he is talking about how you need to subsidize that other technology field to sort of catch up to providing the same capabilities that we can get from the fossil fuel-derived stuff. And there's a role for subsidies in addition to just carbon taxes to doing that. And I think it's the same idea. And then the last part is just that I think people have this bias towards things that were better when I was younger.
Ben Yeoh (10:17): Right, yes.
Matt Clancy (10:17): That's just this continual, that's a sort of this under [Inaudible 00:10:21] [Cross-Talk 00:10:21] in the background.
Ben Yeoh (10:24): Great. Well, that's raised a load of interesting thoughts which we can discuss. So one is on a fact we spoke earlier to Anton house about the difference we put on incremental innovation versus transformational. And I think you're saying incremental is kind of undervalued. And the second is I think what I would think of as what some people might call the Baumol effect, this idea, like you said, on childcare, which would be the same for waste services, arguably the same in some big sectors like healthcare and education, you want teacher ratios and doctor ratios, you can't scale that it's gone, it's can't scale that much further. And then a little bit, I think would be really interesting on that climate fossil fuel piece, because the data I've seen suggests that for instance, the R and D part looking at renewable energy worldwide, both public and private has only been around the order of $15 billion, which is a real tiny drop in the ocean. We spent about probably 2 trillion orders of magnitude on R and D last year and much more on fossil fuel, R and D like you say, I think at least 10 times more, whereas renewable energy pieces are only 15. I'm kind of interested though, I guess, in two parts on the measurement things how do you think this accounts for these kinds of more general-purpose inventions or not, I'm really intrigued by, for instance, a double-blind trial design. So these are kind of what you might call RCTs or whatever. And I'm interested in a couple of ways because one is in theory, we could have invented them a lot earlier. I think actually, even in ancient times, a couple of empires had this, but it didn't stick. And then even famously when it was kind of reinvented around the time of trying to eradicate scurvy with ships going, they kind of lost it a couple of times before it stuck. And then it started in medicine and now is in economics and all over. And I was wondering about how effective you think that those general-purpose things go into these measurements and the second. Actually, why don't you start on that one?
Matt Clancy (12:36): About how you measure these general-purpose technologies, that's one reason that people like things like these really big aggregate measures, just real GDP per capita or total factor productivity, because the hope is that all that stuff spins out into very easy rather than trying to figure out exactly what the benefit of say double-blind testing is, you're just like, well, eventually it ends in better knowledge, which lets us develop better pharmaceutical companies. And that gets reflected in their sales that gets reflected in real GDP, real GDP per capita. And, that's kind of the argument for these really big aggregate ones is that you don't have to get into the nitty-gritty of thinking of all the different ways that any one invention could have positive effects, but then the downside is that it's so far removed. And also, that there are lots of other things that can affect those measures than just the quality of knowledge. And so, it's a real problem. And in terms of the benefits of RCTs or double-blind testing, I guess, I think that's a really interesting problem too, how would you measure the gains from that? I mean, you would look at if I was going to pick a domain for one, so medicine and you'd look at kind of the quality of the trajectory of improvement in diseases that use this method versus ones that don't, that might be one hypothetical way to do it.
Ben Yeoh (14:10): Sure.
Matt Clancy (14:10): But in general, I think it's real.
Ben Yeoh (14:12): Arguably you don't have any modern medicine, so no pharmaceuticals would exist without this. Because you wouldn't know whether they work sort of. So that is essentially all of the medicine doesn't exist without this one general purpose. And then actually it spills over into many other things that could have started in medicine, but obviously, you have it in these other things. So intriguing. Do you think TFP actually accounts for that or does it miss something?
Matt Clancy (14:38): I mean, TFP and real GDP per capita, they both all kinds of, the things they miss are things that are not captured in transactions because ultimately, they derive their value. So if double-blind testing develops better medicines, then people are willing to pay more for it, or even if it has this indirect effect where they live more and then they're able to produce more economic value through their lives that would get caught, that would sort of be in there. But the satisfaction that you get from being alive or from being able to see grandma for another 20 years, that is just missed from these figures. And so, it's a problem. What people sometimes say is well, if it's a problem, hopefully, it's not a problem that changes its intensity over time, because then you can, at least still, you can then add tack on an extra 50% in value. But since it's constant through time we can still look at the trends we have. I don't know. I mean you have to make, for any specific case, you can probably make arguments, whether that's sensible or not. But I do think this idea that double-blind testing can be lost and it's an old, easy-to-understand idea about how come people didn't do it. I think it kind of emphasizes the value of good measurement for just innovation. If you want to facilitate incremental innovation, being able to detect a 3% improvement is really important. Because otherwise, if there's a lot of noise people will just forget it and then nothing gets, we don't build on anything. You don't, we just get stuck trying new things and abandoned them and trying different things instead of keeping the ones that work and then getting 3% better every year.
Ben Yeoh (16:32): That makes a lot of sense. And I guess you partially answered one of my follow-up thoughts would be on one aspect of that. So I'm quite intrigued by pain or pain control. So because actually, it doesn't cost very much now for pain control. So morphine or even headaches, aspirin cost very little. But if you ask people, the utility is quite high to not suffer from headaches, even though we don't pay very much for it in dollars. And then actually you can take it to an extreme where today and arguably say for the last 50 to 70 years, you can now spend the last month of your life if you've got some terminal illness, pain-free. More or less pain-free because you could be morphine or something that. And if you were to ask people and some people do, would you exchange half your fortune or even your entire fortune to now be able to die pain-free in the last month of your life, a lot of people are kind of saying yes, and your family around you, knowing that you can be pain-free is incredibly valuable. And I can understand that's one single step up because we're not going to end that. But I am interested because I'm not sure it is constant over time, although I have no idea how you'd measure it, it seems to me enormously valuable that now for quite a lot of humanity for the rest of time, more or less we can die pain-free I did. That's kind of an unbelievable amount of value, which is not captured in any of these measurements. Is that the kind of correct way of looking at it or are there more complications about how we see value or I guess progress in the wider sense?
Matt Clancy (18:16): I think that what you're getting at is sort of the difference between consumer surplus, which is the gap between somebody's willingness to pay and what they actually pay and then what they actually pay. And in a perfectly competitive market, the price something is charged as the marginal willingness to pay at the last person or whatever, but there are all these other non-marginal people who are driving enormous benefits from it and sometimes you hear. I haven't heard people talk about it in the sense of alleviation of pain. And I think that's a really good example, but it is something that people have wondered a lot about in terms of the quality of digital goods. the value of the internet basically is people don't spend very much, there are still costs associated with it that show up in GDP, transactions that are conducted over the internet advertising that is spent. They spend more on ads if they think people are going to spend more on time, which is in some very indirect way a signal of the value people gets from being online, but it's very indirect and it's, and there have been estimating that the value people derived from say Facebook is I think at the median value in surveys and auctions, where they do these experiments, where they're going to actually pay people to delete their Facebook account. And the median willingness to pay is over a thousand dollars. Maybe that's the mean, but anyway, it's really high. And obviously, Facebook for most people is free. And I think that the ads per person that are spent on Facebook are way under a thousand dollars too. People have looked into trying to say, well, maybe this mismeasurement is why we seem to have this little stagnation since two thousand.
And the sense I get is that most people think that that doesn't really seem to work. And Robert Gordon who wrote this famous horizon fall of American growth book that sort of surveys, the history of innovation and technological change in the United States, going back to 1850, also talks about these uncaptured games. And I think he would argue something, well, the gain from morphine is a lot bigger than the unmeasured gain we get from pain alleviation and morphine, or whatever is much higher than the unmeasured gains we get from the internet. So he would argue that it makes the case stronger for a slowdown, he emphasizes things, people used to wear clothes that were just really uncomfortable and scratchy, and that was their whole lives. And people didn't have air conditioning, you have to live in the ambient temperature. And we don't in the developed world. We don't think about these costs anymore. And those gains were not the full value people were willing to get. Those were not necessarily captured in the GDP and ways of measuring total factor productivity were a lot shakier back then too. So he argues that we are undercounting the gains in the past if anything.
Ben Yeoh (21:24): That makes sense. Maybe a question we will probably not quite ever fully answer, but I can see those arguments, maybe taking one, why is the step touching perhaps into philosophy what do you think about other aspects of progress, which we might think about as progress, but what doesn't necessarily square into innovation? So I'm thinking about social and, or maybe moral progress, maybe most obviously seen for things like say women's rights or minority rights all going back in time even to slavery, where you get a consensus today generally that's better and therefore quite valuable, particularly for those groups say women's and minorities. And again, I guess this is something which generally doesn't go into GDP, although there probably is a GDP element by getting women into the workforce or TFP, generally think there are some aspects of progress and actually the arguments that some of those harder to add to ascertain parts of progress. I guess gay marriage for people who want that now, which is on the increasing rates of a moment, they are seeing some progress and therefore some innovation. Would you put that within the scope of kind of progress or within that, or is it just too far outside what we can measure?
Matt Clancy (22:42): No, I certainly think that in some ways that's sort of the most important types of progress. This kind of progress is ultimately probably more important than a lot of material progress that's made. But it's definitely not captured in GDP. And that's a problem. One way that you can try to get around that is to try to use, I don't know, the measure of, people do try to do these indices of inequality, for example, where they adjust for utility people get maybe from different levels of income, where if you have declining marginal utility, then that means having a couple of really rich people doesn't offset having loads of poor people, but still, that's still ultimately getting into material measures. I think that these other kinds of progress don't follow necessarily a lot of the same dynamics that I study intellectual property rights or emerging from basic science or this or that. But I still think they're really important and undervalued, and I have this sort of thought actually like that we should have. We need to, there's a little bit too much emphasis sometimes on just the hard material innovation and there's sometimes people underwrite this stuff. They focus too much on the material going forward just having cool new sci-fi type inventions and sort of that being the main thing we want, but we shouldn't neglect sort of what I call it, I don't know the social technology that stuff is in some ways more important than for people's lives. When you think about this critique of the little stagnation that people are spending more money on services, instead of things, I feel that's a signal you don't necessarily want to ignore that.
People are all right, well, we've had, thanks for making these TVs, these are great. And thanks for making all this other stuff, but I'll keep all that. But now I want to focus on these other parts of my life and improving kind of these nine material parts seems something maybe since that's where people are spending their money, that's telling us that that's really important. And one thing that's kind of interesting I thought about sometimes these things aren't as unrelated as you think social cohesion or a marriage or something like that. And marriage has been dating and finding partners it has been really affected by online dating technology basically. And so sometimes the material stuff does have significant impacts on the quality of these aspects of our life that is not entirely mediated through economic exchange or I've been a big believer in remote work. And one of the reasons is that it lets people choose where they want to live and how they want to live their lives more based on things like their community and who they know their social ties rather than necessarily having that choice that constrained by this is where the jobs are that's in my skills.
Ben Yeoh (26:10): That makes a lot of sense. And you say, GDP actually correlates quite well with a lot of these other things like happiness or peace, although not perfectly. So thinking maybe upstream about innovation. So Anton house has this idea that it's kind of the improving mindset that you need before kind of innovation sparks in individuals. I was wondering whether you had any thoughts about what might spark off innovation generally, and what are the factors that you think about, people talk about agglomeration effects. Do you have to be in cities that we've got remote work and Anton thinks you have to start with a person and then all of these other things probably not proven any particular way, but I'd be interested in where you think there might be a spark of innovation. Does it start with the person in which case policy should be maybe designed around that?
Matt Clancy (27:07): I really like Anton's emphasis on people changing their mindset and changing what they believe. So Joseph Henrik has this book the secret of our success, which is sort of this long history of cultural evolution in the human species and so on. And he talks about how through most of human history, culture has got more complicated in tech. essentially, we would say innovation happened. It just happened at a really slow rate, but tools got more advanced and stuff. And that all happened without really an improving mindset. I think people just accidentally, maybe it did. I don't know, but for whatever reason, when people were learning how to make tools or whatever, or hunting practices they imperfectly copied whoever somebody else was and maybe they were a tinkerer looking to improve, or maybe it was just a mistake. Maybe it was serendipity, but observing what worked and sort of an evolutionary process allowed innovations to happen, but it was super slow. And I think one of the big things that changed in the industrial revolution is that people got this idea that you can, the way I should preface this by saying, so go back before the industrial revolution. And if you're living in that kind of world where people have a really poor causal understanding of how things work it's the best survival strategy is going to be good at copying what works for other people in your society and not trying to improve it because nine times out of 10, when you try to improve it, you're going to make it worse. And Henrik’s book has these really awesome examples of preparing poisonous plants for consumption.
And there are these extremely onerous. There are lots of things you have to do, it's very time-consuming and very like lots of effort. And many of the steps don't seem to do anything. If you're the one who's is just trying to reason about it, you can't see what the impact of boiling it for 24 hours is for example, but that stuff is really working. It's leaching the poison out and it's leaching it out in a way that you won't be able to detect just by taste and, or getting sick in 24 hours. But it would, if you didn't, if you skipped these steps, you'd pay the prices in by dying 20 years early or something like that. So in that kind of world where there are all these sorts of opaque things going on in the background of how your technology works, the best thing to do is going to be to just copy. And don't try to be smarter than everybody else and figure it out because if you try to take shortcuts, there's a reason thing are the way they are. They were selected that way through being the fittest and you're going to get wrong, I get it wrong. But then by the industrial revolution we had, technology had changed enough that there was sort of scope for just tinkering and figuring things out. I think some of this is just the sort of a new menu of technologies to tinker with that hadn't been optimized over millennia. So there really were some free lunches around. And then at the same time, you had this growing scientific knowledge of how the world works, that made also proactively going out there and trying to figure out better ways to do things like a more reasonable activity. And it's been, I think there's been the switch in the last 300 years were now actively trying to think of better ways to do things is suddenly something that sort of makes sense suddenly in a very long run in for our species.
But I think we still have this baggage from our deep evolutionary past if that's not something we'd normally think of doing. We think we're sort of client bias towards the status quo because for millennia, trying to deviate from the status quo is usually a bad idea. And so, what's interesting is I think in the modern developed world, you can get these, the status quo can become in some places that you just try to do new things. If you live in Silicon Valley or if you live in an area where there are lots of people doing new and exciting things that becomes the thing that you copy for other people. And you hijack these old processes where it's previously I would copy the canoe making techniques of the guy who lived the longest and seemed as I had just had the most prestige in the community and he built really good canoes. He was around because he made them and whatever. But now you copy Elon Musk or something. So it gives that sort of improving mentality. So I think that we're in better shape than ever in terms of creating improving mindsets, but we fight this really big battle because our bias is always going to be towards just doing what everybody else is doing, not taking new risks. And so we've entered and so there's something fragile about it, where if innovation would stop, it would kind of then it could sort of stop. It could get locked in at this frozen level because you suddenly don't have people getting the idea to try to do new things right now. Right now, we're in this really lucky equilibrium where there are enough people trying to invent new things, that is something that is seen as a normal thing to do. And you don't have to rely only on weirdos and outcasts to doing that, I guess.
Ben Yeoh (32:52): Great. I hadn't, I'm going to call that the evolutionary biology idea of innovation and status quo, which I haven't heard before, but I could see, social cohesion during the ice age, you have to do what everyone does otherwise, you're not going to survive. And therefore, we will do that. And now if you're in South San Francisco and you're not working on a startup, you're a little bit weird because surely that's what everyone else is doing. And so, you might
Matt Clancy (33:18): Exactly, but Henrik's bookers is an awesome example of explorers going to these places where indigenous people live successfully for millennia. And it's this natural experiment where these explorers get stranded and they have to figure out how are we going to survive there? And they can't, they usually die. And it's because pre 18 hundred or whatever the natural world is not the kind of thing that the human brain can just figure out in one go, it takes generations of standing on people's shoulders.
Ben Yeoh (33:50): You dropped me in the Indonesian jungle right now. You probably never thing hearing from me yet.
Matt Clancy (33:55): Yes, exactly.
Ben Yeoh (33:57): Maybe moving downstream one piece then. So say you are in charge of a medium, large firm says maybe 5,000 staff and maybe 500 or 500,000. So maybe 10% or 20% of your budget is on R and D. So you're working on innovation. How are you, I guess this is called a corporate lab question. How do you think they should be thinking about organizing that innovation or not? I'm particularly interested in, do you think small teams or large teams, and do you think the sector or the problem makes a difference? So whether you're eating out more, say incremental process improvements versus perhaps a transformational more moonshot idea, does that vary, or should we always be thinking about a small team or a large team, and then I'm going to go on and ask you about glomeration effects with that or not where you can cite it and also size, but let's start with kind of small team or large team or anything you should think about organizational R and D effectiveness now that we've decided that we are going to work on this innovation problem.
Matt Clancy (35:09): Now we've generated a pool of people who have improving mindsets and we're going to figure out how to organize the best.
Ben Yeoh (35:16): Yes. Exactly.
Matt Clancy (35:16): I preface this by saying there's a huge literature of people who this is their life's work to think about these problems. And so, I'll give you my take, but [Inaudible 35:26].
Ben Yeoh (35:28): And no consensus answer because if we'd solve this, we'd be doing it, but yes, I'd be interested in what you think.
Matt Clancy (35:34): So I think I'm really influenced by this theory called the burden of knowledge, but Ben Jones is an economist at Northwestern and he kind of argues that pushing forward innovation is sort of a general rule tends to take more knowledge than what came before. It's not universal. I think because you can have situations a computer revolution where something really new happens and you can make big contributions without having a deep well of knowledge. You can be a Zuckerberg and found a new company. You don't need to have a Ph.D. before you're having enough knowledge to push the frontier. But most technology, most of the time is not in that state, and pushing things further requires more knowledge than earlier because otherwise they would have been solved like problems that are left unsolved are the ones that we didn't have enough knowledge to solve at the time. And so that creates this problem from the management of innovation sector, mature fields that have been around a long time. They need a lot of knowledge applied to them, and knowledge can only really be applied until we invent AI by humans. And so, you have to get knowledge into people's brains and that means basically large teams of specialists. The way you deploy a lot of knowledge at a problem that needs a lot of knowledge, as you get a large team of specialists who are all really deep in one thing, and then you have to coordinate their actions. The problem there is that innovation, breakthrough innovation, I think tends to come from putting old components together in a new way or drawing a connection and seeing that some kind of theory is a useful metaphor for a domain that is different. So we can apply the lessons in this one to another. And if everybody is a specialist, they're super trained in their niche and they don't necessarily, it's harder to make these connections
Ben Yeoh (37:34): Combination technology across adjacent domains or even far away domains.
Matt Clancy (37:39): Exactly. So the way you kind of, the way that companies have tried to get around this is they encourage things like open offices and they try to get everybody talking to each other all the time. That's sort of I think of that as we've got this collective brain and we're trying to sort of share knowledge within it amongst each other.
Ben Yeoh (37:58): But it's that kind of serendipity process almost right.
Matt Clancy (38:01): We're trying to maximize encounters between people because we don't know which ones will be useful. Now I think that that process is probably wise. When we talk about agglomeration, we can talk about whether you can do that digitally versus whether you have to be in an open-plan office. But I think the point I want to make now is that it probably happens less efficiently when it's when all the knowledge is split across a bunch of people's heads. Then when it's all in one person's brain, because you've probably had this experience where you have a problem and you're just mulling it over for weeks. And that's bringing in that process, especially when you go off work and you do other things and that activates different parts of knowledge in your brain that are not, you maybe didn't think were related to your problem. And then you have the moment, maybe or something like that. And so, I think that smaller individual who knows everything who knows that all the specialists know would probably be able to come up with a more innovative answer than the team of specialists. But those kinds of people are just really expensive and really rare. And it takes a lot of time for them to train and it might be almost impossible for them to keep up with the frontier in multiple domains of knowledge. That's actually one reason I started the newsletter is because I feel this is a real growing problem that people are specialized in different domains and they don't necessarily know what's going around or going on in other domains. And so, the goal of the newsletter is actually to try and digest and synthesize this stuff. And if you see a connection, you can go read that paper maybe or so on. But there's actually a study, and it's just one study, but it's a new study that looks at this exact problem of the size of teams versus their impact and versus how disruptive their ideas are. And so, they measure these things differently.
The impact is measured by how much do you get cited? And the disruption is this index they come up with where if your idea is really disruptive, it means people cite you, but they don't cite the kind of work that you cited. You overturned all that previous knowledge and now people just start with you. Whereas if they cite you in conjunction with lots of other people that you cite, that kind of indicates, you're an incremental step in this process. And they find that small teams do a better job at creating these disruptive innovations and large teams do a better job at creating these high-impact innovations. And I think a problem that we have to deal with as our technological base gets more and more advanced and more complicated is how we continue to find these really radical breakthroughs when we need bigger and bigger teams in order to sort of advance things. And I think we're still trying to figure that out. There's some interesting work that I want to read about what happens if you embed a generalist who knows a little bit about everything with a team of specialists, maybe they can be the bridge to see the connections. But I said, I haven't gotten around to reading it, that's stuff that'll be in the newsletter. I don't know sometimes.
Ben Yeoh (41:23): Sure. And it might vary by domain you say, mature or maturing ones, we could see, one of the things, a lot about crop science, if you're working on the crop science for resistant Bali or over genes you'll need to build on a lot of knowledge. But I am still intrigued about even the size of teams because there's one theory, which I guess is stemmed because of Amazon and a few others who have this pizza box theory, which means you basically have eight to twelve people teams, but I think that's an organizational effect. Whereas if you have 300 people, they can't communicate as effectively, I guess it's that domain transfer. And also, you kind of made me think, oh, well, person X is going to do this because they thought about that. And it's not clear to me which one is actually better, except that if you have these small teams filled with what you pointed out these very knowledgeable people who know a lot are obviously key, whether they're in a part of large or part of small. And I do wonder whether it intersects with this idea of, what is it, the social weak ties idea so they can connect others because they've got enough domain knowledge to say, oh, what, Bob and Fred and Sally will need to speak because I kind of see something there, even though I might not know enough about how this works. I know enough to know that these people are there which I guess is an organizational part. But I wonder increasingly to what you alluded to, it's an increasingly important part of how we will do innovation because you need to now know so much and you need this mix of small teams, large teams' domain-specific knowledge, but then now across many domains.
Matt Clancy (43:10): And one thing I always do whenever I talk to sort of people who know a lot about computers or learn, AI or machine learning, I always say this is it, I hope this is an area that, some kind of can be solved with some kind of better artificial intelligence assistance or machine learning assistance where it sort of spots connections between fields and some is able to say, you should really talk to this guy or, there's a connection between what you work on and they work on it, even if you're not connected socially.
Ben Yeoh (43:46): Sure. We have this bunch of people and we're going to work on a problem. So, maybe … it's how to detect cancer in the blood, which is something we're working on now. What do you think the mix should be between say public funding versus private corporation incentive prizes versus say tax breaks? So maybe we can work on even bigger problems. So we're looking at some energy problems, also climate problems, multifactorial global. How do you think you would plan actual incentives, funding, and structure to get various teams, whether public or private, and what portion to work on these innovation problems? Let's assume we've got; they've got the improving mindset. So we've got the right kind of people, but should we, should we be public? Should we be private, large teams, small teams, agglomeration incentive prizes may be purchased commitments depending on what it is. Is that, is it kind of all of the above or do you feel actually from what you read, there is waiting for one or two of these types of structures and texts links that we do or I'll throw in the other one, which I think some people are talking a lot now, Patrick condoms and the others that you just need that diversity of these ideas because we're not certain what hits, but we don't have enough of all of their above.
Matt Clancy (45:14): I definitely, think that's a consensus among lots of people. Citizens, we just need to try a lot more things in a way that we measure. And one, we measure the outcomes and because I said, small changes can accumulate and so they're important to keep track of, but also if we can deploy our experiments in a rigorous way that we can make it easier to learn from them, everyone, I think I'm totally on board with that. And I really hope that we do a lot more of that but on this other question of sort of the general question of public versus private or prizes and sort of how you incentivize work on specific things. So I think that in terms of too there it really depends on your time horizon. All right? So if you're trying to solve a problem in the short term, whether that's winning a war against the Germans and the axis powers or whether that's defeating the Corona virus, then I think your high-powered incentives and specialization coordinating research works pretty well because what we're going to end up doing is we're going to take the best option that we have. That's technologically we believe is technologically feasible and we're going to rant that up really quickly. So it'll have the Manhattan project or we'll have these MRNA vaccines or something that. And, and I think that works because in a short time problem, you don't have basic research is so it takes a long time. It's really risky and you don't necessarily have the luxury of waiting. And so, it's better, in that case, to just taken the option that you have.
And you can I think for advanced purchase commitments work pretty well profit motive is going to mobilize resources that people think are likely to solve the problem pretty well. So I looked at the early days of the coronavirus, I wrote one newsletter about studies on how pharmaceutical companies respond to market demand. And essentially there's a lot of interesting little natural experiments people have looked at where for whatever reason there was an increase in demand or willingness to pay for one type of disease than another. And how does that affect the incentives to do research on that disease? And basically, pharma companies follow the profits, they do more R and B, they run more clinical trials on stuff that's going to be more profitable. But the key is that it has a big effect on third-stage clinical trials or getting a generic drug run through its safety protocols. As you go farther down the chain, it has less and less or down to basically zero effect. So when might think that, for example, with COVID-19, suddenly if you can solve this problem, you're going to be enormously wealthy. And so, you might think that's going to increase R and D across the board for everything by the same amount. But what you've tended to find was there's a huge increase in the number of people looking to see if already approved drugs also had efficacy against COVID-19. You had a lot of people working on vaccine technology, M RNA technology seems miraculous, but we were very lucky that it was already ready for prime time when the disease came. It wasn't they still had to do fundamental research to figure out how to make it work before then. But if you go back, you're not going to see people trying to understand how Corona viruses work necessarily with the hope of developing a drug in 10 years or something that.
Ben Yeoh (49:02): So it's late-stage research, but not basic research. So interesting if you apply that to something malaria or antibiotic resistance, this is why actually it may not be working because of the basic research you might need on malaria or AMR. So new set of antibiotics isn't particularly being incentivized by this type of structure. Is that am I reading?
Matt Clancy (49:29): What I'm trying to say is that I think basically incentives work really well for short-term stuff and they can pull knowledge, that's close to being ready into sort of an actual technology, but if you're not there yet, if you have a longer timeframe or if we don't know how to do something exactly. I don't think incentives actually profit incentives work very well in that case. I think in that case, you need to if you want to tackle climate change, for example, but the existing technology.
Ben Yeoh (49:57): We don't know how to do it or cement it.
Matt Clancy (49:59): We don't know how to do it. And I think that funding, there's a role for funding whatever research you think is most promising for negative carbon emissions, but there's a much, much larger role relative to the other case for just funding anything because the role of spillovers is so huge and the MRNA vaccines the company was developed and had that technology ready to go because they were developing stuff for cancer to do specific gene therapy for specific cancers. But hey, it turned out to have this role here. And I'd done some newsletters about trying to quantify how important spillovers are, which is what we call that when knowledge from another, technically just another person besides the person using the knowledge, has a benefit to somebody else. So, apple doing research that benefits Google or the NIH doing research on a cell-signaling thing that turns out to have a thing for a completely different species and it completely unexpected thing. And in the long run, those things are, I think the most important thing for driving progress forward. If they kind of speak to the fact that innovation is all about doing things we don't know how to do yet, doing new things. And if we already knew how to do them, if we already, well, this is going to eventually lead to this and that's going to eventually lead to that, then it wouldn't be innovation. The whole point is we don't know what's going to lead to what. And so, if you want to tackle any one specific thing, you should just basically fund everything. If you have a long enough time horizon, that's sort of my thought.
Ben Yeoh (51:41): That's a lot more basic research and things. And my intuition is that seems a bit, right, because, we don't really have malaria. Malaria vaccines are kind of in trial and we might have MRI and A for malaria, but arguably it's been a bit slow. And that is essential because countries that have malaria, aren't going to pay very much money for it. So there's this profit thing, even though you've got some advanced purchase because it takes you 10 or 20 years and it's super risky and there's some stuff, we don't know that doesn't overcome the fact that you could develop something for malaria or you could do flu or cancer vaccines probably equally risky within what we know, but with a much bigger profit part attached to it. I actually think incidentally on over the long run, it might prove a net slightly advantageous that MRI and A came through now because my current reading is if it wasn't for something acute, the regulators would have asked for so much data now on how much MRNA data would you have wanted before you could prove it would work for you. If they're going to ask for 10 or 20 years of data that would have actually set back everything you can use MRNA for, which will actually now include cancer, malaria, and all of this would have much preferred not to have a pandemic and have a, maybe more entrepreneurial regulator, but it doesn't look, look that way, which might kind of thinks that then we thought about the corporate lab a little bit and this need for teams and specialists and things that.
Matt Clancy (53:19): And I just interrupt really quick to say the big thing is, if you also have this really long run perspective where you wanting to maximize spillovers, then I think that says a bigger role for things public agencies and things that involve the free sharing of knowledge. So maybe prizes rather than intellectual property rights, because you want to maximize people using other people's knowledge. And when you're looking at a short run, pulling something, that's close to prime time into the market, then you can start using targeted profit motive stuff because I think that works really well anyway.
Ben Yeoh (53:53): I agree if you're thinking about, say global inequity or even globalized growth, you can see it too. They could have got then actually tech transfer from richer countries to poor countries in the areas where they could do with that tech transfer actually raises everything. You diminish the inequality and you graze global growth with that would be much better suited to a public agency to do those sorts of spillovers than private which actually brings us to. So with this idea, you've got obviously us ARPA, DARPA we might have a health ARPA I don't know, you probably haven't seen, but there's been a French paper out by bland shard and Tyrol suggesting a French or European ARPA, EU ARPA. The UK has this opera called Aria. So everyone wants their own little mini-innovation agency. But if you think about the UK ARPA which I think we're calling aria, it's going to have about 500 million pounds. But it's going to need an executive director and is going to need, it doesn't really act. So there's not really looking to be missioned aligned, at least not yet. Maybe it will be, but you've been put in charge. You're now going to be the executive director or the person who appoints the executive director. What should the UK innovation agency be looking at? I mean, I guess, is it funding everything, should it, and is there a particular structure organization? How should we be thinking about innovation through that? You can choose us health ARPA fit seems the same for that.
Matt Clancy (55:35): I think the lack of a mission focus is in theory that's fine because you can just pursue whatever maximizes the public good at the best. But I think that having a mission sort of does a nice job of positioning and agency and helping it, it seems it helps coordinate it because then you draw related types of people together who have related but not over. And you have a nice metric for success and then make a nice justification for ongoing funding in the long run. Right? So if you're not going to do that, then if you're going to fund basic research, for example, basic research takes a really long time for the fruits to become clear and even, or to emerge. And they may never be clearly tracked back. So the political economy of keeping an organization going I think without a mission is sort of tricky, but in general, for all these ARPAS my position is it's good to do whatever we want. We should be spending a lot more as a society on R and D than we do. The returns are really high when you look and think of them in social terms. So we should, we could probably easily double or triple the amount we spend on R and D before we would have to worry about whether we're [Cross-Talk 57:07] wasting money. But I don't when I criticize them, I don't want to be, well, so we should just close that down. But I do think that there's sort of an excessive hope and faith being put on the ARPA model right now.
The evidence-based that we have to draw on is for two reasons, one is we've got the old ones, DARPA, we can talk about them in another, in a minute, but then the newer ones, it takes a while to figure out if these things have worked. And especially if their goal is sort of breakthrough innovation, then you need a lot of data. Because if you only need one hit every decade, then you might need decades of data before you can tell if it's a good investment, but with DARPA I'm always wondered, all right there's a couple of things that I wonder, do they make it seem it was more innovative than it really was? And then, so we were sort of doing a cargo cult thing where we're taking the form of DARPA and just assuming we'll get the same results. So one is they just spent a lot of money and maybe that was the secret sauce is just spending a lot of money, two they were able to recruit the best scientists who were really motivated by the mission of DARPA. They really believed in this survival of the United States. They were people who the people who worked for DARPA were really believed in their mission. And so, you have really smart people motivated with resources. And I think that in that kind of setting, giving them more free reign to use their judgment and not tie them up in grant reports and stuff. That sort of makes sense because they really have sort of this altruistic at least from their perspective motive, they're not going to run away with the funding, or wasn’t it.
They're really concerned about this and then the resources, and then they've got the talent. So you're going to get great results. If you don't have a mission or whatever, maybe that's harder to achieve. And then the other thing about DARPA is because they were a defense agency, they were able to keep everything classified. So all the failures we don't necessarily hear about. So there's a bit of a selection bias. And then the third thing is since everything is classified, they can work on something for decades. And then when it finally works, it sort of appears to us, XD hello, out of nothing. And it's oh my God look what they did. But if [Cross-Talk 59:42].
Ben Yeoh (59:42): It took them 27 years.
Matt Clancy (59:44): If they had not been a classified agency, other people would have been following it along and then somebody else would have picked up the so I read this one book, the Pentagon's brain and talk about how they were working on drones since the Vietnam war. And if all that stuff had been open and out there maybe some other company would've picked up the drone technology and advanced it a little bit and somebody would have been an ecosystem of people doing it. And we would have got there earlier, but it would have emerged more incrementally instead of being seen as this amazing reveal by DARPA on this particular day. And then I think the other thing DARPA had going for it was that because it was just working on different stuff than everybody else, they had a different mission. And so it was kind of working on stuff that nobody else was working on or not giving the same resources to. And so that generates also the potential for surprises just by being funding, weird stuff. Because they're not pursuing their profit motive. They're not doing basic science. They're doing some kind of orthogonal to both. And since you never know, what's going to work trying something that nobody else is trying is occasionally going to deliver these big successes that everyone's going to be really surprised by because nobody else is looking at them. So all of those factors, I feel don't necessarily hinge on the product manager with a lot of freedom to operate. And so that's kind of my hesitancy about are we putting too much faith in these things because maybe what made them work so well was a little bit idiosyncratic and is a little bit hard to replicate.
Ben Yeoh (01:00:20): These are the facts. So I agree and I think I would even put it more strongly that we are under-invested in innovation across all of these things. I would be going much, much higher given the scale of the challenges and returns to all of these aspects. And that is because I partly do think we don't capture some of the value, I said, on pain, that’s not captured within it. And I think you will see those benefits and I'm not utterly convinced that DARPA was necessarily more successful than say US venture capital. So it's the same sort of thing. But to your point, they were working on other ideas, and maybe they had a longer time horizon.
Matt Clancy (01:01:58): I would agree with that.
Ben Yeoh (01:02:01): But they did produce some other good stuff. So I just think, why don't we replicate that, but it might not have been anything special about DARPA. And I haven't heard your point about the kind of secrecy non-spillover effects on that. But I'm reflecting that quite be quite a good thing. I was always taken by two instances in deep pattern history. So one is on Japanese semiconductors where you might be aware that the Japanese patent office held up IBM semiconductor pattern in order for its local industry to get a boost up. And they really did so you, so you can see that. And actually, even going back further in time into the 18 hundred, there's a lot of work on how the dye industry for colors in say France, England, Germany around that time what patterns and things we're doing around there. And in fact, a lot of people moved to Basel in Switzerland and why we've got such strong pharmaceutical companies in Basel in Switzerland is because the dye manufacturers at the time moved over from France to avoid patents so that they could do more work. It was processed at the time and things. So I do wonder about that interplay between how long we should give patents for incentives versus the spillover effects that you get when you've got new ideas coming through. So I guess that, go on.
Matt Clancy (01:03:31): I mean, that's a classic problem. And there's this old Quip about if we were deciding today whether to start a patent system and we didn't have one, we wouldn't have enough information. We wouldn't have enough evidence to suggest that it's a good idea to one, but since we do have a patent system, we don't have enough evidence to argue, to shut it down. So might as well keep it going.
Ben Yeoh (01:03:54): My personal view is actually patented for certain things are too long and certain things that you short, but because only I can judge that on an individual basis and I feel I have that amongst certain drugs and software things, because I can see there's much more value in one and the other and then centers needed, but I can see how the system could no way ever be able to cope with differential pattern life, some things I think, that was probably only worth two years. But the incentives for new Alzheimer's, I'll give you 25 years if you can crack that because that's going to be so super valuable whereas a new type of headache pill is probably at this point in time not going to be that valuable.
Matt Clancy (01:04:36): I completely agree if we knew a good way to get to instruct a bureaucracy to implement that kind of thing, that would be ideal, but we don't know how-to, we haven't figured out a way to do it, that it wouldn't be captured or something.
Ben Yeoh (01:04:52): Exactly. I guess I have two questions that come to mind. One is in hiring people, how much do you think you need formal education for this? So this is kind of the specialized thing. And then the second nicely is an intersection of what we just said is do you think, so Anton house thinks copyright should be reformed. There's kind of a lot of copyright thinkers who think that at least copyright's a little bit too long because of that balance. And I was wondering if you had any views on patents, although I guess we've said that, that it's probably, we are where we are. We don't really have evidence one way or the other that may be tiring to the hiring. So there is some evidence, at least in history that formal education wasn't really a barrier to the invention. And in fact, lack of formal education might have led you to more invention, but more recently, to your point, it does seem that you need a little bit more specialized knowledge in certain areas. But you say, you do need patchouli this sensitivity to cross-domain ideas, which is sort of they're not formal education in whatever it is this other domain is. So I was wondering what you think about the level of formal education that you might need.
Matt Clancy (01:06:05): I think that it goes, it's related to this burden of knowledge thing in some fields you need a lot of knowledge to make meaningful contributions. And I think that sometimes I get frustrated by people who I think draw too many lessons from what we did in the distant past distant, but just in the 18 hundred or first half of the 20th century and this sort of, oh, look at all these tinkerers, these lone inventors they were. And I think, well, I think the nature of innovation for a lot of industries is just changed since then. It was possible in some days to just without formal education or without advanced degrees to just understand the mechanical system, for example. And I've really gone to these when I used to live in London, no, it was in DC. They had this thing on the longitude prize where they made these clocks to tickets see, to measure longitude. And I love those things because they're right, they're so intricate, but you can still look at them and piece out exactly how it's working and understand how the pieces perform their function. And a lot of things are not that anymore where somebody can just come in and figure out what's going on and make a meaningful contribution. There are still are fields that. So in that sense, I think that formal education is probably more important now than it was in the past that said on the other side of the equation, there's a selection effect where the kind of person who has the improving mentality or is going to want to start a company, develop new technology.
They may be, selected out of pursuing enough formal education and the kind of person who can finish, it could be self-denying enough to do they're under that they're masters to Ph.D. and two rounds of post-docs and then apply for all the grounds, play the game long enough so that you're in a position to affect change may not be the kind of person who is, there may be a very rare type of person who can do that. And also advance it, be an entrepreneur is going to advance things. So when I listened to your podcast with Anton, I remember him saying, you can have teams basically, you don't have to have, you might need some specialists with formal education and you might be able to have some other people who don't. I think that's a pretty reasonable compromise. I mean that you can have both and at the end of the day, it's not necessarily the credential that matters. It's just that you do need a lot of knowledge to advance some of these domains. And sometimes you can teach yourself and sometimes you can get an experience and work in an environment where you can learn that without needing the credential. But I think that a very common mainstream way to get the knowledge you need to solve major problems is through school.
Ben Yeoh (01:09:03): So ideally you need both an improving mindset and a reasonable degree of knowledge, if not a whole lot of knowledge, but you might not all be in the same person at the same time. Great. If you have, if you have it in your team, that might be okay too. Excellent. And then would you reform copyright or patent? Or both.
Matt Clancy (01:09:22): I mean, for copyright, I think copyright is outrageous. It's a scandal how long it is, 70 years past that, it used to be 28 years and then you could renew it once to 56. I thought that that seemed to reason. So the issue is copyright lasts right now an enormous amount of time, I think, an insane amount of time. But a patent lasts 20 years, you might ask well, why is one of them so much shorter than the other? And I think that copyright is a weaker form of protection in the sense that it protects not a patent protects sort of the idea itself. I can't just take your same thing and make some small changes and, and, I'm still violating your patent, but with copyright, I can take the idea of a boy wizard who has to fight a guy who stole some of his soul and goes to a wizard school. I can take that idea. And as long as that could be the core thing that makes that series of books popular. And I can just steal that as long as I'm not calling my characters, Harry Potter and not making them super identifiable is supposed to be the same thing. So in that sense, copyright is a weaker form of protection. So it makes sense that you get a longer period to capitalize. But I still think that the current length is crazy and is driven not by. So I remember I did this calculation once where say you're at, I can't remember what I said, say you're at 20 years and you're going to extend it to 50. If you discount the future value of money at 5% per year, what's the difference between that, as you extend it, you only get the present value of an income stream that says your book is going to sell $10,000 a year in perpetuity, the present this value at 5% is I don't know. I think it's 500,000 or something. I can't remember.
I'm doing my math wrong, sorry. But anyway, if you only have it for 9 for 20 years versus 50 years is a very small difference and so the person who it really matters to is somebody whose copyright is about to run out for them. There's an enormous difference between zero more years of protection and another 15 years of protection. And so they have an intense incentive to lobby to extend copyright, and they've been successfully doing that around the world. And the problem is that it has no impact or such a small impact on the reason we have copyright, which is to motivate people to create new things because of all that. And I think it just locks up ideas that would be profitable, recycled and remixed, and stuff like that. So I would love to go back to drastically shorter copyright I don't know, 50 years at the max. And I would, I don't know, I could be convinced that's something that is appropriate, but my bias is towards something that's more the length of the pad as for patents, I don't have as much to say, I think it's a lot fuzzier what the appropriate length is. We've already kind of alluded to some of these things. I think that they do lock up ideas. I think that a lot of patents are a net drag on innovation rather than a benefit, but there are some domains where patents are really important and innovations probably wouldn't happen in some kinds of fields without patent protection.
Ben Yeoh (01:13:04): No, I agree. I think on copyright special interests, essentially publishers and the maybe artists in the top note, 0.1% have this laptop, the average says artist or writer, it does seem to be harmed by it. And then you've got these orphan works and all of this, and it hasn't kept up to date with technology, the way that all I think that's all right, but I am intrigued on patterns because actually if you go back before we had tripped and all of the patent system, much wider variety on pattern length terms. And you say, actually in hindsight, you can identify what maybe deserves longer and deserve shorter, but I don't know of any bureaucracy, which could, which could handle that.
Matt Clancy (01:13:53): Yes, I mean, because, you would want to have it be based on something, you want to provide sufficient incentives. So, you can have longer copyright if you need more money if you need more time to recoup your investment at a reasonable rate or something that. But who's going to tell you the honest length of time it took them. How are you going to monitor that?
Ben Yeoh (01:14:12): And exactly, I mean, I spent quite a long time looking at biopharmaceutical patents, and there's maybe there's any handful of us in the world. You can really look at and assess that you've got process patterns, substance patterns, and crystallization patterns. Then you look at health economics. Was it incremental? Did they really do it? Was it a family? Were they building on other classes and things and all of this knowledge and we judge about whether it's valid or not, or how much prior art and things. ? And this is the whole specialist knowledge that's maybe for each individual drug class has maybe a thousand people who really understand what that value is and they can reasonably disagree within some balance.
Matt Clancy (01:14:50): So, this is a good point for me to ask you a question then. So, if I was asked what's an area that really seems to depend on pants? I would have said pharmaceutical. Innovation pharmaceutical seems one of the best arguments you could make for the value of patents. And I would've said it's because most of the value created by a drug is the knowledge that it is safe and effective, not necessarily how you manufacture it. And so, it's hard to protect that knowledge. Once it's out there you need a patent to protect it because otherwise, everyone does oh, this compound, we now know cures headaches, and doesn't cause any side effects so we can make it and sell it too. But since you're a person who reads patents and pharmaceuticals and this whole, what is your take on the value of patents in medicine?
Ben Yeoh (01:15:38): So, I agree. There are quite a few complexities. One of the most obvious ones actually is that roughly 20 years living on your pattern from when you discover, say the white powder drug or say the MRNA formulation, but on average, it's taking you eight to 12 years to get it through regulatory development. So, your effective commercial patent life on average is getting can vary. It's actually only 10 years. So even when you compare it to the software family patterns, because your development cycles much quicker, you've got way less time. And if you compare it to say software patterns, you've got this kind of a family pattern and you build on that and you can kind of keep hold of that as the large tech companies have almost in perpetuity, right? Which would you call them do within pharma, at least with white powder chemicals, slightly different for biological, but for white powder, once you have the chemical structure and that's known, and then your pattern is gone, it can be copied for you say, the actual chemical manufacturing is extremely cheap, quite cost you a few cents for something then that is not where the value accretion has occurred?
But then because it's only actually 10 years commercial life, you have all of these, the stranger effects of where you try and play essentially patent games to get more life because, because of that value. But fundamentally if you could, you can tell what we would what I would call sort of unmet medical need. Or you could translate that into health economics, where you could see that if you had a cure for cancer Alzheimer's, or malaria, that is more valuable than a headache pill today. So arguably I can give you a 30-year patent and say you've got a cure for Alzheimer's. I can give you a 30-year patent and the NPV, the net present value to humanity is still going to be enormously large, whereas headache pills, given what we have and where we are today, I'll give you two or three years to do something.
If you could do some incrementally better, but quite frankly, you're not that much better than ibuprofen or aspirin or something else. So actually, I don't really need to incentivize you to do anything else because society has something decent, but we have nothing for Alzheimer's is nothing for malaria. And in fact, you can go on because there's quite a lot of health care needs. And so, I would want to incentivize you more. Yes, you can do this maybe with some advanced purchase commitments and other things, but they are way not as good as patents to do with the fact that you need also them to do. As you alluded to earlier, the basic research stuff in some of these areas, particularly in say the brain sciences, Alzheimer's schizophrenia, depression, and you can go on and on where we still don't understand that basic research and some of that's coming out of universities and other publicly funded, but some of that could come really nicely out of corporate labs if they had the incentive to be able to do that because of the long time horizon and also an incredible which people don't understand.
So again, 10 years that it takes you when you discover the white powder drug and before you've put it in humans, your average chance of success is around about 1%. So, you've only got a one in a hundred chance and it's going to cost you anywhere between 500 million to 2 billion to develop that. So, this is why also it probably doesn't suit governments because governments can't afford to spend a billion on something and turn to the public and go, well, you know what? We knew as a one in a hundred chance, the MPV would be great if it was outsiders, but you know what, we, we wasted that billion. We did get some spillovers, but you can't count that. And so again, patents uniquely do that.
Matt Clancy (01:19:17): That reminds me so there's this, the typical view of why we have patents is you do all this R and D, somebody can copy you. And so, you need protection to charge a price above costs for a little bit of time, make a profit. … To make a profit, to compensate you for the R and D you did it. They call this the reward function of patents, but there's this other perspective by a guy named (Edmund Kitch) that he calls the prospect function of patents. And that's sort of what you were saying, where one of the roles of pens is actually to carve out space for an organization to develop a product that actually isn't ready to sell yet knowing that.
They can kind of work freely in this domain without worrying that somebody going to take their ideas and also to prevent the duplication of effort. So, for example, if we gave Alzheimer's patent for, you said, maybe 40 years to some company, and that company has a long-time horizon that it knows all research that it does relate to the mechanism that it has the patent for is it's to work on. And so, it's kind of the likens it to giving a land claim for a mine, the person who owns the mind has the incentive to develop it and go down and dig and do all that stuff because they know that they get to own, whatever comes out of that. And by giving a patent to an organization, to a specific pharma company, we prevent duplicative wasteful effort, people trying to do research on the same mechanism at the same time, they can do other things, knowing that that company is doing its own thing. And so, I've heard people say that that function of patents is actually pretty, it works pretty well still or coordinating research. So, what those guys think.
Ben Yeoh (01:21:08): So, it works really well and pharmaceuticals to take the Alzheimer's example. So, you've got something called amyloid-beta, which is one big hypothesis within Alzheimer's. So, say you locked up the Abita hypothesis or other people could do other APETA mechanisms. But you were given a 40-year claim on that. A lot of people would then look at outsiders, but look at another mechanism that wouldn't compete within that class. But you know that Alzheimer's is really valuable. There's actually a very current debate because USFDA has approved an Alzheimer's drug, which scientists debate its value for in that particular drug. But it's really interesting that you can see it has very much excited companies and investors for follow on drugs. So, the actual benefit to society is at least in my view, unlikely to be in this particular drug, which has got pros and cons, but with the follow-up incentives, because people look on, what I'm not canceling my face to 50 50 Alzheimer's program, I'm going to invest in it more because I naturally now know that the regulatory barrier to this was not as high as I perhaps thought which is a kind of interesting thing.
The flip side saying that maybe that mechanism of action is not the one to go for. So, this brings me to we can do a little bit in honor of emergent ventures and Tyler Cowen, a game of underrated overrated. The little trick here is what we do is what you might've heard this on one of the others one is what may be phantom Tyler might think so you can try whether you think it's underrated, overrated. And then what do you think that Tyler would think? So, this is the idea that we don't actually need the real version of Tyler. We can just have the version of Tyler in our head, and he might scale maybe a little bit better than you would have thought he doesn't, he can't actually be everywhere at once.
Matt Clancy (01:23:07): All right. Exciting to get to play overrated underrated. I never got to play this one.
Ben Yeoh (01:23:11): Let’s start with agglomeration effects. So maybe city agglomeration effects. Do you think overrated, underrated?
Matt Clancy (01:23:21): I am, I totally think these are overrated. This is my whole thing. That doesn't mean that I think that they're zero or that.
Ben Yeoh (01:23:29): They're are just overrated.
Matt Clancy (01:23:31): I think they're overrated though. I think though that the world is catching up a bit with having seen that remote work functions, not just for six weeks, but for over a year without notable huge downsides or anything. And I think that the thing that's interesting is you can see this, I had written a paper trying to synthesize a lot of stuff about remote work. And one of them was about knowledge spillovers and very few papers try to measure how knowledge spillovers, which is people learning from other people who are geographically close from them, but maybe not in the same firm or, or whatever. Very few papers try to have tried to measure how that has changed over time, because there's just kind of been this assumption that, well, what's interesting is what is it? And then we just kind of assume that it stays that for decades. But when papers do try to measure how it's changed over time, they tend to find that it's been declining for decades. The propensity to site patents that are local declining, the propensity to incorporate words for new technologies used to be much higher in big cities rather than media in cities. Other stuff, I don't know, you can look at academics, citing papers that are close or far away in their department. The value they get from moving to a high department has been in terms of their own personal productivity has been declining.
I think there's lots of evidence that the internet and travel have basically eroded the value of being physically co-located in the same city as other knowledge workers. They haven't eliminated it, but it is a lot less than people think. And the reason that's important is that there are countervailing benefits to not being agglomerated that when the benefits of agglomeration begin to fall, those other benefits may start to become dominant. For example, the ability to pick just the right person for the position the right specialist, if you're not limited to your local labor market, I think becomes even more important as this burden of knowledge problem happens, I need to get just the right person who knows just about the right kind of thing. And they may not live in my city. And if we don't benefit that much from being physically co-located, it might be a lot better for me to get the right person rather than the person who I could work with close. So that's my little rant on agglomeration. Great. So, you'd
Ben Yeoh (01:26:06): What would tie say? Well, I guess that means you're then cautiously pretty optimistic about remote working and probably all of the fasts on innovation clusters. It's a thing, but actually, it's of a diminishing return. I don't know what would Tyler say.
Matt Clancy (01:26:21): I don't know what Tyler was saying, I don't know if he's written about agglomeration effects before. I'm trying to think. But I think that he would say they are either, I think he's maybe opposite from me or at least thinks that the conventional wisdom is correct. He talks about the Vienna circle and sort of the value of these highly innovative cities and trying to figure out what made them special. And so, he might be more cautious than me about.
Ben Yeoh (01:26:48): I know he's very pro traveled, but I'm not sure about, I'm not sure about the collaboration. Well, we'll see. Go for GDP as a measure.
Matt Clancy (01:27:03): GDP as a measure, I think I've got the boring consensus. You have an economist, which is terrible, but except for everything else, it's correlated with lots of stuff that people do care about. And so that should give us some confidence in it. It misses a lot. I think if you're aware of what it's missing, you wouldn't want to fetishize it where you start to focus only on it and ignore everything else. But it's very incorporating that other stuff just added extra layers of complexity and subjectivity. If you're trying to incorporate measures of wellbeing, there is some kind of that reminds me of a quote from there's an innovation economist, Alberto Galasso. And I don't think you'll get mad if I share it, but I was asking him once, what do you think about patents as a measure of innovation? Because this is a big debate among economists who study innovation, that patents are this terrible. Are they good? Are they good enough? Is there a super misleading measure of innovation? And he was sort of, this really frustrates me because, nobody bothers all the economists who use GDP, but it's not that's perfect. So, we always get this question, but I'm not sure those other guys do anyway. Is it overrated or underrated? I'd say it's correctly rated. Most people have that view.
Ben Yeoh (01:28:33): That’s a fact.
Matt Clancy (01:28:33): And what would Tyler say. I think he would, I think echo a lot of the same concerns. He's got his wealth plus concept in stubborn attachments. So that's, I think he would probably be, have a similar take that's my guess.
Ben Yeoh (01:28:50): I'd agree. Although I do think he notes that GDP does correlate with so many things that we do actually think are great. So, he would probably say, show me something better, which I guess is the same with TFP, I guess is the same with democracy. Right? Show me a better system. We know it's pretty imperfect, but until I see something better, I'm going to.
Matt Clancy (01:29:09): I don't know if there's not much appetite at the moment for overturning it and changing it to something else., I don't know. I haven't heard a lot of calls.
Ben Yeoh (01:29:19): There is a New Zealand living budget, which I think is the cooker that you're going to get to look at something which is GDP plus, which is being done seriously by a country and looked at by that government economists and things to look at that. But essentially, it's what I say, essentially it is trying to put natural capital and human capital elements and how they think about it. So that would be a step on, but it's not, completely different to what, to what you would think. Over genes or what you guys would call eggplant particularly maybe with OB-GYN yields or crop yields.
Matt Clancy (01:30:02): I don't know that much about it, I know that there was, I believe there's a famous paper about GMO over genes that found that they were the GMOs was spectacular for a developing country context. And so that's good in general. I'm pretty happy with GMOs as for Aboriginals. I think that they're probably a little underrated, at least in the USA, there just as a food. They not part of the staple of any main meals are prepared.
Ben Yeoh (01:30:43): Pretty good. And I get it, it'd be pretty good for yields as well. I don't know what Tyler would think about everything.
Matt Clancy (01:30:51): It's hard for me to guess because he's having too many sophisticated views on the food, I guess.
Ben Yeoh (01:30:57): I would go underrated because I think he thinks a lot of food, in general, is underrated. NFTS or non-fungible tokens. This is blockchain-type thinking.
Matt Clancy (01:31:14): This is tricky, there's obviously with all crypto things, there's this huge polarization where it's probably massively overrated by people in the committee or at least overrated to some extent by people. I don't want to say massive necessarily. And then maybe underrated by other people. It's something because there's this huge polarization of people who think that it's just, I don't agree with people who think that they're just a STAM, that there's nothing to it. But also, I guess one view I have on crypto is that I think people who are crypto enthusiasts maybe underrate how long it's going to take these ideas to become mainstream, have a significant impact. New technologies just tend to take decades to wind through an economy and for people to figure out how best to deploy and use them.
And over that timeframe can be really significant. And we've kind of got a weird case where remote work has leaped forward probably several years in the future because we had these crazy events where we all had to use this new technology. If we had done in the 18 hundred that there had been some disease that came from steam engines, but not from electricity or something, how that would have changed, accelerated the transition to electrification or whatever. And that's an outlier, how quickly people are adopting remote work and moving to these remote-first companies. I think that's a really weird thing. And other technologies NFTs are going to take longer, but there's this. I read this newsletter from a guy packing McCormick and he used an NFT to try to allocate revenue from a product from a newsletter. He wrote to people which I thought was an interesting idea.
Ben Yeoh (01:33:11): What do you think Phantom Tyler would say?
Matt Clancy (01:33:15): I think, so I think he's often said, crypto stuff he's waiting for a really good use case for them. And this being a potential use case maybe he thinks that this is a good thing, on the other hand maybe he wants to do with his other crypto usual concerns. So maybe he thinks that they're slightly underrated, but I could be, I have low confidence in that.
Ben Yeoh (01:33:49): I think you'd think they are underrated, but I agree with you that they probably are overrated if you are a guy with laser eyes, but actually for the average person in the street actually don't really touch crypto. So, for them, it is underrated. And I do think NFTs do have a genuine use case. I don't know how large it will be, but it is. I think it is genuine. Maybe the last one on the overrated, underrated science fiction.
Matt Clancy (01:34:17): Oh, science fiction. I think I think for it to be underrated, it would have to mean that people who would like science fiction are not reading, are avoiding it because they underrate it. And if they just gave it a try, they would really like it. And there's some newer science fiction. There's like newer, maybe it's not new, but there's science fiction that's being done by authors who do not traditionally are not story genre writers and they've started dabbling in it. And I think that stuff appeals to people who don't traditionally science fiction. But I do think that people who think of hardcore genre science fiction, which I like, I should say, I'm organizing an econ, Twitter, science fiction reading thing and we're reading a deepness in the sky, a book by burner Benjie, which is amazing. And everyone should read it who is interested in this kind of thing. Because that's the thing. I think that a lot of people who think they wouldn't science fiction, I mean obviously if they could get the right book point, but if they picked something off the shelf, I wouldn't be surprised if they are, I don't like it because it does have genre science fiction is very preoccupied with certain kinds of things that if you're the kind of person who's also preoccupied with them that are fantastic. But if you correctly understand that you're not, you're probably not going to like it that much deepness in the sky is very great if you're interested and how very long run science and technology, how that's going to look for the human race or kind of economics has innovation questions and stuff that. But if do you think that's all boring and you just want a story about somebody's internal journey, that's not the place to go.
Ben Yeoh (01:36:25): Surely that means it's underrated because more people should be thinking about the very long term and innovation question.
Matt Clancy (01:36:30): That's a good point. That's a good argument to say that it would be good for more people to read optimistic science fiction or at least I would say I think that's probably true. And I think that it would probably be good if we could make more high-quality optimistic science fiction television, that will probably have a bigger impact than more reading.
Ben Yeoh (01:36:53): And Phantom Tyler.
Matt Clancy (01:36:58): I don't know, Tyler, it's really hard to guess Tyler's reading.
Ben Yeoh (01:37:02): You think.
Matt Clancy (01:37:05): But I think for the reasons you said he might be optimist; I agree with what you said
Ben Yeoh (01:37:09): I'm a strong conviction on science-fiction he would definitely think science fiction is underrated for all of those reasons. And also, because I think he is a science fiction reader, but that's my guest.
Matt Clancy (01:37:23): Cool, cool.
Ben Yeoh (01:37:24): So maybe turning to the slightly personal career journey is what do you think maybe physics and religious studies have informed you to do anything with economics or also because I know you studied some of that, this idea that you can, I guess there's a kind of liberal arts thing, but this idea that you can swap specialisms or have things across domain has that been useful to you?
Matt Clancy (01:37:52): I when I was in high school, I was really interested in world religions basically, what do all the different religions and sort of this view that all the different religions are sort of trying to find the same core truth and they're all approaching it from different angles. And so that was a question I really wanted to pursue in college. And I went and did religious studies but my parents were very wise, but this is a poor career strategy. And they encouraged me to do something quantitative. And so, I compromised by doing a double degree in religious studies and physics, because I figured this would be a pincher, a talk on truth anyway, we'd learn about the nature of the universe from both sides. And I think that that experience definitely informed who I am I'm not particularly religious right now or a physics person. So, I don't know, read into that as you were, but what did I take away in terms of economics? The reason I went into economics is I didn't really know anything about what economics as a field was. When I was an undergrad, I didn't know any economists really, and I never read popular economics books. It was just something that existed on the wall street journal page or something. And I read Atlas shrugged and guns, germs, and steel when I was a junior, trying to figure out what I was going to do. Because I had decided I wasn't going to be a religious studies scholar or a physics scholar.
And it's not I buy and rants objectivism or anything that, but it was the first time I'd been exposed to thinking of economics as this sort of engine that runs the world as it's super important. And then guns, germs, and steel wire, you can study big, important questions about society, very quantitatively. And so religious studies are the study of world religions and we read a lot of texts, but it's, in some ways it's not a quantity. It wasn't a quantitative field as I was taught, physics was super quantitative. And realizing that these two things could kindly go together and that they might go together in the field of economics, was what brought me to that field. It's an important thing and it's quantitative and rigorous, not that other stuff can't be rigorous too, but that was my bias, I guess. And that's what led me to choose to become an economist. And I think that's all been proven correct. What do I still take away from it? I mean I still read pop physics books. I try to keep up with physics and I still do think about very ultimate questions. It's not I read about them on my sub stack, but I don't know if they're better than anybody else's, but I still think about those things. I think they give you a useful big picture about what do you want to do with your career and your life?
Ben Yeoh (01:41:02): And I think you studied in London and Cambridge. How does that compare to Iowa or even the US versus the UK, but there is something specific about it, I see the London, Cambridge, Oxford Southeast is probably considered the only UK agglomeration thing on that, but very different Cambridge, I guess.
Matt Clancy (01:41:27): They were very different. I mean, how I got into it was I knew that I wanted to study economics. I had no training in economics except for one class that I took and an online class that I took and books that I read. And the University of Cambridge offers this program called the diploma and economics, which is designed for people in that position. It's a compressed one-year undergraduate education in economics for people who have an undergraduate degree in something else. And then the goal is that it's the first year, then the next year you can go on and get a master's or something. So, I did my master's at the London school of economics, and I worked as an economic analyst. So that's, that's what took me to the UK is just to, did this degree.
But also, I had studied abroad in Ireland and really enjoyed the experience of living somewhere besides the United States. And so, I was really open to trying something different. It’s a very different system. The US system is very piecemeal or spaced out. We have lots of little assignments that are evaluated and count towards your grade throughout the entire semester. And so, the final is maybe it wouldn't be uncommon for the final exam to be, 15% of your total mark and everything else is from midterm exams and homework that you did and participation and a bunch of other, a huge mix of stuff. And then both universities of Cambridge and London School of economics, that was all just an end-of-year exam was pretty much the entirely everything. And so, the first year at Cambridge, I didn't do particularly, I did fine, but I didn't do great on my exams because I had no idea how to sort of study for these kinds of high stakes, everything exams.
And so that was a learning curve, but I think I got over from London school of economics, I figured out how to productively study for that kind of thing. But then the funny thing is, so now I teach at Iowa state university and I have slowly gone from a more UK style to a more US-style and how I teach my own classes. Because the thing I about the UK style is it's very I don't know, there is a moment in time when you sort of, in theory, know everything that, you studied in that year, whereas in the US you can count on things to kind of filter in one ear and out the other, they go on the exam and, maybe you suffer a penalty because, at the end of the semester, you can't remember, but that's only 10% of your grade or something that. Or you cram and you get a little bit, but with the UK system, you really know everything at that moment. And it's a nice system, all right, at least at this one moment in time, this is a good measure of what this person learned over the year. I tried to do a system like that in the US where most of the exams for almost all your weight, and then you could get some extra credit from doing the homework as a way to incentivize people, to keep doing the homework in a culture where people are not used to the idea that if they haven't had the experience of not doing it and then getting burned at the end of the semester, and it didn't work at all, [Inaudible1:00:44:28] was too hard, the culture, the expectation of students is just too different, in a class where they're not being continually evaluated on things they have opportunity costs and there are other classes where they are being evaluated on stuff at the moment. So the stuff in the class that it's only the final gets shunted down the way. And it just, and then they don't have experience with knowing how badly that will burn them. And so, then I didn't, I don't do that anymore, but I still think that there's merit. I think there's merit to it, but it's also very high stress and high stakes.
Ben Yeoh (01:45:10): It's different, but I do think, I think there are pros and cons about the UK system is because you do need to basically assess yourself whether you think, something all the way along the year to get to that endpoint. And actually, that sometimes reflects things in companies because depending on the companies or even whatever walk in life is you need actually a particularly brilliant manager or whoever's above you to be able to give you constant feedback, to tell you that you're doing. You could do this project this month, a week. Usually, the average worker doesn't get told anything and they might have a big project or not, but generally, you don't know what you're doing. Maybe you have an annual review. It's more companies that really even have annual reviews. So, unless you've got a built-up some way of knowing how you're doing, you actually then tend not to do that well, and that's something you have to do on the job. On the other hand, you’re having the big project at the end that doesn't always the same as all companies, but that sense of being your own self-evaluator as opposed to I just took a quiz every week. So, I knew and had to do it is a kind of different way of thinking, which is interesting. I don't know which ones better.
Matt Clancy (01:46:20): I that the UK system also builds in autonomy and you have to self-motivate and you have to decide on your own study. And I think those are good skills to learn for life. Whereas the US system can just follow the plan laid out for you by the instructor. And then when you graduated and you don't have a plan anymore, you might in [Inaudible 01:46:43].
Ben Yeoh (01:46:45): Great. A couple of last things, I think one I'd love to hear you on about remote work. But maybe before that, because I know you studied a little bit in Sub-Sahara Africa, also some countries there I've been interested. Do you think there's anything we particularly misunderstand about Sub-Sahara Africa? I'm sure there are lots but something which you kind of see, if you could understand this it would just give you a lot better insight because I guess most people don't.
Matt Clancy (01:47:15): It's tough. I mean, for Sub-Saharan Africa, that was my first job. And it was kind of assigned to me. I went to work for this company to be an economic analyst and they were, you will have these countries. And so, I didn't, have a pre-existing knowledge or expertise in them. And I was always nervous about, if I had sufficient depth of knowledge over this thing, I think one big thing is just that, it's an enormous diverse place. Africa is not a country, its many countries, it's enormous. It has that probably one big misconception or I mean people know that I suppose, but they don't realize it and they think it's just sort of Africa. But that all leave it at that because, to be honest, this company never even paid for me to visit. And I don't know. I don't want to speak out of term basically.
Ben Yeoh (01:48:24): And so remote work, so you've done some interesting stuff, actually, what you've got to find off a separate blog, because it's not all on your substack, but your I get the sense, pretty cautiously optimistic about both the emerging work and what it might mean for people. So, I don't know whether you want to sum up what your current thinking is about how remote work might work in the future for us?
Matt Clancy (01:48:48): I've been working on remote work for a couple of years before the Corona virus. And it started because I live in Iowa, which is in central United States. It's a small primarily agricultural state with these problems of declining population. And what is what jobs can be done, especially we don't have a major urban city, they don't have a big city the biggest city is under a million people. And there've been all these, this is sort of in the era when Trump had been elected and there was all this talk about the urban-rural divide. And there were a number of plans that have been put out about ways to sort of revitalizing parts of America that are not you're muted there, but ways to revitalize the parts of America that I've been left behind by agglomeration forces and stuff. And I just thought that most of the plans that were suggested were, I wasn't really happy with them. So, there were plans to sort of make research centers, instead of in a few cities make pick 20 different research centers, or maybe even just 10 around the country and pour research money into those places, or revitalized small liberal arts college towns and, or just give people money to move to the big cities. And I didn't think that these were very good, but I thought remote work was potentially useful. If you can do the jobs of the agglomeration economies without having to live in them. This is a way to sort of tether left behind regions to the modern economy. And so, I started getting into that and I knew I had some personal experience working remotely in my previous job in the department of agriculture.
And I had been always paying attention to this issue of knowledge spillovers. And so, I started looking a lot more and I thought that there was basically a good case to be made that the outlook for remote work was a lot stronger than people appreciated, because I felt the arguments for remote work, this is all pre coronavirus were fragmented across many different social science domains. And there weren't people who had attempted to pull all the threads together, you have economists, or labor economists to study remote work directly. And they do experiments where they sort of see how well do people work when they work at home versus in the office. And they do experiments where they randomize people to one or two arms of this treatment. And they tended to find that remote work was pretty effective.
Often, they would find more effective than people in the office. So that was one strand of literature and they were kind of advocating for it, but they weren't. It wasn't they were super passionate about it. Then you had other people who were working on online labor markets and developing algorithms to match workers to firms or working for Up work business guy, John Horton who's worked on experiments on the design of the Up-work platform to match remote freelancers to things. And other people who've looked at the effect of the internet on job search and just all this stuff, basically that finds that how does the internet facilitate job search? And one of the problems with remote work has been well, even if you can do the job remotely, how do you find the job that's a good fit? And one of the promises of remote work is that you'll be able to hire just the right person for this job.
They're not local, but if you don't know where they are, if they're not, how do you find them? And with online labor markets and algorithmic recommendations that problem starts to get solved. And then on the third side, you have people who do study online social networks, and they're looking at the geographic footprint of our social networks and how that's expanding. And so, whether these things let us. How many of my friends live close and how many of them are far away and I don't think there's exactly work on remote work there, but just the idea that people tend to meet people who live nearby. But when they move away, there's a lot of those connections that seem to persist pretty well. And I think it's because of all the modern technology that makes it easy to stay in contact. And so, a lot of jobs are found through these informal social networks. And if your informal social network is now spread out geographically, that becomes another avenue. And then the last piece of the puzzle was just the knowledge spillovers I knew had been falling for a while. I tried to write a comprehensive survey of all these different things and put them in one place and say, look the case for remote work is a lot stronger than any one of these individual groups recognizing because all these things go together. And it's basically the case that the technology for performing work remotely has gotten a lot better. And so, you can, many jobs can be done remotely pretty well. We've done experiments on different kinds of works. And that seems to be the case. And even if I'm only 95% or 90% as effective working remotely as I am in the office if you can find the person who's 20% better, because they're not local, then that's good for the firm.
Or you can think about if I'm only 90% as effective, but I don't have to commute at all then on a per hour basis, that makes sense for me to just stay home and you can change my compensation maybe or something so that you're getting the same value for money. And basically, I wrote the case for remote work. I had started it and then coronavirus hit and I better get this thing out quick. And then I feel that has been so vindicated that it's almost boring now. Maybe it's a little bit too much, too soon to say, but I think everyone’s, everything that's been followed up, there's been a lot of work since then. And workers are pretty much across the board. It works better than we thought and is productive. And a lot of people say they're more productive working at home. Companies are realizing this and I don't know. It's a very weird case. I thought this is going to play out over decades and that we should incrementally make policies that make remote work easier. But instead, we had this huge shock that happened. Looking forward I think the biggest unknown about remote work is about the formation of social networks, informal social networks. Right now, we're kind of riding on the capital, the social capital developed from lifetimes of working in the office and in cities. And that has all these benefits of I go to parties with my coworkers and I meet their friends. And if it's a city that's a cluster, they might work in the same industry or a related industry.
And we build this whole network up. And that's a really valuable thing because that can give me a job recommendation. That can be a place I turn to for advice about, should I hire this person? Are they really good? I'm sorry if I have a question about an area of technical capability, I know people who I can ask, I am aware that there might be technical capabilities that I could use for some new innovation because I talked about it at a party. And if we all scatter which I don't think we're going to all scatter, but is that going to slow down innovation? And I think the big unknown is how much of that can be recreated with the internet. I have made a lot of professional contacts through Twitter and Twitter is in some ways a bar that you hang out and randomly meet people. And the thing about Twitter is a lot of the interactions are not as good as if you were co-located. Face-To-Face they're more fleeting. It's harder to form a deep connection, but you are interacting with an order of magnitude, more people. And so, I think we don't know to what extent those things balance out. And then, you have things other social networks, you have conferences, people would go to, and they could maybe do all their socializing in one place there. And then, meet a bunch of people. And in the past, I would have met them, but nothing much would have come of it. But now I meet a bunch of people and we keep in contact over Twitter and Facebook or email. And so, I'm able to form informal social networks in that way, almost as efficiently as if we all lived in the same city. And I think that we just don't know, this is the big unknown, but I think it is probably important. I still think that even if it turns out we can't form informal social networks as well online, or, maybe that's a big demerit, there's still a lot recommending remote work because it has all these other advantages that are sort of compensating. But that's my current big concern about the future of remote work is that side of it.
Ben Yeoh (01:57:52): I think that's fair. I think there will be these compensating, but that weak tie, social network, the serendipity, and also for a category of jobs where you need what I would say you need to face to face works better. A classic example would be a master carpenter. So, if a master comes to write their stuff in a book, probably you can't really follow it. Maybe a bit. You probably can learn a reasonable amount if you follow them on video. So, YouTube would help, but you obviously added a lot more, if you can actually be in the same room and follow them. And there might be a class of the things, the soft skills, knowledge process, and know-how that you need from master carpentry, but actually, that master carpentry category of jobs might be smaller than we thought, or rather that the category of jobs where the video is good enough.
The video might not work for master carpentry, but it probably works for baking a cake, might not work for a chef, but it might work for these other things. I think that's true. The one other thing I would say that actually, I think seeing your business leaders who, what I would say maybe for more old school industries or have older school ties are less keen on remote work. Hybrid. I don't know because I think that's where the compromise is going to be, but actually, if your tech Silicon Valley or you're, whatever, the CEO of gravity payments or Coin base you want virtually remote anyway, beforehand, you've done that. But actually, the CEO of JP Morgan, much more reticent. So those are my two reflections on that.
Matt Clancy (01:59:32): I think that the people who grew up who became CEOs in the old system obviously are people who know how to flourish in a certain domain. And you wouldn't expect that they're necessarily going to be the people who would also flourish to the same extent in a different kind of organization. So that was one reason I always thought there's going to be a slow thing to play out because it would be led by companies forming from the beginning. And then you'd have this ever-growing cohort of people who have experienced managing and leading remote companies. And they would spread out. Now we're in this weird situation where you have this. We're having companies that existed and were set up for one way deciding to switch because for these kinds of reasons, and they paid the switching costs in terms of the developed processes to do things remotely. And they've invested in the equipment coronavirus forced them to do that. But in some way, they being remote-first isn't necessarily in their DNA in the same way it would be for another company. And as to your first point, that's probably something I should have said right at the beginning, I don't think there's ever going to be a hundred percent well ever in a long time. But in the next few decades, I don't see us going a hundred percent broke remote at all. I sort of target well, 20 to 30% of workers being full, being mostly remote. And I think that's thing is still a huge change. You had some new industry that was 20% of people and it could be done from anywhere. It seems that's a really good thing for places like Iowa and so on.
Ben Yeoh (02:01:08): And whole categories of jobs, call center to the extent that it wasn't going to go AI, and you still think actually, what a person would be better than the person at home can completely do. Maybe that remotely on that. So, I do think that's true.
Matt Clancy (02:01:23): What you're saying also about the value of being a face to face. I totally agree with that. That's one thing that I changed my mind about doing all those researches is when I first started, what do you really need to be face-to-face for? When in the era of zoom, it's just people being Luddites or something, but reading more and it just continually comes up again, that even knowledge workers who are comfortable with it, they like to have occasional in-person things. WordPress has been fully remote for 20 years or whatever, but I think they have quarterly things where teams get together in person because there's a study of Google and they have people working all over the world and they talk about how valuable it is to have when you're starting a new project, to get everybody in one place to get on the same page.
And then you kind of figure out the roles and the project, the tasks that need to be done. And then it's a lot easier to them in the sense you're sort of breaking the problem down into discrete chunks that then makes sense to do remotely where people can focus. But I think that a fully remote, it's not everybody it's never going to be everyone and it's not even going to be that people never see each other. They just will see each other intensely for short periods of time. And then most of the time they won't at least in person, that's kind of how I think it's really to evolve. And then as you said, for the hybrid. Hybrid is, I don't know, it's a little bit risky. It's sort of in some ways it's halfway between the two, but you can very easily end up in a worst of all world's situation that hopefully, firms are going to try hard to avoid because you have to live in the region, you have to live within commuting distance. You're limited to the local labor population. And you have to, you have to commit to establishing processes so that remote workers can do things effectively. So
Ben Yeoh (02:03:15): You might as well be in person. It's kind of all that.
Matt Clancy (02:03:18):
So anyway, that's, that's the risk.
Ben Yeoh (02:03:21): I can see that. Actually, I didn't think about the risk burden of hybrid as much. Great. Last question, which is what does a productive day or week look like for you?
Matt Clancy (02:03:37): I am trying to write a new post for new things under the sun about every seven business days. It’s about 10 days on average. And I do a mix of reading for that and then writing for that. And then I have other projects and I think I don't have a hard time staying focused because I have three kids and when I dropped them off at daycare, versus when I picked him up is the only time to really get things done. And I worked pretty intensely for that middle period. And then I'm too frazzled to do much outside of those areas. I'm usually reading or writing. I think one thing that's interesting about doing the newsletter is I'm trying to synthesize and write about existing work. That's, invariably very high-quality cause that's why I chose to write about it. And I didn't really realize this until I was working on my own research project recently and had that feeling of thinking things are going really well. And then about three hours later, realizing that you made an error and everything was a waste. And I didn't realize, but the safety or I don't know, the risk is lower of me, I'm always producing stuff that I think is valuable. I'm doing it more often now rather than doing these really uncertain things. I think that's what a productive day looks like for me. Drop the kids off, read or write until five answer emails and go to meetings is less productive. It feels less productive. It probably needs to be done anyway. And that's it.
Ben Yeoh (02:05:24): And the occasional podcast.
Matt Clancy (02:05:25): That's right. I love it.
Ben Yeoh (02:05:27): Great. Well with that. Thank you very much. I think that was an amazing conversation. And please check the link down below and subscribe to Matt's substack. Thank you very much.
Matt Clancy (02:05:39): Thank you very much.
