Luke Burgis Newsletter
Audio & Recordings
Anti-Mimetic Salon #1: Byrne Hobart Applying Mimetic Theory to Financial Bubbles

Anti-Mimetic Salon #1: Byrne Hobart Applying Mimetic Theory to Financial Bubbles

Financial analyst Byrne Hobart talks about his paper "Manias & Mimesis: Applying René Girard's Mimetic Theory to Financial Bubbles"

Thank you to everyone who turned out (live) for the Anti-Mimetic Salon with Byrne Hobart last night. I look forward to doing many more of these with you.

I know there were many who wanted to attend but could not make it. So please find here:

1) Audio/Podcast version (above)
2) Video of the Zoom call (below)
3) A full text transcript (beneath the video)


Video of Zoom Call:

Full Transcript:

Luke (00:01:55):

Awesome. Well, welcome everybody to a great event with our guest Byrne Hobart. Byrne really glad to have you with us today. Excited about this. I've been following your work for a really long time, and you are clearly very familiar with Girard, as I think everybody on the call is at least somewhat. I said a little bit of contact. So I think we can dive right into a pretty fascinating conversation about manias and mimesis. And I think the thing that interests me the most about the paper is that it was written in October of 2019. And that timing is interesting because I want to know what the 2.0 version of that paper is so much has happened. You had no idea what would happen in March of 2020. You talked about, I don't know, at least a couple of Bitcoin bubbles, but we didn't get what happened over this last year. So I'd love to get your thoughts on that.

Luke (00:02:54):

But first, just a few words of introduction. This is part of a series that I'm starting to bring in interesting guests. Next month will be Demetri Kofinas from the Hidden Forces Podcast. Planning to do at least one of these a month and record them. Byrne Hobart, our guest is the writer of one of my favorite Substacks called The Diff which you can find at I don't even know... He does something that's really unique. And he writes about the intersection of finance and tech, but brings in history, philosophy. I would almost call it a metaphysics of markets in some ways. Uncover to use one of Girard zone terms, like some things hidden beneath the surface that help explain some of the inflection points that were seeing in technology and in the financial markets.

Luke (00:03:52):

And he writes pretty frequently. I don't know how he does it, but they're always fascinating. So I would encourage you to check out his Substack. The format for today is very laid back. I think my working title for this is the anti-mimetic wine club, because the vibe here is as if we were just having a glass of wine after work, a fireside chat environment, very casual. Feel free to raise your hand if you have a comment. It's cool to jump in and answer each other's questions. Byrne and I will chat for... I've got a couple of questions for him, two or three, just to give you an intro to who he is a little bit about the paper. If you haven't read it, we'll just do a very high level introduction. And then at the absolute latest at 6:30, I would love to open up the floor because I'm sure there's going to be a lot of questions after we introduce the paper and the topic.

Luke (00:04:49):

So just to get started Byrne, I don't know the answer to this and I'm always curious to know, how did you first come into contact with Rene Girard's work and what was the genesis of this paper, which you co-authored with Tobias, by the way, who's not with us today. So your first contact with Girard and then where did the idea to write this paper come from?

Byrne (00:05:13):

Sure. Yeah. And first note, Tobias is not with us in the sense that he is in an inconvenient time zone and not that he's dead. Still alive.

Luke (00:05:21):

Good distinction.

Byrne (00:05:21):

But yeah. So how did I come across Girard? Like many people, probably most people who are in tech or tech adjacent, it was through Peter Teal. Specifically, there was an early profile of him in like 2007 or something, right about the hedge fund mostly. And it mentioned a bunch of really interesting stuff that I decided to look into more. And one of the things that it mentioned was, his fascination with Girard. So I went to the local library, picked up a copy of Things Hidden and, or no, it was Violence and the Sacred. And I read about 100 pages of it and it didn't click at all why this was such a big deal. It seemed like a lot of weird dark anthropology. There were all of these rifts where he is arguing with academics, who I also just didn't recognize and didn't have a start opinion about.

Byrne (00:06:12):

So I dropped it for a while and then I started coming back to Girard probably six or seven years ago, just older, more life experience, recognized a few more of the anthropologists who Girard is arguing with, but not all of them. And I started rereading it. And part of what appealed to me was just that Girard is an incredible analyst of the writing of other great novelists. And you sometimes you can read a great book and realize it had an impact and try to ask yourself why it had such an impact. And that's a difficult question to answer, like you know that reading say Crime and Punishment can change you, but it's hard to say what in the story, what in the ordering of these words actually cause that. Why do these characters resonate and feel so meaningful when other characters and other books just don't?

Byrne (00:07:11):

So I started doing more work about reading, and just more serious thought about his fundamental ideas. And I started like a lot of people, I think the time when it clicks is not necessarily when you're reading him, but it's when you're noticing the Rene Girard dynamics in other people's behavior. And then you realize, well, if this is true of any of the 8 billion people I can observe, it's probably true of me just as much. So yeah, Girard, he's a very useful checkpoint. A very useful way to evaluate your own behavior and other people's behavior. And I think turning back to the teal thing, there is something very mimetic about people seeing that this guy got rich and is allowed to do a bunch of really contrarian things.

Byrne (00:08:05):

So we're going try to reverse engineer it. And what we will do that is read all the source material and see if we can rebuild the engine that way. And then I think if you read enough of the source material, you get the stuff about conspiracies and secrets and realize that there's probably other influential stuff that is not going to be put on a syllabus reading list, but is still important.

Byrne (00:08:27):

So the mimetic effort there sort of Peters out after a while, but you do end up... Reading Girard, it's a really powerful mental exercise. It's a really powerful workout. And it's changed the way that I read fiction and has changed the way I look at human behavior. And then you can turn that to the markets question. And I've been fascinated with markets for a very, very long time. And I think one of the things that's convenient about them is just you have so much visible information that is always being exposed. Markets are this machine for instantly interpreting and reacting to every single thing that happens anywhere in the world.

Byrne (00:09:14):

And if you are trying to make money trading, you're obviously trying to do that better than everybody else, but your default assumption in general should be that when something important happens in the world, prices roughly reflect that. And it happens roughly instantaneously. So you can use markets as this real time look at what's happening, but then they have all of these other features because there's this glorious, beautiful machine, but it's entirely made of people and people have their normal human weaknesses and emotions.

Byrne (00:09:45):

And because everything's so quantified, you can actually look at a chart and you can see mania happening in a way that you couldn't really do just looking at someone's behavior. You can't quite see the peak. You can't see the collapse. You can't see when they're feeling this burst of hope that maybe they weren't wrong after all. And then, "No, actually I was wronger than I ever thought." So it's just a great source of information on the human condition. I think between novels and markets, that's your distribution of, if that's 20% of the material you can absorb in the world, it's maybe 80% of the knowledge about human behavior that you'll get. So yeah, that's where I got to thinking about the two.

Byrne (00:10:28):

And then in terms of combining the two, that was actually Tobias's idea because he got introduced to me through mutual friend because I'd written some things about bubbles and how I find bubbles really interesting. And he had also written some things about bubbles and the idea of bubbles as innovation accelerators that a lot of money gets wasted, but they also bring together a lot of talent and thinking more about that, I realize that bubbles are this fundamental important force that you can trace a lot of technological and economic progress back to bubbles. You can look at a lot of booms that's happening in the after effect of a previous bubble.

Byrne (00:11:11):

So I think like the super obvious version of that is that if you look at the internet companies of the 2000s, a lot of them were able to take advantage of the infrastructure that have been overbuilt in the 90s. So YouTube was able to take advantage of lots of cheap fiber, because there was lots of fiber people spent way too much money on in the 90s and Google was able to buy it for cheap.

Byrne (00:11:35):

And that goes back pretty far. The US still has some of the world's best freight rail infrastructure. And that is in part because people overestimated the amount of passenger rail demand in the late 19th century and built way too many railroads, but the railroads still exist. And in modern terms, what's more important is that the rights of place still exists. You can't build a physical thing in a straight line between any two big population centers without running into at least one endangered species or an AMBI. But if it's already built, it still exists and it's going to be of monopoly forever. So yeah.

Byrne (00:12:12):

I think bubbles are really important and there are a lot of different theories for why they come up about. You can talk to people who are, I guess you could think of it as progressive flavored skeptics where they'll say that bubbles are this just perfusion of greed. And then you have this rightest, or maybe libertarian leaning version where it's no, bubbles are always everywhere, just due to an increase in the quantity of money that eventually flows close to markets because they're able to produce this indefinite supply of things that you can trade money for. I think all the bubble theories out there, and there are many of them, they're all useful. They all point to something interesting and important.

Byrne (00:12:53):

And then what I think Girard adds to it is that there is just this consistent theme in bubbles that whatever people are betting on was a good idea when it started. And that the people who figured that out were generally really, really smart people who are also really ambitious and just the kinds of people you want to be in charge of organizations to try to change the world.

Byrne (00:13:12):

And then you end up with a lot of fakers. By the end of the bubble period you have everybody making these weird knockoff clones, or maybe the clone seems a lot cooler on the surface, but it's actually fundamentally pointless and doesn't do what the original does or doesn't even understand why the original is so great. But you have these two levels of mimetic desire. There you have one, someone who's starting a starting in the late 90s, they really want to be like Jeff Bezos. And then another level you have the investors who really want to have invested in Amazon at the IPO. Amazon had already gone up, but they can find us other thing.

Byrne (00:13:48):

So you have a lot of people who are copying each other badly, maybe copying their most visible traits. Someone who's copying their worst traits, and this happens a lot with the cult of Steve Jobs where like the things that set Steve Jobs apart, where he's a design genius and he's [inaudible 00:14:03]. One of those is really, really easy to do. And one of those is really, really, really hard. So which of those do people have more success in copying? It's probably the easy one.

Byrne (00:14:13):

And I think you can look at, you can look at a lot of companies that get copied and try to figure out what is like this core platonic thing that the company figured out or the individual figured out. And then the question of copying it is not how do you catch up to where they are today? But if you had that insight and had that attitude, what problem would you be solving today? If you were, if you were in their position today instead of where they were five years ago, and that's I think a healthy copying.

Byrne (00:14:41):

So there's this riff in the paper on Google and Amazon actually being deeply similar companies. And the basic idea is that one of the things Amazon figured out was the internet is a really good place to sell things. It's a really good place for people to find products they're looking for. But to do that, you actually need to categorize things in some searchable, readable tractable way. The really nice thing about books is that you have a bunch of metadata already. So you've got the title, you've got the author, you've got the genre, you've got the publication date. You can pull in a bunch of other data about the book. You basically have a lot of the search stuff done for you. So they were just able to slap at HTML front end on a database of what all the book wholesalers had in their inventory.

Byrne (00:15:27):

And then they had to do the shipping, which is non trivial problem, but they did it. And then they have the core of this working business and then they can move into less and less legible products. And because they understand what they're doing, they always can move in that direction of collapsing that legibility gap and creating more ways for people to spend money. And then you could look at Google as having seen another case where there's this latent abundant data source that exists, but it's not being turned into a useful product. In Google's case, it was link graph. So you can look at this set of all sites and the set of all links between those sites as this collective waited vote on what is important for someone looking to learn about what topic?

Byrne (00:16:12):

And it's not a coincidence that the Google people came out of academia where it is citation counts are really important. And you can do these algorithms that look at the citation count of a paper, waited by the citation counts with paper that cited. So they had some of the theoretical work done for them, and then had the insanely ambitious idea of doing this to the web, which as insanely as ambitious as it was, it would've been more ambitious a year later. I think at the time the web was growing faster than hardware was getting cheaper.

Byrne (00:16:47):

So it was actually the time to do that. As soon as you figured it out and with business, it was the same thing of like he realized the internet was a big deal. He realized it's growing insanely fast. It'll never be growing this fast again. And no other trend is would've matter as much as this trend. So they tried to get his company to do something about it. They said, no, he decided to go do the thing and the rest is history.

Luke (00:17:12):

So Bezos sort of bet on the mimesis of the internet in general and it didn't matter as much what they sold. I mean, books happened to be the thing that made the most sense for a variety of reasons that you talk about in the paper, but he just realized this is a mimetic phenomenon. It's going to continue to grow. And he wanted to get in early. What is... Yeah, sorry.

Byrne (00:17:32):

Yeah. I mean, there's a question of how much he internalized the medic nature of the internet, how much... I mean, some of it is just if you look at a line on a chart and the line is like, I think at the time the internet was growing 20-30% a year. So whatever is driving that it is important enough to jump on. But yeah, I think certainly one part of what drove the internet's growth was that people realized that it was cool. Their friends thought it was cool, so they decided to try it. That is one of the things that bubbles do really nicely is that they get everyone to build different components of a particular version of the future at the same time, so that you build everything you need for that future to happen.

Byrne (00:18:15):

So everyone who was over optimistic about e-commerce or online media, they start a company in that space. And then everyone who's over optimistic about the ISB business, they start laying cable and sending CDs to every household. And between those two the over optimistic ISP actually validates the e-commerce retailer or the online media companies business by adding more internet users. And then the over optimistic content companies and shopping companies give the people who newly signed up at that ISP, something to do online. So as long as everyone is over optimistic about different variants of the same thing, they actually end up building something that makes it not over optimistic after all.

Luke (00:18:59):

Sure. And then crypto and Web3 in general, it's an example of how you have a bunch of people building a bunch of different things and writing this thematic wave. And I think one of the most fascinating parts of your paper that I want to ask you about. And one of the things I don't think people realize about crypto is how leveraged it is. That the amount of leverage involved. I just want to talk a little bit about the role of leverage because I think there's a really interesting connection between debt and mimetic desire. And you quote somebody in the paper. I can't remember his name who says that leverage creates legibility, I believe. And my interpretation of that when I reread the paper last night, was that legibility, what does that mean? Well, it allows us to be able to read who's who a little bit better. And one of the things that leverage may do in my own words is to amplify differences, which makes it easier to see who's over-leverage and who's not. Who blows up. What's your take on the relationship between leverage and the mimetic desire?

Byrne (00:20:09):

Sure. So I think there are two parts of this. There's the point on leverage generally, which is that there's this taxonomy of bubbles where you have bubbles that are a bet on some wild divergence from history. And then you have bubbles that are a bet on the end of history, that things are going to be basically the same way that they were just more so and more so and more and more predictably. So if you compare the dot-com bubble and the housing bubble, dot-com bubble is the internet is going to change the way we do everything that turned out to be roughly true. And here we are on Zoom.

Byrne (00:20:41):

And then the housing bubble was nothing's going to change. People are always going to want suburban McMansions. The thing that'll change is we'll get better and better at predicting default rates. And because we'll be better at that, there'll be more supply capital going into buying new houses. And that means housing prices will keep going up. And all of this will be predictable. We'll have these incredibly elaborate financial products that are basically built with some understanding of what the underlying risk of the system are. And that allows them to enable more leverage than what otherwise exists.

Byrne (00:21:13):

So that is one piece of the leverage story. The other piece though, which is really interesting is that leverage is a way that you can catch up to someone who was smart before you were, but didn't bet as boldly as you're willing to bet. So the way to think about it is that in theory, given enough time, and it's going to be a lot of time. But given enough time, if you are always levered two to one, and you always own Amazon and you never get a margin call. If you maintain that level of leverage, eventually you will own more of Amazon than Jeff Bezos does because he's not levered. You are. So every time the stock goes up, 50%, you've doubled your money. You're compounding faster.

Byrne (00:21:56):

So leverage is just one of the ways that the mimetic desire is expressed that people realize I can't catch up to this person unless I am taking more risk than they are, taking the same fundamental risk, like making the same fundamental bet, but doing it in a more risky, more amplified way. And I think that can be literally true where you have people who come in late and the only way to be a Bitcoin billionaire by buying Bitcoin would be to use a lot of leverage unless you're already a billionaire from something else.

Byrne (00:22:27):

So the late entrance who are trying to catch up, they have to use more leverage to do that. And then I think what that does, like one place where legibility comes in is sometimes leverage reveals who didn't understand what the risks were. And there's this interesting phenomenon that people who are early to a business that really takes off, whether it's a category or a company, they tend to have a more measured view of that company or category than the most wild, ambitious speculator, or like the late joining speculators. In part, because they have more information. They've seen more of the risks, upsides downsides, and they just have this more realistic view. If you're looking at Tesla from the outside, you're not super well informed. You can have all these imaginary opinions on what they could do next and how lucrative it could be.

Byrne (00:23:23):

And what you can't know is how many of those ideas are things that Musk told someone to work on this for three months, back in 2015. They tried it out and they found out that for reasons, X, Y, and Z, this is actually not a very good move for us. So shouldn't be part of the long term plan. You can have more of this imagination driven upside view if you're not deeply involved.

Byrne (00:23:43):

And then when you're deeply involved, you realize that a lot of the business is a grind and a lot of these big tech companies, they have tens of thousands of employees or over 100,000 employees. In some cases, it's not like all of them are doing amazing, brilliant stuff all the time. A lot of them are just answering customer complaints or tweaking some algorithm to make it perform slightly faster, or slightly more accurate or whatever. They end up losing some of the glamor that you would get from the outside, because there's a lot of schlep involved in building anything great.

Luke (00:24:18):

No doubt. Just a reminder to everybody listening. If you have questions, please put them in the chat. And I've just got one more, because this is really for you, not for me. And we talked about this a little bit at the beginning, you wrote your paper in October of 2019. I would call it a metaphysics of markets trying to get underneath the surface and look at some first principles drawing on Girard. Would you update anything? What has changed? Anything new to say since the paper was written... What would you have included if you wrote the paper today, or is it just identifying fundamentals that are just there?

Luke (00:24:54):

I mean, I always think, how new is the meme stock thing? It was an unprecedented situation, but one of the things we talk about in your paper is there were guys like Jesse Livermore in the 1920s who was forming pools of like stock investors. And it reminded me of Reddit forms. I mean, it's the same thing, but it's just using a different form. Now we have technology to do that. So to what extent do you think that something unprecedented has happened versus this people have always been this way and this has been fairly perennial human behavior.

Byrne (00:25:32):

Yeah. I think for pretty much any market phenomenon, you can name, you can go back and find some version of it that happened way, way earlier and that was similar. Usually it's not as extreme as the one that you're thinking about just because the extreme things with the salient ones. So that's necessarily what you're comparing it to. But yeah, I think there are a lot of precedents for meme stocks in the 20s. And it is interesting that a lot of the meme stock stories, they were partly about the company fundamentals, but a lot of it was about how we're all in this together.

Byrne (00:26:07):

And I think the story in the 20s was a little bit different. It was more like you are an insider to this group of outsiders, we're taking on the establishment and we can control the prices. We can move RCA stock up if we want to, and then we'll move it back down where we feel like it. So in that sense slightly different story. But I think if you took a 20s era leveraged speculator transport them to early 2021, get them access to WallStreetBets. I think they'd fit in pretty well. I think they'd understand pretty much what was going on. They'd know how this game is played. They'd have a sense that it's not going to turn out well for everybody, but it's really fun while it's happening.

Byrne (00:26:55):

And then even in the 90s definitely had online forums that would try to manipulate stocks and that did have cheerleaders and haters and all this stuff. In some cases, well, not as big an impact necessarily in terms of market value, but there were definitely cases where small companies stocks would just go crazy. And actually, because I had been writing about that 20s era manipulation, and thinking about other cases of retail mania, I actually, I wrote a piece on WallStreetBets in, I think February of 2020, and talked about some of the shenanigans on the Yahoo finance message boards in 90s and how we're bringing back some of that stuff in a more modern way and at a larger scale. One of the differences between the nineties and today is that internet penetration and average time online for internet user have both gone way up. So everything that is an internet thing is just a bigger deal because the numbers all involved are all bigger.

Luke (00:27:56):

No doubt. Yeah. I mean, the paper covers so much ground and we don't have a whole lot of time. I mean, I think one of the statements that comes to mind immediately is people think of markets as impersonal, but they still indulge in ritual in your words. And you don't just find the mimesis in the markets. You find patterns, like ritualistic patterns. You even touch on Satoshi and Bitcoin and that there was this immaculate conception that everybody's now tried to copy. And perhaps one of the smartest things that he did was to disappear so that he couldn't very easily be singled out as some scapegoat, if things didn't go right.

Luke (00:28:38):

So the paper touches on almost every aspect of the mimetic theory, which is very broad. So we could go a lot of different directions with this. Quite a few people have now joined since we started. So I'm going to go ahead and step back. Now we already have one question from Jeremy rather than me read it. Jeremy, do you want to just jump right in and ask it yourself?

Guest (00:29:04):

Sure. Sorry. I'm cooking dinner as I listen. I'd love to more, obviously there's plenty of conversation around negatives of bubbles and all that and the boom and bust cycles. And people who are over leveraged, I think, especially in crypto, but I'd love to hear more about your thoughts on both positive releases at play because obviously every bubble, it seems like there's great things that come out of bubbles at the end and great companies that get filled. But also the way that it seems that things are just rapidly improving and these ski cycles just are just going quicker and quicker and quicker and where it will eventually end in your opinion, if at all.

Byrne (00:29:46):

Yeah, sure. So I think that one would think about positive mimesis would be that there are a number of tech people who are pretty thoughtful about tech history. And I think the further back you go in tech history and the deeper you go, the more that you'll see evidence that a lot of the things we think of as new are actually old. And I think that's especially useful because a lot of young entrepreneurs who are in a growth industry, they just can't internalize the fact that eventually every industry gets big enough that it's cyclical. And you have to have some sense of what that cycle will look like, because if your business model is built for growth and your pessimistic case is that the industry grows 10% this year and the base case is plus 30%. Well, you're not going to be well prepared for the year that things go down 30% instead.

Byrne (00:30:39):

And I think that the Collison's have actually done a pretty good job with that. Hold on. I'm on a call. [foreign language 00:30:51]. Sorry. The Collison's have done a good job on that. If you look at Patrick Collison's website, he has a ton of stuff on book recommendations on early tech history, like semiconductor industry and early software history. And I think that stuff is it's really valuable because the chip industry was the original growth tech industry and then turned into a more cyclical industry and went through some brutal downturns where a lot of the snaggy companies just got wiped out. And that does happen to every industry.

Byrne (00:31:28):

I think turning into the question of the broader social utility of bubbles, like definitely true. Bubbles create a lot of value. And in part the question of whether or not they are worthwhile, it partly comes down to discount rate. So if you look at them initially, they just destroy a lot of capital. And then over time things get built. There was some kernel of truth to the bubble, and then eventually the things that get built become very valuable and are often worth a lot more than what the initial cost was to build them, whether that is the wonderful internet we have today, even like the housing stock.

Byrne (00:32:04):

The US has housing shortage if you compare average historical rates of household formation to new housing starts. Like we have a housing shortage now and in some ways it's quite fortunate that we've just built tons and tons of housing in the mid 2000s. And going earlier, looking at things like the, say the conglomerates level that's written about in the paper, that was a case where there were a lot of practices at big companies that were just not fully modernized.

Byrne (00:32:37):

And it was useful to have some force that swept in, and that caused some underperforming companies to get bought by companies that were run by people who maybe weren't great managers, but had figured out some of the basics of accounting and understood what understood some things about how to run a business more efficiently. And we see this happening today with private equity, where a lot of people who go to work for private equity companies are extremely sharp, extremely hardworking, and they are the kinds of people you'd want running businesses. And in fact, I have this meta theory that if you think about what someone's life path might have been, if they were smart, interested in business, and it's like 1955, they might have tried to work their way up to running a paper mill or something.

Byrne (00:33:26):

And now it would be incredibly uncool to get a Harvard MBA and then go work for a paper bill and hope that in 20 years you are the COO. But if you work for a private equity firm and they buy out a paper mill and you are the person put on that deal, then that is cool because you're working for a private equity firm. You're still doing the same things you would do in the other scenario. So it's basically this conspiracy to get talented Mid-westerners who go to Ivy League Schools to go back to the Midwest and run just normal person industrial companies, instead of doing something like just working at Goldman trading currency directors or whatever.

Byrne (00:34:06):

And I think moving human capital around is really valuable. That's one of the things bubbles do. It's a huge function of the financial industry. The financial industry moves capital around, but that capital is especially in tech is largely used to hire people. So really the industry is trying to distribute the talent to where talent should go, but you can't do that top down. So you have to do that by allocating capital instead.

Luke (00:34:28):

Thank you. Any thoughts? Comments on that? Okay. Couple more questions. If you missed the beginning, this is very open. Feel free to jump in. Otherwise, we'll just keep moving through questions. David, you've got a question on the tension between access to investors for certain people in a company and walling them off.

Guest (00:35:00):

Yeah. And you just helped me see with new eyes this tension between where you were giving examples or talking about where there have been in different bubbles. It's almost been a discipline or an anti-mimetic focus tool of bringing walling off the development team at different points so that they only feed or mimetically imitate each other versus letting them have access to the customer base, to the investors, to the marketplace or even to other competitors or other imitating companies. I hadn't thought about that formally as a method of almost not anti-mimetic, but more focused mimesis. I wonder if you had more thoughts since you had written the article or more examples over the year, positive and negative on how that tension can work positively, negatively when it's generative, when it becomes destructive.

Byrne (00:35:59):

Yeah. So I think that there's an interesting example in finance, which is Renaissance Technologies, the famous quant fund, where they are located near New York city, but actually inconveniently far away. So they're way out on long island. And I think part of the decision was just that's where the founder was. So that's where he put his company, but it turned out to be this way to separate them from the culture of the rest of the financial industry. And you can see some companies try to separate their employees' culture from other companies through... Sometimes it's stuff like dress code where, and this didn't go either way. I forget there's some tech company where they... I don't remember. I think they were a chip company where during the period when business casual was getting really big, they actually made all their employees wear suits.

Byrne (00:36:53):

And you could really tell that those people were at this particular company, but the usual way it goes is that you're in a business where everything's pretty stuffy and informal. And this is the one company where they have bean bag chairs in the office and everyone's wearing t-shirts. So sometimes enforcing that separation can work. Sometimes it is about location. And I think in early tech, there was this tendency for people to... The West Coast tech companies did want to maintain that West Coast vibe. They didn't want to become the next IBM. They were already viewing IBM as a business in decline. So they did try to keep their employees on the West Coast. There's this anecdote about how Intel would send employees to the East Coast for one conference a year that was like, it took place in... It was in Buffalo, New York and February. So it was basically giving people a once a year warning of, if you go work for one of the big East Coast companies, winter is going to be terrible. So come back to sunny, California and enjoy.

Byrne (00:37:55):

I think you can enforce it that way. Sometimes you do have levels of secrecy within a company that might enable that. I think Apple has that very hermetically sealed design approach to getting people on different teams. There's some story on Quora where someone had to... He was working at apple and he was given an NDA and the NDA only covered telling him the code name of the project so that he could read the NDA for project purple. And then once he had signed the NDA for project purple, then he was allowed to know that it's an iPod.

Byrne (00:38:33):

So yeah, they had a lot of layers of secrecy and that does mean that you get fewer ideas floating around, but you also get more people focused on this one thing they're building and probably the weirder the thing is the more of a departure is from the rest of the company. The more that you want them to focus on what that singular vision is rather than making lots of incremental adjustments based on what the market thinks at once right now.

Byrne (00:38:57):

And I think for a lot of those big projects, whether it's something like the iPod or then the iPhone or Gmail, for example, if you did market research, you'd end up lowering your ambitions because like if you take Gmail, for example, I think at the time that Gmail was launched, Hotmail had just increased free storage from two megabytes to four megabytes. So when Gmail offered gigabyte, that was just a huge deal. So they launched at April 1st. People thought it was a joke, but I think if they'd done more market research, they would've realized that they could launch with 50 megabytes.

Byrne (00:39:30):

But what turned out to be true was that if you give people what seems like an infinite amount of storage, they will eventually fill it up and then you'll be able to charge them for that. And you also get just tons and tons of data from having more email. So it turned out to be the right decision. But I think the more that you think about it as how would I do a clean room design of what this should work like? And the extent that there's market feedback, maybe it's market feedback within the company, because Google was a very email centric company.

Byrne (00:39:59):

So people had strong opinions on email. They were really good at using their inbox as quickly. And they, they prioritized speed in that Gmail launched with really great keyboard shortcuts in part, because a lot of the tools that a hardcore email user would've used at the time, would've been things like, I think it's pine. It's like a text based terminal based email reader. And it's really, really fast. And if you know the keyboard shortcuts, you can get through an inbox insanely quickly.

Byrne (00:40:26):

That feedback of this is the stuff that actually matters to us on the margin and we are the users who have the strongest opinions and we will actually use the product. I think that can be useful feedback. But in that case, it's not so much... Not really the mimetic. It's not like they're saying, here's what I like about Gmail. They're saying I wouldn't even consider using the web-based products that exist today because here's what I like about the thing that I'm actually using. So here's what a new product has to satisfy.

Luke (00:40:57):

So couple of questions that both touch on culture from Brendan and Patrick. Brendan says this in zero to one, make your team look different in the same way. You mentioned in the paper Byrne and I think one of my favorite lines in the whole thing is that everyone seeks to be the most original looking copy of a model that they don't fully understand. So we're all copying something, but there are parts of that model that are hidden. There's this weird cargo cult thing going on, where we're imitating all of the wrong things. I think Patrick... We do Patrick first and then Brendan. But Patrick's question is related to culture. Patrick, you want to ask or elaborate on that a little bit?

Guest (00:41:37):

Yeah, for sure. And just a little context here and thanks Brendan for the follow up as well. So I actually had a positive back and forth engagement on Twitter. Someone had tweeted about there hasn't really been a lot of study about the PayPal Mafia and just the overall trajectory of the company and why can't it be recreated? And I said, it reminds me a lot of an NBA super team. You can stack the deck as much as it looks good on paper, but they're not really sustainable over time. So just digging deeper on that and having read the founders as well, understanding just the randomness basically of everyone's story. A lot of immigration, people coming together at the right place, right time in life. I just think it's impossible to recreate and yes, you can certainly try, but I don't know if there's actually a good example of something similar or maybe even really bad examples, which I'm sure there are too.

Byrne (00:42:42):

Yeah. So on the PayPal Mafia, I wrote a piece on this a while ago asking why business mafia's are so rare. And there are a lot of ex-Googlers out there, but there's not really an ex Google Mafia. It's not like if you worked at Google, you can definitely raise money from all these people who worked at Google or still work at Google. So what I realized was there is actually one pretty comparable this or pretty comparable case study, which is PEGA Management. So was one of the most successful hedge funds. One of the largest in the world. Shut down in the year 2000, and then basically the founder shut down the fund, but kept his lease and told his best employees keep working, start your own funds. I'll give you seed capital, keep in touch with each other, et cetera.

Byrne (00:43:26):

And now those funds in the aggregate managed far more capital than Tiger Management ever did. And some of them individually managed more than Tiger Management ever did. So that was a hugely successful mafia. And I think there are two points in common here. One is when Tiger Management shut down, it was because they had, they were pretty value focused and they had made some bad bets on macro, in the late 90s. And then they just owned a ton of old line industrials and were betting against tech. And that combination gave them really poor returns in the late 90s. And then had they held onto those positions, it would've been a phenomenal performer in 2001, 2002, but they didn't. They shut down.

Byrne (00:44:11):

And then with PayPal, when they sold there's good early Quora stuff, you do want to do some Quora archeology. Back when Quora was invite only a lot of exp-PayPal people got invited and some of them basically put together informal oral history of pre-IPO PayPal. And one of things they said was they were always worried that eBay would somehow someday find some way to crush them. So that's an existential risk that they were so reliant on eBay. And the other thing was because they had so much data on eBay transactions, they could see how eBay was doing.

Byrne (00:44:44):

So part of the reason that they sold was apparently that they actually detected a deterioration in eBay, which would've been a deterioration in their own performance later on and figured we can sell, get a premium now and just go do our thing. So that's one piece is bull companies have this exit that is well before they fulfilled their full ambitions. And it's because they hit really serious speed bump that could have been an existential threat and they wanted to avoid it.

Byrne (00:45:10):

And then the other thing is if you read PayPal's last 10K, when they disclose their employee count, I think it's roughly 150 or 200 people who are in the Bay Area office. And then they have a bunch of customer service people in Nebraska. And then Tiger was also about 150 or 200 people when it shut down. So what's neat about that is that those numbers are pretty close to the Dunbar number of how many real relationships could you have? People always quoted it as 150, but I think Dunbar gave a range of like 100 to 250. And if you ever read about, it's a fun study looking at primate social groups relative to primate skull size and inferring from that the natural size of a human tribe is 100 to 250 members.

Byrne (00:46:00):

Now, if you think about what it's like to work at either a hedge fund or at a rapidly growing tech company like PayPal, you would often join when you're really young. So you probably went to college and then you moved to that city to work at that company. So you abandoned a lot of your early social ties, and then you form social ties, but you're working all the time. So the social ties you form are not people in your neighborhood or people you've bumped into at the place where you get long lunches every day. No, it's people who work with you. People who work with those people, et cetera.

Byrne (00:46:33):

So your Dunbar number gets fully saturated only by people at that institution. And then they all leave. And because of the timing, they generally leave with enough money that they could do a lot of different things that they don't have to work. But not enough money that they feel like they're done and they've accomplished their life's work. So the ones who left with a lot, they fund everybody else. The ones who left with a bit of money, maybe they know that they could start a business. They don't have to pay themselves a salary for the first year or two. So they actually have a little bit of personal runway. And that starts the cycle of all these people starting businesses.

Byrne (00:47:08):

So I think that's why they're rare is that good companies are hard to kill. So the only way to get a mafia is a company that is like lucky enough to get killed. Although, the luck thing is tricky. Until recently it was true that the total value of all the businesses in the PayPal Mafia, while it was big, was actually smaller than the total value of Google. So it may be, this is really small sample size, but the mafia story may be cooler, but the story where there's no mafia because the company never sold or never shut down is probably the story where those people make a lot more money.

Luke (00:47:52):

Cool. Brendan. Anybody else? We've got about 13 minutes left? I've got plenty of questions, but I want to keep the floor open.

Guest (00:47:58):

Yeah. Mine wasn't really a question. Just to comment on separating your people from investors or other companies.

Luke (00:48:09):

Okay. Yeah. So this question, Tiger calls to mind all of the different investors that are imitating Tiger. I mean, just in the last few years they have a very specific model. I had somebody tell me six months ago that the numbers work. They're running on math. There's stories about them never having met a founder and sending them a term sheet while the founder's out for dinner and telling him he's got 12 hours to respond or something like that. And the breakneck speed, I think that's all changing. It just makes me wonder, when it comes to imitating these models, it seems like if a firm like Tiger or a really good investor would have certain parts about what they're doing that they want to hide or that they want to be transparent.

Luke (00:49:01):

I don't know if you have any thoughts on this, but this and there's something that, because you'd mentioned. We don't fully understand the models that we're trying to copy. And to what extent are certain investors, because we're really talking about largely about investing in markets today. So I'll focus on investors. To what extent are investors welcoming imitation? And in what ways are they trying to put forward a front, but keep their most valuable secrets to themselves? And what are the consequences of this weird dynamic of imitating things that we might not really fully see?

Byrne (00:49:44):

Yeah, that's a good one. And I think it's a big topic. So I think you have a couple things going on there. One is that there is just this tendency for a lot of investors to share ideas. And I think it's somewhat healthy. You could end up with a lot of crowding, but a lot of investors, they want to know what the case against their view is because they'll lose money if they don't know it and it turns be true. So you do have a lot of ideas being shared. There was a time when... I forget where Tiger is located. But in the like 2017 period, I was telling someone about a stock and he said, "Yeah. That's a 40 Park Ave name." And we meant was there are all these Tiger spinoffs that all were located at 40 Park Ave because that's where Tiger's original office was.

Byrne (00:50:33):

And they all talked to each other, they all know the case for this company. They all have the same case and are using roughly the same way to think about it. And they all owned it. And that's great when they're doing well, they're raising money. When they raise money, they add their positions. So the stock goes up. It's not so great when the stock goes down and they all have to liquidate at once and they're all pretty levered, or they're all somewhat levered, so goes very badly. So you do have a case of crowding like that.

Byrne (00:51:02):

I would say in terms of investors how much they disclose. Yeah. There's definitely some back and forth there. So at one level, they want to be transparent enough that they can make the case that their process is repeatable because what a lot of hedge fund investors are looking for is some repeatable, predictable source of excess returns. And if you tell someone, well I think really hard about what's going on in the world and then bet on what the future will be like, "Well, we don't know if you did that and got lucky. We don't know if you're brilliant. We have no way to reverse engineer that process."

Byrne (00:51:39):

So they often want to provide some legibility to their investors to explain how this works. Legibility also helps with internal scaling. It does mean if you have some process that works really, really well with software companies, maybe you figure out what parts of it analogized to say retailers, and then you spin up a team doing that.

Byrne (00:51:58):

And now you've been able to expand, but you've also expanded your monoculture a little bit. So if you think about these companies in similar ways, but they are fundamentally different, that can run into problems. And often if you were able to convince yourself that these businesses are fundamentally similar, it's probably because they're both doing well at the time. And then when they do badly, you learn that they do badly in very different ways. So you can run into that problem.

Byrne (00:52:23):

I think there is another layer of indirection or misdirection that happens less on the discretionary side, but a lot more on the systematic side, because with systematic investing where you are building an algorithm that is going to consistently trade some asset and make a return, well, a lot of the work goes into things like cleaning the data that turns out to be a really big deal. But some of it goes into pursuing a bunch of different ideas, finding out that 95% of them don't work. And then you find one thing that for whatever reason works pretty well.

Byrne (00:52:57):

So most of your output is a very small function of your input, which means that if someone knows what you're doing, they can save a lot of time in implementing it. So that makes quant systematic traders just incredibly cagey about what they do. And sometimes it's pathological. I'm actually working on a piece that I'll be writing about a pretty well known quant firm. And I found an article talking about one of their founders where, they refused to give an interview and refused many times to give an interview. The journalist said they called four times, but also he refused to disclose the names of the co-founders of the company, which is insane because they're on the website.

Byrne (00:53:40):

So it's like just as insanely reflexive insistence on never telling anyone anything, because everything you tell them, even if you say here's what we don't do. Well, that's a clue on what doesn't work. So the more you respect this person, the more that tells you should be looking at some other opportunity instead. So what some quant funds will do is they'll have fairly simplistic strategies that work or some of them will. But then if you ask them what they do well, they'll say we aggregate ex terabytes of data and we use all these fancy machine learning algorithms, and we have all these brilliant people. Sometimes that's part of what they do. And it certainly sounds cooler than what they actually do. But if they've actually found some fairly dumb rule that actually works pretty well. It's way better for them to tell people a harder version of it, rather than an easy version.

Byrne (00:54:33):

And that is a useful heuristic that goes beyond finance because you'll see this with tech companies, when they give you technical specifics on what they do, like Facebook has done this on their machine learning for detecting the quality of comments and the tone of comments and things. And Google has done this with a lot of their AI stuff. When they disclose things, it's almost certainly something where you couldn't copy it, except at their scale.

Byrne (00:54:59):

So Google is never going to say anything publicly that could make Bing better. But they can say publicly things where if Bing had 30% market share, this would be a really good idea, but at 10% market share, it's actually a huge waste of money because it's so expensive. It doesn't scale well. So that's I guess, a useful way to read this stuff. It's a really cynical way to read the way that tech companies talk about their technology, but probably accurate to say anything they're telling you how to do, you should not actually bother copying. It's only the stuff that they don't tell you how they do it, where there might be some fairly simple magic trick that could in principle be discovered, and that you'd be able to achieve some level of parity with something that they do. You're probably not going to build the next great search engine, but maybe.

Luke (00:55:48):

Got you. Patrick got his hand raised. We've got a few minutes.

Guest (00:55:53):

Yeah. May maybe quickly just the other big current event is this whole like Elon and Twitter deal going on. To me, it just seems like more of a psychology battle in a negotiation tactics or philosophy versus a financial transaction. So we'd just love to hear you weigh in on what's going on there.

Byrne (00:56:15):

Yeah. I mean maybe one way to look at it is that a lot of rich people own media outlets. They often own media outlets that they have some affinity for or that are just really recognizable brand names. Musk obviously loves using Twitter as a product. I think Matt Levine has done really good coverage of this. And he's pointed out that if Musk does buy Twitter, he will be the only member of Twitter management who actually uses and likes Twitter as a product. Musk is always hard to figure out because he does have a weird personality, but he also knows that's an asset. So you have to wonder for his behavior, how much of this is Elon Musk, the person making business decisions and how much of this is Elon Musk, the celebrity doing celebrity things that continue to improve the brand, and that might be useful for other future business decisions.

Byrne (00:57:06):

It's kept him in the news a lot and kept him a lively topic on Twitter and probably helps sell more Teslas. So I don't know that it's worth 44 billion. I don't think it sells 44 billion worth of Teslas, but it does help with some of that. But I do think that to the extent that you want to bring it back to the mimetic question, if he looks at Bezos and says, "Okay. Bezos gets the post, times is not for sale. What are the big media outlets that you could actually buy and have influence over?" And Twitter happened to be available and happened to be one he was familiar with.

Byrne (00:57:46):

And also, I wonder if there is some runaway dynamic where he buys some Twitter stock as a troll, people get mad, he decides to escalate, which is if you're a troll and you're not worried about getting banned, which the SCC has not been able to band Musk. If you're not worried about getting banned, the right thing to do is always keep on escalating because you can make people madder at you faster than you get mad at them. And then they'll eventually do stupid things and you'll have a really good time.

Byrne (00:58:15):

So I think you can view some of it as potentially trolling, just the troll you can do if you're a certain billionaire that is not available to the average person. And then there's always the weird possibility that Twitter is notoriously not that well run. Their launch cadence has historically been pretty slow. They've fixed some of that. Their monetization has historically been weak. They've always had this huge cultural impact that is not reflected in their financials. And Facebook is weirdly the other way around where for all their cultural impact... The financial impact is a bigger deal. It's funny, if you go back and read... There's this book, the Facebook Effect that came out right before the IPO, that's a story of Facebook.

Byrne (00:58:59):

And it opens by talking about the really wonderful, amazing, positive impact that Facebook has in the world by affecting the outcomes of elections. And it's talking about a Brazilian election that I think no one in the US remembers in 2011, 2012 and how people used Facebook to get attention for their candidate. And it was able to build this groundswell of support for someone. So Facebook's been doing that for a long time and is probably not doing that much more of it than it was, or maybe pro rata for its share of usage and time maybe doing less because more of that time has moved to Instagram.

Byrne (00:59:34):

So in some ways, Facebook has become a bigger financial phenomenon relative to its cultural phenomenon status. And Twitter has remained big cultural phenomenon that just as a financial thing has been unimpressive. And Musk was able to start a car company that reported an accounting profit. And it's been generations since someone has been able to do that in the United States. So clearly knows something about business. He was able to get launch costs. Launch costs had stagnated, I think from the 70s through SpaceX. And now they're going down faster than before. So he's smart and able to tackle really hard problems. And it may be that he views monetizing Twitter effectively as in the category of problems that are too hard for most people, not too hard for him. So he'll do it.

Luke (01:00:25):

Yeah. Thank you. Yeah, we can't really forget the weird mimetic rivalry thing going on between Bezos and Elon. As soon as the news breaks, Bezos gets on Twitter, got some great memes, gets political on Twitter around the same time that Elon does. So we could talk about that for a while, but I've got a whole theory behind that.

Luke (01:00:44):

Byrne, thank you so much for, for doing this. Man, your paper came out like while I was in the depths of writing my book and it was hands down, head and shoulders above anything else that I've ever seen with mimetic theory applied to financial markets. It's the best thing out there. If you haven't read it yet, I highly recommend it. And you've just been so generous with your time. I learned a lot tonight. I hope everybody else did too. I know you got a family waiting for you, so I'll be respectful of your time. And just want to thank you again on behalf of everybody here. It's been great.

Byrne (01:01:18):

Yeah, absolutely. Thanks for having me. I really appreciate the questions. This was a really, really fun discussion.

Luke (01:01:24):

You bet. Thanks again, Byrne. Everybody else will have Demetri Kofinas from Hidden Forces next month. And I will circulate this to the premium subscribers in the next couple of days. Have a good night, everybody.

Byrne (01:01:36):

Cool. Thank you.

1 Comment
Luke Burgis Newsletter
Audio & Recordings
For premium subscribers, this is a collection of audio recordings (my reading of essays), community Zoom calls, and an occasional podcast.