Byrne Hobart joins Jordan McGillis to discuss his book, Boom: Bubbles and the End of Stagnation.
Audio Transcript
Jordan McGillis: Welcome to 10 Blocks, I’m Jordan McGillis, economics editor of City Journal. Since the 1970s, advanced economies have experienced slower growth than they did in the first two-thirds of the 20th century. The reason according to great stagnation theorist Tyler Cowen is that we’ve already picked the technological low-hanging fruit that could deliver big returns to quality of life. Now, we’re eking out better efficiency, but with real breakthroughs fewer and further between. Into the economic malaise discourse has now arrived a new theory for how we can escape the great stagnation. Byrne Hobart and Tobias Huber argue for an unlikely remedy to slow growth: bubbles. In their forthcoming book Boom: Bubbles and the End of Stagnation, Hobart and Huber make the case that under the right conditions, bubbles are the ideal exit ramps from stagnation. Joining me to discuss this counterintuitive idea is author Byrne Hobart. Byrne, thanks for coming on.
Byrne Hobart: Absolutely, great to be here.
Jordan McGillis: Let’s start with the premise, the great stagnation. Can you concretize that for us and give us examples that we might be able to recognize that show we have slowed down?
Byrne Hobart: Yeah, absolutely. So I think the place to start is that just the default state of humanity is we’re very poor and over our lifetime, that does not change in any visible way. That’s been the human experience from pretty much the dawn of time through about the mid 19th century and depending on exactly where you were in the world, it hit a little bit earlier, it hit a little bit later, but we did reach this positive inflection point after a long period where the GDP per capita, it wasn’t just flat, it was just compounding at an extremely slow rate and it was limited by information transmission mechanisms. It was limited by different technologies, having dependencies on one another. It was also limited by the fact that plagues can destroy lots of concentrated populations, that political upheaval can reverse many generations of material progress.
So we were kind of bouncing around this oscillating around this GDP per capita mean. I think today we would describe as just barely subsistence if you’re lucky, and often worse than that for most people, and then things do start to accelerate, and there’s this combination in the industrial revolution of one, just more opportunities to actually save capital, turn that capital into useful productive assets, and then actually reap the gains from those assets, but also there were these gains in productivity and productivity is this very nebulous thing. It’s because really the way people started talking about productivity in a rigorous way was there were these very simple but edifying models of economic growth where you have some amount of labor input, some amount of capital input. As those grow, basically the product of those growth rates is the overall growth rate in the economy.
And it turns out that if you track that over really long periods, if you try to say, “Okay, the amount of fixed capital in the economy grew X percent and the number of labor hours per year grew Y percent.” You have this residual, you have this gap and the gap is we are richer than we should be on paper. When you ask yourself, “Okay, how could same number of work hours using the same equipment lead to better outcomes?” Productivity is a nice word for that and productivity is a word that’s going to encompass many different things.
It’s going to be everything from you are spending the same amount on say machine tools at your factory that you did a generation ago, but they’re much more efficient, much more precise. They use less power, they need fewer people, fewer operators, etc. So even though your inputs are the same, your outputs are higher, but they’re also social technologies, just ways of organizing any kind of human endeavor, whether it’s a business or it’s a branch of the government or a charitable organization or even kind of a more abstract entity that has interest and pursues goals.
And that productivity growth number in the mid 20th century had just this incredible run where it was growing at an average of like 2% a year. So if you held all of the inputs fixed, the economy will grow 2% a year just because of primarily better technology and also better institutional structure organization, etc. And that productivity boom, it really started to kick in the ‘30s and there are lots of interesting questions and there are very few periods of economic history that have more moving pieces going in more different directions than the 1930s because you have these incredible productivity enhancements in manufacturing, you have a large number of new discrete products, so things like appliances, radios, washing machines, etc. All these things getting more widely deployed. You have more widespread electrification.
And then you also have more intermediate goods, so a lot of chemical compounds that we use all the time, we don’t need to know the names of them, we just kind of notice that they’re there. A lot of that stuff comes out of DuPont and its competitors, a lot of it happened in the 1930s and then continued at a slower pace thereafter. So we had this incredible boom in productivity and then it’s been widely observed that since roughly 1970, that productivity growth has really petered out and there just are growth, the growth in productivity roughly halved since then. There are some promising recent signs that productivity growth might be reaccelerating and it’s a good idea to keep an eye on those, but productivity, it’s a frustrating thing to track because year to year, you just have lots of swings that are pretty noisy. So for example, measured productivity will be a lot higher in the first year of recession because a lot of people get fired.
The first people to get fired are people who were less productive and so just mathematically you end up with higher productivity. You also just have fewer people using the same base of equipment. So the laborer’s output per hour is higher. So you have some noise like that in short time periods, and that’s why I’m very cautious to extrapolate the recent more promising numbers, but yeah, that is the general sweep of things and I do think the most defensible theory of this, like your base case should be that there were a lot of very useful general purpose technologies that could have existed at any time, but it started to become practical to discover them and implement them in the late 19th century, and then they scaled massively over the 20th century and then we ran out, and then there’s another view that is maybe a little bit more of a policy view, which is we sort of regulated our way out of progress and I think there’s something to be said there.
There’s also a demographics piece that if you look at a mostly agrarian economy with very little socio-economic mobility, you tend to have an economy that’s going to be misallocating a lot of human resources just because there are going to be a lot of really bright kids who grow up to be the smartest farmer in the whole county instead of discovering a new branch of quantum physics or something, but over the course of the 20th century, education became much more accessible, became much more universal. Years of schooling went up and schools started to get a lot more rigorous. It used to be they were elite schools because there was an existing elite, and that elite sent their sons mostly to this small set of schools so that became elite.
But then over time, they became elite in the sense of you need to be very smart and hardworking and have some dose of either luck or be a legacy to get into these schools. So they became elite in a very different way, and that meant we just had much more aggressive sorting of human capital, but that’s a one-off benefit. You can do find all of the smart kids growing up in farms or growing up in tenements and get them into Princeton one time and then their kids go to Princeton and you don’t get additional gains. Sure, you do have a better sorted elite, but you don’t get those incremental gains. So there are a lot of moving pieces and I think it’s anytime there’s a mono-causal theory of a very complicated economic phenomenon that we only really found because it looked like a weird measurement error, those it’s right to be suspicious of.
But it’s also by the same token, that means there are a lot of things that you can incorporate into a model of this behavior to make it better and the one we focus on in the book is bubbles and government mega projects, and we kind of treat these concepts as more similar than they are usually seen to be because they actually have a lot of underlying commonalities, one of which is just you concentrate a lot of capital and a lot of talent into a particular task in a narrow frame of time, and that turns out to be a big qualitative change from what you would get if you spread that out over time and didn’t treat it as this intense one-off thing.
Because what happens when you have this huge concentration of talent is you have very high quality serendipity, and we wrote about this with respect to things like the Apollo program where when they were looking at the specs for different things that they needed for the launch, they realized that one thing that was really important was they needed guidance computer, it needs to be very reliable, shouldn’t use too much power. It needs to be fairly lightweight and the standard way to do computing was vacuum tubes, which they did the job, like you can implement logical operations in vacuum tubes, but they burn out, they’re very heavy, and they use a lot of power and you really don’t want your space program to be limited by the question of, “Okay, how many spare vacuum tubes do we launch up with these people launching a pound of mass and orbit?” It’s expensive now. It was really expensive back then.
Jordan McGillis: I want to home in on the similarities and the distinctions between the government mega projects that you bring up, notably Apollo and Manhattan Project and then private sector bubbles. How are they similar? How are they different? What are these characteristics that you identify that can accelerate innovation?
Byrne Hobart: Sure, so one of the differences is just you have a very different system of checks and balances. So the public sector mega projects, it is not a coincidence that these tend to be chronologically earlier than some of the private sector bubbles that we talk about. The US was just had much higher state capacity from the 1930s through the 1970s basically. The really simple argument for that is in the 1930s, if you were young, ambitious, just entering the workforce, there’s only one employer who’s really hiring, and if you are restricting yourself to jobs where you can very rapidly move up the organization chart and be in charge of a lot of people who have serious responsibilities, it makes even more sense to work for the government. If you go work for GE, just wait patiently for your promotion.
Jordan McGillis: So you’re coming as we’re mired in the Great Depression of the 1930s is what you’re referring to there? Okay, and it was interesting to me that you reference the ‘30s as a time of great innovation given that we were in the depths of that economic struggle and people would usually say that that was a result of a bubble. That’s the common narrative. Explain how that’s wrong and how you conceive of bubbles differently than the average person who thinks 1929 stock market crash.
Byrne Hobart: Yeah, so definitely 1929 is a really interesting phenomenon. The boom of the 1920s, the bust of the 1930s. One of the drivers of that boom in the 1920s, it was not purely money flows. It was also that investors were genuinely and correctly optimistic about electrification and that presenting opportunities both for investing in electric utilities and also investing in manufacturers that had newly electrified. So historically, manufacturing, one of the big limitations is you need some kind of external power source and if you can’t use electricity, you are generally using something like water, which means that it doesn’t just mean that you have to locate your plant near running water. What it also means is that you’re designing your entire plant around power being mechanically transmitted. You have this one central pole that rotates everything is just attached to that with various pulleys and gears and levers and things.
And so you just have this one source of a finite amount of power, and that means it’s actually very computationally difficult to figure out how to upgrade a factory. If you are adding an assembly line that uses a little bit more, a little bit less power, you have to rebalance everything else that this facility does. It also means that the factories tended to be very three-dimensional structures. So the optimal structure is actually a cylinder, but that’s hard to build. So what you actually end up with is a giant cube, and a giant cube adding additional floors gets pretty expensive and you’re putting this cube just wherever the power is. So you can’t really optimize for this is a really good place with commutable place with cheap land, etc. You are stuck with a lot of constraints that are just kind of annoying constraints, but those constraints guided how everyone talk about these things.
Jordan McGillis: There’s some excellent cylindrical factories in Richard Scarry’s What Do People Do All Day? where the water’s coming in is just down and down and down and down, amazing.
Byrne Hobart: Yeah, the Richard Scarry books are great for just making a lot of this stuff more tangible. I think The Economist had an article a while ago about how the Richard Scarry rule is if you’re a politician, never condemn any job category that appears in Richard Scarry book and so I think the corollary of that is like we need new Richard Scarry books that have high-frequency traders and people who write ad targeting algorithms and people who design GPUs, all those contributors to our current GDP. Anyhow, so once you can electrify a factory, you can choose how much power to consume. You also aren’t restricted in building it around one specific power source, and so you can actually sprawl your factories out, and that means that if you are running one of these companies, it actually makes sense to continuously expand rather than expand in discrete increments because you started with one factory, now you’ve built or bought another one.
You can just add one more assembly line, you can upgrade some of your equipment. You can just continuously adjust what you have, and what that means is that now these companies can actually grow and it makes sense for them to retain their earnings and expand. It makes sense for them to do long-term research and development. I think the historical investor attitude towards public equities was this is basically a bond with even fewer investor protections, even more risk, and so it should pay you a premium, and now the finance nerds will still say equity owners are the residual claimant, and that’s the value of equity, and that’s also the definition of equity, but if you are thinking of it from an investor perspective or general capital allocated perspective, the interesting thing about equity is you have this uncapped upside and it can just compound forever.
And that was a lot of what investors were excited about. Now, if you look at the financials of some the hot stocks of the 1920s. For one thing, a lot of them are banks and the valuations don’t make any sense, and they weren’t getting great returns on equity. A lot of the big industrial companies were not getting great returns on equity, but it was starting to creep up. A lot of investors just weren’t thinking about returns on equity because it was less salient to a world where companies don’t retain capital for growth. So you had this change in what is the optimal way to think about a company? What’s the optimal way to run a company? You had this category of companies that were actually growing.
For a long time, investors would sort of anchor to the par value of a stock, they’d say, “This company went public at $100 a share.” It has $100 a share in net assets or it’s supposed to, and its earnings are going to fluctuate around some reasonable return on $100, but some of the utilities, their earnings might be $8 a share of one year and then $10 and then $12, and then $15, and investors start extrapolating and start putting a high multiple on them. Now, the utilities growth story did not go great in the 1930s in part because they just faced a lot more regulatory restrictions, and some of that growth was that they were massively levered. There was just a lot of fun things you could get away with in financial markets a century ago you can’t do today. So the utilities, they weren’t great as a growth story, but they did actually replace railroads as the default safest security, and they actually did outperform the broader market during the crash.
And it was a lot of the electrified manufacturers though who were the true beneficiaries of this technology transition. They were the ones who were able to just mass produce lots of goods and produce them at lower prices, make them more accessible, and you actually had this feedback loop where the existence of a lot more appliances meant that it made a lot more sense for more houses to get electrified. The existence of more electrified houses means larger market for these appliances. So you have a nice virtuous cycle that allows both of those industries to grow, but the size of each industry kind of constrains the growth rate of the other one. So they actually grew at a reasonable pace. If you look at a list of the largest appliance companies, what you’ll usually find is that they were founded around when their home country started actually electrifying.
The ones that are founded later, because their addressable market is just every household that has an outlet, they often have to expand really fast because their competitors are going to copy their product, and because they’re expanding so fast, there will be some point at which they overestimate demand and they exit the holiday shopping season with way more inventory than they could possibly sell, and that sometimes kills them. So it was this golden period for building really big companies in that area and then that period does also transition into the post-war period. So I guess the war economy is its own thing. It is actually amazing in retrospect that the US was able to put such a huge fraction of GDP into fighting a war and then have a kind of normal functioning economy to go back to at the end of all of that, and they were able to move millions of people out of the workforce and then put millions of people back into the workforce and did not cause a depression. So it is like a minor economic miracle that all of this stuff worked out.
Jordan McGillis: How much of the postwar boom can be attributed to our physical capital being intact compared to the other industrial players at that time?
Byrne Hobart: Yeah, that is exactly where I was going with this. Yeah, we dropped bombs. We did not get bombed all that much and so at the end of the war, the US was pretty much the place where fighting hadn’t been happening, but the building of equipment had been happening. So you had a lot of facilities that were owned by the government auctioned off to the private sector. So lots and lots of spare capacity, but also lots of pent-up demand, and because the US still had this manufacturing capacity intact and because the rest of the world largely didn’t, there was an export market for American manufactured goods, and it wasn’t a huge export market. A lot of the growth was still just internal demand, but there was enough of that demand and just if you have the economic capacity to produce enough tanks to replace all the tanks that are going to get blown up on the front lines for either front, because the US was supplying a lot of the USSR’s military equipment too.
If you have that capacity, you have the workers who are just used to building lots and lots of motor vehicles quite efficiently. It is not trivial, but not impossible to redirect that towards building lots of cars. So it did work out pretty nicely, and that is along that timeline. So over the period where the US has transitioned to a war economy and is investing huge amounts of capital in military production. There’s also the Manhattan Project we talked about where again, you have this definite vision of the future of we can build this unimaginably powerful bomb is theoretically possible. We know that the Germans have uranium. This was actually one of the early things that alarmed the White House was they learned that the German government had quietly put some restrictions on exports from Czechoslovakia’s mines, many of which had access to lots of uranium. So there was this awareness that this weapon could be built, and it’s possible that the same research that you do for this military application could be used to generate cheap electricity in the future.
That motivated a lot of people and this actually brings up one of the commonalities between mega projects and bubbles, which is the idea that both of them, because of the time constraint, because of the capital and talent abundance, they encourage you to run a bunch of things in parallel that you would otherwise do serially, but where if you tried to do them serially, you would just run out of money and run out of enthusiasm and nothing would get done. So with the bomb, there were multiple proposed designs using different kinds of fissile material. It was unclear which of those designs would work. It was unclear which materials we’d actually be able get access to. We knew that it’d be very capital intensive to get the right isotopes in sufficient quantities, and also no one had done that before. No one had gotten bomb level quantities, and so what we did was we actually just tried a bunch of different techniques in parallel.
So there were different uranium and plutonium processing facilities in different parts of the country. Some of them worked, some of them didn’t and enough of them worked that we were able to build a test bomb and then two different bomb designs that were actually deployed in the war, and if you think of the counterfactual where the budget’s a little bit smaller, you don’t have quite as many incredibly effective people at every level of the org chart, maybe you would’ve picked just one of those techniques and if you picked the wrong technique or you picked the wrong bomb design, then instead of the Manhattan Project, you have the Manhattan boondoggle, and after the war, there are a bunch of congressional hearings on how these physicists were able to trick the government into giving them a lot of money for a science project when red-blooded American boys were dying on the beaches.
So we really lucked out in that this stuff is actually physically possible in this world and that kind of persistent luck where making huge investments in what seem like slightly deranged visions of the future, it does tend to work out, and part of the way it tends to work out is that these deranged visions of the future, they form a sort of, you can think of the idea of an industrial cluster, like an industry cluster where if you want to get into movies, you should move to LA. If you want to make it big in finance, you should be in New York. You can also create such a cluster in time where if you want to build this thing, if you want to build an atomic bomb or go to the moon or make America energy independent or create a new form of money that isn’t controlled by any government, there’s a time that you can actually work on that and it will either get done then or probably not get done at all.
And so when people talk about bubbles, they’ll often talk about this idea of FOMO, fear of missing out, and that’s one thing that drives people to make really irresponsible decisions in a bubble, but if you are allocating your time and talent rather than your money, missing out is actually the thing you should fear the most. The thing you should fear the most is that something amazing was getting built, but you didn’t quite believe in it, and you knew that people you respected believed in it more than you did. You couldn’t quite get on board, you worked on the second or third most important thing, and that thing turns out to be a footnote and your friends end up making history. I think it’s a healthy thing to fear. We all have finite lifespans, so we have a limited number of shots to make a really massive impact on the world.
And so when things start to move and they start to have this internal logic of their own that keeps pushing them towards more and more extremes, that is actually a sign that the next equilibrium they reach is going to be radically different from the current one, and even if you’re somewhat skeptical of the boom, it can often be worth it to be as close as you can to that boom specifically to see how things are playing out. So if you look at the AI boom, there are people who are really hyper-optimistic about AI. They are very dismissive of existential risks, and I’m not a huge believer in those risks, but I do try to take them seriously and try to discuss them, keep them in mind, etc. So there’s a set of people who are just wildly optimistic, do not believe that the risk is real.
There’s a set of people who are extremely pessimistic, but because they’re pessimistic about AI because they got interested in it early, they thought about it a lot, they learned a lot of the details of how these systems work, how they’re trained, etc., and decided that the risks are unacceptably high and a lot of them work at big labs. So side by side, you have accelerationists and people who really wish there were a practical way to pause AI, they’re working together to ship the next model because each side wants to see what’s next and also wants to keep an eye on how fast things are progressing. They have very different utility functions with respect to the speed of that progress, but they care about the same thing, and as you dump more capital and talent into one specific field, you do end up bringing together a bunch of people who have wildly different motivations, but they can still work towards the common next step, and so you just get more productivity out of that same set of people and capital than you otherwise would.
Jordan McGillis: Your book introduces a new concept to me at least, and you’ve referenced it indirectly here a couple of times, technologies of transcendence, nuclear power, leaving planet Earth, artificial intelligence. Can you expound on this idea you have?
Byrne Hobart: Yeah, absolutely. So this gets to another concept, which is just when I think about bubbles and I think about technology and economic growth, part of what comes to mind is these deeper questions about what is my purpose? What is our purpose collectively as human beings? And one framing of that purpose is that a lot of what we’re here to do is figure this stuff out, and you can take that in a completely secular direction and say that the observable reality is all that there is and therefore figuring that out is what it means to live, and then you can take it in a religious direction and say, “This is creation. It was created by something omnipotent and omnibenevolent and omniscient, and therefore anything you observe in nature is the cleverest way that this thing can get done,” and I think that either of those end up motivating this same kind of deep curiosity about how the world works and what can be done.
And when we discover new things, when we increase our output per hour, what we’re also doing is just reducing the number of hours that we have to spend doing just the basic subsistence things. So modern America, we have plenty of economic problems. We are also just an extraordinarily wealthy country where people can achieve a standard of living that is just ridiculous to the rest of the world without having to work as hard as people may have to work in other parts of the world to get something similar, and so we are just over time transcending more limitations as more of the energy that we consume is not human produced energy as more of the insights that we consume are not our own insights. So you can think of something like a smartphone as just this, it’s crystallized intelligence from everyone who worked on every little detail of this phone, and it would be unimaginably expensive to just put a bunch of people to work building you one copy of a new iPhone.
But if tens of millions of them get sold, then it is actually worth it for someone to work full-time making some little component 0.01% better because it’s amortized over such a large group of people and it does make all of their lives that increment better, and yeah, over time, I would say from an ancient perspective, we live in a post-scarcity society. We’ve transcended all of our limitations. It’s just hard to imagine something like recording this podcast and then transmitting it. It is just not remotely worth doing in a poorer society, and the richer we get, the more things there that are actually worth doing just because the cost has gotten low and that means there’s just more room for self-expression, and again, you can hold in your mind this tension between the more secular and more religious interpretation where the secular interpretation is you yourself have your own unique personality, have your own unique interests and wants and needs, and that is kind of your moral North Star is like, “What are the things that you actually want to do? Can you do more of those things and more cost-effectively?”
And then from a religious standpoint, you can flip that around and say, “No, if we’re all made in God’s image, we’re all unique. There is something we were each put on this earth to do, and the more effectively we can do it, the more we can actually follow that plan.” So I think either way, you can have these more cosmically significant interpretations of the extent to which GDP growth cannot be explained purely through changes in working hours and the amount of physical capital out there and I think that’s pretty cool.
Jordan McGillis: Talk to me about the cosmic significance of science and the way people interact with it and hold it consciously.
Byrne Hobart: Yeah, so when you think about the world, you’re always thinking about it in terms of what you can see, what you can interpret, and how you interpret that stuff and so the more knowledge we have collectively, the more we actually see, the more details you notice. I actually noticed this at a really low scale. Years ago, my job description slightly changed and I was now an analyst covering airline stocks, and the next flight that I took, I just could not stop being aware of all the little details of this entire process that are designed to either make boarding a little bit more efficient or find one more way to sell people things during the flight or get people to sign up for awards programs. It was actually a much richer experience just because there was a lot more context around what’s going on, what are their end goals, what are their constraints, etc.
So the more that you can have that feeling just everywhere you look, the more that you can look at things and actually explain them coherently, actually the more you can observe because you are compressing all of your observations down to what is the minimum thing you need to know and you can extrapolate the rest, and you’re also just noticing things that you would not otherwise notice. So I do view that scientific knowledge, it is among other things, a way to experience reality more directly by seeing past the details, seeing what is underlying those details and it makes you notice when things don’t make sense. The more theories you have about the way the world works, the more salient it is when your theories are obviously not holding true.
Jordan McGillis: And when you observe the state of science today however, you’re seeing stagnation in the same way that we do in the macro economy. Can you talk about stagnation in that field?
Byrne Hobart: Yeah, I think looking from the outside, it’s very hard to see exactly where things went wrong and when we talk about this slowdown in the progress of science, we do want to talk about how there are many different causes and it’s very hard to identify which ones are the biggest unique contributors, but there is this general burden of knowledge problem, it takes more time to get up to speed on any important field and as those fields become more established, they get a little bit more calcified and there’s more of a linear path to being able to make incremental contributions, and that’s a necessity when you have a large enough discipline that not everybody knows everybody else, and where you don’t have people who are selected in purely based on passion and being able to make material contributions right away because there’s so much low hanging fruit.
You do have to systematize things a bit, but what you do end up with is a more metrics-driven approach for something that is just fundamentally not amenable to a metrics-based evaluation process and that leads to pretty strong incentives to in some cases, outright falsify data, but in many other cases to just ignore a lot of complicating factors in order to publish something that looks real, and that’s a pretty pathological incentive and then we do have other domains where there are pretty strong incentives not to do that, and so there are just very different academic practices when you’re looking for statistical significance in say a social science evaluation of people’s attitudes, etc. There’s a lot of data mining that is socially acceptable to do within those fields, and then if you are trying to do a similar kind of analysis, but it’s over what is the conditional probability of someone making a purchase after clicking on an ad and you’re testing different things? The people who do those studies really do not want you to just p-hack your way into saying, “Yes, it’s significant.”
Maybe you get promoted once, but at some point, someone catches on that you’re faking your data and then you get fired and they do actually look at the metrics. They have faster feedback loops and it’s a hard problem because in science you have really slow feedback loops, and especially in domains where you want people making big contributions when their fluid intelligence is high and it peaks at a frustratingly early age, I’m well past that peak. So the amount of things that I know is still going up and the number of people I know is also going up.
So between just accumulated knowledge and the social graph, I’ll be fine, but it does mean just my mental horsepower for making big contributions to mentally demanding fields is not what it was, and so if you are trying to get people who are at their peak, you tend to be promoting them before you really know how good they are, and that’s a pretty high risk thing to do and so as the practice of science gets more institutionalized, there’s just more of an incentive not to do that and more of an incentive to have an exceptionally smart and hardworking cohort of people who are also probably more agreeable than you really want a scientist to be, and actually perhaps less open to new ideas than you’d expect.
Jordan McGillis: I want to swing back to the commercial side of things. You spend a lot of time in the book writing about the golden age of corporate R&D. Can you talk about some of the standout companies from this mid-century era? I know as an Acquired podcast listener, Lockheed is one of the legends as well as of course, Fairchild and the Bell Labs and these sorts of places. What are your favorites?
Byrne Hobart: Yeah, so I would say Lockheed was a really fascinating one because it was almost this passing of the torch from mostly government focused to more private sector focused where obviously there was not a very big private sector market for spy planes and things, but Lockheed, it was a private company. They were trying to hit their quarterly numbers, pay their dividend, etc., and they were working closely with the government because they also really wanted to win the Cold War, and they managed to create this structure where there’s surprisingly little accountability except for when it really, really counts. And the accountability was often at the level of making sure the feedback loop works, making sure that people who are designing something are actually on the shop floor, trying to see how their designed would be realized. They’re talking to the engineers and mechanics who can tell them which ideas they have are totally wrong, and they’re trying to make sure that information flows back to the Pentagon as efficiently as possible, which meant cutting a lot of corners, skirting a lot of rules, but also it meant that the iteration cycle was really, really fast.
And when you’re dealing with a situation like that where the Soviets had some advantages in terms of radar and in terms of anti-aircraft defense and the US had some advantages in terms of the planes, you really want a fast cycle. You want to make sure that there is not a point at which the adversary has a multi-year head start before you can catch up to them. So you always want to be guessing what they will do in response to the thing you’re working on now and working on the countermeasure to their countermeasure, and that actually ties into one of the other bubble dynamics we talk about, which is this idea of positive feedback loops where when you have one side building something on the assumption that the other side is going to build the complement to that thing. So I think of the ‘90s, the internet bubble. AOL sent out a lot more disks with one month free or whatever because they knew that Yahoo is categorizing all the interesting sites and that Amazon is selling more and more useful products that CNN is putting its stories online. New York Times is putting their stories online. There was just more and more to do online, and you could assume there’d be even more than that over time and so it made sense to sign people up early and meanwhile, from the Amazon perspective, made sense to invest a whole lot if all the ISPs are desperate to get everyone in the US spending as much time as possible online. So each side builds something knowing the other side is going to build a thing that makes that initial investment worth it, and that collective decision, it’s delusional if one side is wrong, but if they both happen to be deluded in the right way, they sort of escape the prisoner’s dilemma and actually invest for growth.
So going back to this golden age of corporate R&D, the one I have a soft spot for is actually DuPont. So for a couple of reasons, one of which is that part of the story of DuPont and corporate R&D is this long bout of soul-searching, basically like the first Iron Man movie, so the people of DuPont, the DuPont family, their name is on the building, etc. They really did not like the fact that people were calling them merchants of death after World War I And they also didn’t like the fact that it was basically true. DuPont’s profits would go up a lot if there were a military conflict, more death in the world, more suffering in the world means more explosives, means DuPont when they sell mostly gunpowder makes more money, and so they wanted to sell things that just wouldn’t make them merchants of death and some of those things were not great. So they made really important contributions to leaded gasoline. It turned out not to be a great thing to make material contributions to, but they were able to-
Jordan McGillis: Quick pause there, what were the big benefits of leaded gasoline? Why did we gravitate to that?
Byrne Hobart: So my understanding, definitely not an auto mechanic. My understanding is that your engines would sometimes knock, they would get out of sync, something like that and you’d have to restart the car if that got too bad, and so you just couldn’t drive very far with unleaded gasoline and I think some combination of engines getting better and us finding new anti-knock additives that did not cause serious brain damage. Yeah, replaced that.
Jordan McGillis: Okay, back to the positive contribution.
Byrne Hobart: Yes, back to the positive thing. So stuff like Teflon and Rayon, all of these materials, they literally change the texture of the world. If you are just in your house and you’re picking up objects like you’re cooking dinner or something, you’re touching a bunch of compounds that just do not exist in nature, did not exist until some clever person probably at DuPont ran some experiment and realized they’d found something cool. I keep forgetting to look up which compound it was. There was one, might’ve been Teflon, might’ve been Kevlar where the way they discovered it was that it was actually the residue from some other compound they were trying to synthesize, and the lab techs complained that whatever this stuff is, it’s just impossible to scrub it out of a beaker or something, and they realized that yeah, if you have this unscrubbable thing like once it’s attached to some surface, it sticks there forever, this could actually have some useful applications.
Cellophane was another one and cellophane is a really interesting one because part of what it changed was its second order effect was it totally changed how we buy food. So in a pre-cellophane world, the cheap cost-effective way to package food is either one, if you are fancy, you put it in a brown paper bag and you have a label or something or you have a clerk telling people what’s in this bag, and if you’re less fancy, you just have a barrel. The term Cracker Barrel, that was a thing. You would buy some crackers, they would scoop them out of the barrel and there are your crackers, but with cellophane, it actually became cost-effective to show people the product they were buying and that actually meant that now everyone cared a bit more about how the food looked. Obviously very easy to take that too far.
There definitely trade-offs where the better it looks, the worse it tastes for some kinds of produce, but it’s also true that one way it can look bad is just if it’s rotting, and if you’re grocery shopping, you see this adventure of okay, what am I actually going to end up with or how much time am I going to spend inspecting every single piece of food, that’s pretty annoying, economically efficient, and can lead to food poisoning, and if everything is very visible, you can inspect it just as you look. It changes how you shop. It changes which products you buy. It means that there’s actually more room for things that taste good and look interesting. It means there’s more room for prepackaged foods. It also meant that it made sense for stores to actually add some nice interior illumination to expect people to enjoy spending some time browsing around.
So totally changed the shopping experience. It actually collided with another feedback effects bubble, which was the automobile where once car ownership was ubiquitous, it made sense to think of grocery shopping not as something that you do pretty much daily and you probably do within walking distance of your home. Instead, it was something where you could do it once or twice a week. You could drive for a bit, load of the trunk of your enormous Cadillac and drive home with a whole lot of food, and that meant that if you built a much larger grocery store further out from town, you could amortize just the fixed cost of running that store over a whole lot more sales and make a ton of money. Again, if you have all these nice plastics that allow you to cheaply package these products and make them look good, then you are in a really good situation. You have lots of positive swill over effects from multiple financial bubbles.
Jordan McGillis: The big downside of cellophane of course is that you can no longer read newsprint off your pork chop.
Byrne Hobart: It is a tragedy, yes.
Jordan McGillis: All right, so you’re a big DuPont fan. How about today, what’s going on that you’re enthusiastic about? How do you feel about the corporate investments into AI, for example?
Byrne Hobart: Yeah, so I’m feeling pretty good about that. In the book, we did try to not spend a ton of time on AI and part of the reason was just I think there are many other potential bubbles. There are many things in healthcare that could get massively better very fast, and GLP-1s are obviously a big part of that, but definitely not the only part. I think actually healthcare is a case where just getting tight feedback loops and getting aligned incentives does a huge amount of good. Space is actually, it’s becoming an increasingly interesting one just there are a lot of hardware limits, but then there are a lot of just real-world limits that actually go away once you are in orbit and beyond, and the short-term ones are there are just are some manufacturing processes that are sensitive to things like gravity, and so if you can get rid of gravity, the whole problem is a lot easier. There are some use cases for that where it does actually make sense to go to the trouble to get gravity out of the equation. Obviously, for communications-
Jordan McGillis: Pharmaceuticals especially, there are potential benefits from not having various helixes collapse, that sort of thing?
Byrne Hobart: Yes, so I think that one is really interesting and then AI, it’s very easy to track. It’s very much in your face right now and I do continue to think that AI has a lot of very positive implications and the tools are also just really fun to use, and I think that is also just an underrated aspect of new technologies, but when I read about the early days of any new technology, it’s amazing how much fun people are having. There’s an anecdote from a biography of William Shockley where it’s talking about being a grad student in physics in the late 20s, early 30s where he says something to the effect of, “Every fall the new textbooks would arrive and if you read faster than your professor, you knew more physics than your professor.” And so you would be in a better position to write the final exam than your professor was.
And that’s got to be just electrifying, but at this point, you have to be a pretty extraordinary person to know all of physics by the time you’re starting your PhD in physics. It was kind of a reasonably practical goal of you know all the most important parts a century ago. Now, that’s really not feasible, but the fun part is actually a really important thing. That is part of what keeps people really motivated is that they know that even if they’re not necessarily building something lucrative, they are building something really cool that they personally will be proud of and it’s often people who have the motivation to do that who are the ones who keep on building and do end up stumbling on something that turns out to be very valuable to everybody else.
Jordan McGillis: Our guest today has been Byrne Hobart. Byrne is the author of a forthcoming book, Byrne, where can our listeners find that book?
Byrne Hobart: Sure, so best place to find it would be stripe.press/boom and you can also search Amazon for Boom: Bubbles and the End of Stagnation.
Jordan McGillis: Terrific and elsewhere on the internet, where can they find your work?
Byrne Hobart: So I write a newsletter on trends in tech and finance. It’s called The Diff. So go to thediff.co and you’ll find it.
Jordan McGillis: Byrne Hobart, thank you so much.
Byrne Hobart: Thank you.
Photo by Andy Roberts/Getty Images