Dwarkesh Podcast
Dwarkesh Podcast
Byrne Hobart - FTX, Drugs, Twitter, Taiwan, & Monasticism

Byrne Hobart - FTX, Drugs, Twitter, Taiwan, & Monasticism

what happened at FTX, how drugs explain past financial bubbles, whether Musk Twitter will succeed, and much much more!

Perhaps the most interesting episode so far.

Byrne Hobart writes at thediff.co, analyzing inflections in finance and tech.

He explains:

  • What happened at FTX

  • How drugs have induced past financial bubbles

  • How to be long AI while hedging Taiwan invasion

  • Whether Musk’s Twitter takeover will succeed

  • Where to find the next Napoleon and LBJ

  • & ultimately how society can deal with those who seek domination and recognition

Watch on YouTube. Listen on Apple Podcasts, Spotify, or any other podcast platform. Read the full transcript here.

Follow me on Twitter for updates on future episodes.

If you enjoy this episode, I would be super grateful if you shared it. Post it on Twitter, send it to your friends & group chats, and throw it up wherever else people might find it. Can’t exaggerate how much it helps a small podcast like mine.


A huge thanks to Graham Bessellieu for editing this podcast.


(0:00:50) - What the hell happened at FTX?

(0:07:03) - How SBF Faked Being a Genius: 

(0:12:23) - Drugs Explain Financial Bubbles

(0:17:12) - On Founder Physiognomy

(0:21:02) - Indexing Parental Involvement in Raising Talented Kids

(0:30:35) - Where are all the Caro-level Biographers?

(0:39:03) - Where are today's Great Founders? 

(0:48:29) - Micro Writing -> Macro Understanding

(0:51:48) - Elon's Twitter Takeover

(1:00:50) - Does Big Tech & West Have Great People?

(1:11:34) - Philosophical Fanatics and Effective Altruism 

(1:17:17) - What Great Founders Have In Common

(1:19:56) - Thinkers vs. Analyzers

(1:25:40) - Taiwan Invasion bets & AI Timelines

Dwarkesh Patel 0:00:00

Today I have the pleasure of interviewing Byrne Hobart again. He writes at thediff.co and the way I would describe Byrne is — every time I have a question about a concept or an event in finance, I google the name of that event or concept into Google and put in ‘Byrne Hobart’ at the end of that search query and 9 times out of 10, it's the best thing I've read about that topic. It's just the most schizophrenic and galaxy brain takes about how Machiavelli's discourses relate to Big Tech or how Soros’ theory of serial reflexivity explains hiring in finance and tech. Just very interesting stuff. I'm glad to have him back on again.

Byrne Hobart 0:00:47

Great to be back.

Dwarkesh Patel 0:00:50

First, I really want to jump into the FTX saga. What the hell happened? Let me just leave that as an open ended question for you.

Byrne Hobart 0:00:59

I think the first thing to say is that there's a lot we don't know and there's a lot we may never know because so many of the decisions at FTX were made through auto deleting encrypted chat. So there are some holes we will never be able to fill in. The lack of accounting is also going to make it tough. 

Basically, I think you can tell a bunch of different stories here. The really obvious one is fraud and you can debate over exactly when it started. One version of the story which is getting some currency is that SBF had this entity Alameda and it was supposed to be this really hot crypto trading fund but maybe it was a Ponzi scheme all along. And then at some point that Ponzi scheme started to run short on cash so he decided to start an exchange and the exchange got more cash and then he used the cash to pay off previous debts. 

That's one version and then the maximally exculpatory version, which actually is still really bad is — Alameda was a real company. They really made money trading. They took tons of risks. SBF has talked about why he thinks that's a good thing. FTX cut some corners when they were raising money and they had really bad internal accounting. And the extended entity of Alameda and FTX sort of lost track of whose money was where and it ended up with Alameda spending FTX’s customer money. 

One way to look at that is you think “Fraud is twice as bad as just incompetently losing money.” Well, it's not as if we had a $4 billion fraud instead of $8 billion fraud, everyone would be like, “Well, that's fine. That's normal. Why are you giving SBF a hard time?” It's bad no matter what. Running a big company that is systemically important in crypto and then having that company completely vaporize over the course of a couple of days is really, really bad and it’s worth understanding what happened. 

It's also partly worth understanding what happened because there are just different solutions that present themselves depending on what you think the story is. If the story is fraud, it's actually a lot harder to solve because there are just a lot of people who are willing and able to commit fraud and to lie. If the story is bad accounting, then that's actually a lot more solvable because then you could say things like, the solution is make sure you never invest in a crypto exchange that doesn't have a real auditor and make sure that they have their proof of reserves calculation and it's happening consistently and that you can audit that. There are different solution sets. 

I think the actual story is going to be somewhere in the middle. Extreme risk tolerance plus extremely poor accounting plus fraud at some point. But I suspect the fraud, if it happened, actually happened pretty late. I think there's like a 80, 90% chance that there was some level of fraud versus pure incompetence. But if so, I think it may have happened fairly late in the story and as kind of a last desperate move. 

I think part of what drives the response to what happened with FTX and Alameda is that if you think the story is pure fraud, it's very easy to say you would never do that. I can say very easily, I would definitely never start a Ponzi scheme and then start another bigger Ponzi scheme to pay off the first Ponzi scheme. That's not me. That's not most people. But I think if you draw the scenario where they discover at some point, like a couple months ago or even a month ago, they realized, “Hey, we actually there's a billion dollars plus that was supposed to be customer money, but we thought it was Alameda money and we actually spent it and now it's gone. We've lost it.” What would you do in that circumstance? And I think the ideal answer is, “Well, I'd immediately come clean and step down and commit myself to getting everyone paid back and made whole.” But there's also the possibility that the realistic answer is more like, “Well, I would scramble and try to make sure that that didn't cause the company to collapse and then try to pick up later.” And so at that point, you've sort of backed your way into fraud through earlier episodes of incompetence. 

I think one of the problems with the fraud story is that frauds have to be good at accounting. In a very rough schematic sense, they have to be twice as good at accounting as everybody else, because not only do they have to have the real books that tell them how much money the business has and whether or not the next check they send will bounce, but they have to have the fake set of books and they have to have a way to make those tie out with one another. So actually accounting frauds tend to be fairly sophisticated. They tend to really dive into edge cases. 

I was reading up on MF Global which was a big futures brokerage that collapsed in part because they were dipping into customer funds and making some investments they shouldn't have. They did a lot of clever and shady stuff. There was one point where they were transferring money at the last minute out of their customer funds in order to make margin calls and what they would do is, they would send the wire from the customer account to a different company account and they'd send it a couple of minutes before the wires closed for the night. And then they would send this email right after the wires closed saying, “Hey, we just realized we sent this transfer from our account. Got to reverse tomorrow.” But that gave them at least one night of enough liquidity to survive. Now, you can only do that kind of fraud if you are actually keeping really close track of where your money is, where it's supposed to be, what the rules are, so that you know exactly how to break those rules. 

I don't think FTX was in any position to commit that kind of fraud. I think that if they tried to do something like that, they would wire the money from an account that didn't have any money in it or something or send it to the wrong account. There are these stories about them accidentally burning a bunch of USDC by sending it to an address that didn't exist or something like that. The operational slip-ups actually make it harder for them to have committed fraud and it's unquestionable at this point that their record keeping was very bad.

Dwarkesh Patel 0:07:03

Yeah, to your point about the fraud being harder, it's like the classic story about if you just tell the truth, it's just gonna be much easier for you. You just don't have to keep track of that many things. 

I interviewed SBF for like an hour and before that I tried to do quite a bit of research into how FTX worked and what was going on. I had this impression that this guy was the most competent genius that had ever graced finance. And this was a common impression, it wasn't just me. Then it turns out that out of sheer incompetence he loses track of billions of dollars, the internal operations are just him putting together spreadsheets and throwing them around and putting emojis on slack messages asking for payments. 

I want to understand how it is that this guy put out the impression out there that he is just hyper competent and it turns out that it's the opposite. It's not even that he's mediocre, it's the opposite.

Byrne Hobart 0:08:09

Right. I think you can tell a couple stories there. I know I've been saying that you can tell multiple stories a lot but there are multiple stories that fit the facts. We have lots of different weird things to explain and therefore many different weird explanations that fit them. I think one version is, he's never all that smart and decided that he could just play up this weird, eccentric genius thing and that would help him get away with it. There are these anecdotes about how someone told him to cut his hair and he said, “No, I have to look kind of crazy for this.” so that fits in. And it is kind of an MIT thing to do that. To play up your eccentricity because you know there are these super brilliant, very eccentric people and you can be like them. 

It's kind of like how a lot of people read about Steve Jobs and they're like, “Well, the secret to success is to be this brilliant perfectionist who can always see the future and also be just a giant asshole to everyone you meet and I'm going to try to do both of those things.” And it turns out one of those is really, really easy to do and the other is really, really hard and you have to do both to be Steve Jobs. But you can sort of give this surface level impression of Jobsianess by just being really obnoxious to everyone. 

Some of it is that but the other is that if you get really good at just very narrow domain specific stuff, you might miss what other stuff people have to be good at for that skill set to be valuable. Thinking about his previous background where he worked at Jane Street, a prop trading firm, and seemed to do well there. They're very, very selective with who they hire, very hard to get in and they're very profitable. So it’s good to get in. It's entirely possible that part of what happened was just that Jane Street has its operations people and they have their trading people. There may have been enough siloing within that, that if your job is just identify discrepancies in ETF prices and take advantage of them, you don't actually have to know things like — How do we figure out which counterparties are credit worthy? How do we make sure we have enough liquidity? How do we have backup plans upon backup plans upon backup plans in case something goes wrong with our liquidity situation? They're very opaque in terms of their trading operations but part of the model seems to be that they want to be the trader who is there trading and making a market when everything falls apart. And what that means is that the way you make the most money in trading is when markets are insanely volatile, volume is very, very high, and you're still trading. But the reason that markets get really volatile when prices collapse and there's a lot of trade going on is that other people who would love to be trading can't trade because maybe the broker they use is suddenly insolvent and they can't get to a new broker, their money is frozen. So if you're planning to be there when everybody else is out of the market, then you have to have lots and lots of contingency plans. It's not enough to buy lots of deep out of the money put options as Jane Street does, you also have to make sure that you're buying those options from a counterparty who will actually send you the money when you need it. You want to structure those things so the actual cash gets to your account at the time that needs to be there. And that maybe is something that a prop trader should not be spending most of their time thinking about. 

It's one of those things where it's like, if you own a house and over the last 24 hours you learned a whole lot about electrical wiring or you learned a whole lot about how plumbing works or how septic tanks work, that's not good. That means something very, very bad happened in your house. It could be nice to be an expert on those things but if you suddenly became an expert it's because somebody else wasn't doing their job. So I think you could be a trader like that where they can be very good at finding little pricing discrepancies and have just no awareness of what the operation stuff is. Especially because the better the operations team is, the less anyone else needs to be aware of them. You only email them when something is going wrong. So if nothing is going wrong, you never email them and then you forget they exist.

Dwarkesh Patel 0:12:23

Yeah, that's a good point. In fact, in the interview I did of him, I asked him what is the difference between Jane Street and FTX? And he mentioned that at Jane Street, there was this button he could press to buy and all the intermediaries, all the servers and all the bank accounts were just taken care of. What was really funny is he then said, “Let’s talk about that. Just getting a bank account is so hard when you're independent.” Apparently it turns out that it's so hard that you might have commingled funds because you couldn't manage to separate them out. 

You had this really interesting take. At one point we were talking about how every single market crash can be explained by the drug that was common in the industry at the time. And we finally achieved the hypergrade meth stage of that Emsam patch he was taking that’s stronger than Adderall.

Byrne Hobart 0:13:23

I think I was saying every crash can be explained by the drug they're taking at the time. That takes it a little far but I do think that the impact of new drugs on financial markets is underrated. You can have examples of this going back pretty far. 

There is some connection between caffeine consumption and extroversion and risk taking. You temporarily get a little bit more willing to do deals when you consume caffeine. Before Lloyd's of London was this insurance consortium, it was Lloyd's coffee shop. So you do have some history of coffee shops being associated with financial centers. 

And then you have to zoom forward because we just haven't had that many novel stimulants. Depressants, deliriums, and other drug categories probably just don't lead to that much financial activity. I don't know how someone would trade differently or invest differently if they had a really strong acid trip or took ecstasy or something. But the stimulants where people can just consistently reuse them, they keep people alert, they make them active and wanting to do things, it seems those would have a connection to financial markets. 

The theory is — if you look at the 1980s where there were a lot of these hostile takeover deals where someone would find a company that's underperforming. When you look at the spreadsheets and say this company is underperforming, what you're often looking at is a story that is more like, this company believes that they have this social obligation to the community where people work and that they have an obligation to give their customers a fairly priced product and maybe they give them really good customer service that doesn't really pay for itself and it's the right thing to do. 

Well, if you are a coke head with a coke head morality, you decide well that's not the right thing to do at all. You should actually just take the money and we should fire these people and replace them with cheaper employees. So levering up a company in order to buy out a bigger company and then firing everyone and shutting down the pension plan and distributing the surplus to shareholders, it is just very standard cokehead behavior. Whereas if you look at the mortgage backed securities boom and structured products generally, in the mid-2000s, the way that people made money in that was just by being very, very detail oriented and being able to make these incredibly fine-grained distinctions between different products that were basically similar but one of them pays 5.7% and one of them pays 5.75% and if you lever up that difference enough times you're actually making really good money consistently. It's super boring but maybe with enough Adderall it's actually very tolerable work that you can enjoy. 

So I do think that within stimulants the difference between short acting stimulants and long acting stimulants does mean the difference between a hostile takeover boom and a structured products boom. 

The drug in FTX’s case is called Emsam, which is a Parkinson's treatment and there's some evidence from pretty small sample size studies that one of the side effects of this drug is compulsive gambling. There have been very, very fun tweets about this claim and there have been these official denials from the company doctor on the other hand. 

If you're a company that has a company doctor, maybe that says something about the level of medication you're consuming and maybe the company doctor's job is partly to say, “As a doctor I can assure you, I would never give someone three times the normal dose of Adderall just because their boss hired me to do that specifically.” Dealers don't exactly have patient confidentiality norms, doctors do. So maybe you hire a doctor instead of a dealer specifically to get that plausible deniability.

Dwarkesh Patel 0:17:12

Other than drugs I also want to ask you about the phenotype of the founder. FTX had a barber who would come in every Tuesday to cut everybody's hair. SBF could have just sat in line and gotten his haircut. The way he dressed was completely unnecessary and it was very purposeful. If that archetype of a founder who's in a t-shirt and shorts has been priced in and is beta instead of alpha now, what is the new phenotype and physiognomy of the founder? Where are you looking for alpha?

Byrne Hobart 0:17:58

I would draw the distinction between the physical type of someone versus their presentation and their dress. I'm sure someone could run some interesting numbers on that but I don't have a good sense of what exactly they'd get from that but in terms of how people publicly present themselves, my guess is that yeah, there will be this swing towards investing in people who look a little bit more formal, a little bit more boring. These things are somewhat cyclical. 

Part of the norm on investing in treating the suit as a negative signal is that a lot of investors have this view that when the MBAs come into an industry a lot of the alpha is gone. There's a decent market timing signal that if a lot of people from Harvard Business School go straight into some field that field is probably peaking. There's a little bit to that where the suit is some example of conformity. On the other hand wearing a suit in Silicon Valley is an example of non-conformity and outside of New York, within the US, most of the time wearing a suit as a tech company founder would be this weird sign that you don't know what you're doing, you don't know what the right signals are or you're about to testify to Congress and that's why you have a suit now. Not generally a great sign but maybe it is a sign that you are willing to do some more conformist things and that you could pay attention to details even when the details are boring and also that you are making some kind of financial investment in that particular appearance. 

I would guess that there will be a tilt away from the hyper informal founders. But I also think that if you treat that hyper informality as either this attempt to game the system and just say, “I'm going to try to remind people of Mark Zuckerberg circa 2005 as much as possible so I can raise money and pretend to be the next big thing” That is one thing people are signaling. And then the other thing is they're just accidentally signaling total indifference to anything except the thing they're working on. Maybe that's a good thing in unregulated domains and a really, really bad thing in regulated domains. If you're investing in a medical devices company, you probably don't want a founder who just cannot focus on anything except the product because there are rules and norms they have to follow and it gets bad if all they're focused on is this one element. 

If the hyper focus is just perfectly calibrated that's good. So maybe adjusting your appearances this way to say that you have correctly calibrated your hyperfocus and you're going to get one thing right and it's going to be really really right and you've identified what things matter and what things don't is good.

Dwarkesh Patel 0:21:02

You'll also keep track of your bank accounts. That's the dress itself but I also want to ask about the other characteristics. You had this really interesting point in that blog post about how when you try to scout for talent when the talent is young, you're over indexing for parental involvement. I'm curious, if you had to identify somebody who had to be under the age of 18 or under the age of 20, what is the metric you're looking at that least indexes for parental involvement where they're being forced or encouraged by their parents to do it?

Byrne Hobart 0:21:35

I think the closest you could get is something that is either totally illegible to the parent's understanding of status or something that is actively low status. It’s hard to enumerate those and not just get swamped in — should this thing be low status? high status? Is it actually terrible to say that you ever want to hire someone who was really good at x for some value of x? 

The origin of that point was that I was arguing that when you look at people who are at some percentile and they're in their 20s or 30s, a lot of it has to be that they have some combination of talent and have tried really hard. There's probably been some element of luck but over time the luck starts to wash out hopefully. But the younger you go, and this is probably just my experience of having kids, if you talk to your kids every day about multiplication they will start doing multiplication at a pretty early age and it's not that they are really really smart and they got to multiplication a couple years early. It's that you push them in that direction and they are able to do it early. So the earlier you go the more you are over-indexing on what the parents did, what they emphasized and also what they told the kids was just part of the script. 

I remember anecdotes about people who grew up in lower middle class or below circumstances but would have one distant relative who owned a business and that made them aware that they could own a business and this is a thing they could do. It's part of the script now. That wasn't the only reason that they would have started business but it could be a reason that they decided to do that when they did. And you have to imagine that for everyone who had one uncle who owned a scrap dealer or something maybe there are five or ten or fifty people who grew up in similar circumstances with a similar level of innate ability and just didn't have anyone in their social circle who demonstrated to them that this was something you could actually do. 

Getting back to the talent identification problem, part of my thesis there was that it's really hard. And it's getting harder now that you have Y Combinator going after the relatively young talent versus what the medium VC was going after when YC started. Stuff like Pioneer and Emergent Ventures is going even younger. The younger you get, the more it is this luck driven thing that is about what they got exposed to, with the exception of prodigies. I'd like to think that if I encountered an eight-year-old Mozart I would be able to identify this person as just an extraordinary talent where even if their parents were making them practice ten hours a day, they couldn't be that good without talent. And maybe something similar with the Polgar Sisters where if I encountered a six-year-old who can routinely beat me at chess and so I google some chess books and then go back and try to beat them again, and they're actually better and they're laughing at me and at some point you decide that this is actually natural talent.

For a lot of other domains, there's just so much room for parents to push one thing and do some combination of their kids talent and their own emphasis to get their kids really good at it and that's very hard to adjust for. Especially because if you ask the parents they're going to underestimate how much they overemphasize things because to them this is just a normal thing that everyone should be interested in. So you won't get a good signal from asking parents and then you won't get a good signal from asking other people because they don't know how this family spends time at home. If the median family has more YouTube and Netflix time and less math practice time, that family's just going to assume their behavior is normal.

Dwarkesh Patel 0:25:25

It's a bit confusing because you also want to potentially include parental involvement in your estimate of how good this person will end up being. If you think for example that giving somebody a shot to get started programming early is actually a big factor in putting them on that sort of loop where they get better by practicing and they enjoy it more, you might expect momentum more than mean reversion in that kind of early start.

Byrne Hobart 0:25:54

Sure. So I think part of what this gets to is the question of — What are you optimizing for when you're doing a talent search? Maybe one reason there could be some alpha left in talent search among people who are super young is that a lot of the academic institutions that are doing some form of talent search are optimizing only for how this person does over the next year. So if someone is a math prodigy and they get to join the math team at that school, the school is not trying to optimize for: will this person be proving novel theorems when they're 25? It's really, will this seven-year-old be doing algebra by the time they're eight? And that is still very tied to parental involvement. 

Parents, like kids, like structure and if you tell them this is the appropriate next thing to do with your kid then they're more likely to do it. So you can coast on that momentum for a while. But I think the trap you can run into is that you identify people who are like 95th percentile talent with 99th percentile just super aggressive parents and that combination gets them to 99th percentile performance until they leave home and then they never do whatever that thing is ever again because they didn't really like it, it was just something their parents pressured them. 

Maybe the ideal would be when you get 99 percentile on both. So the parents are putting them on this trajectory but the parents are actually aiming a very powerful rocket ship and it's going to go right in the right direction, which is ideal. I think there is some level of imprinting that young kids have where a lot of kids learn about programming when they're very young and that's something that they do from a very very early age and then it becomes the thing that they work on for their entire career. 

Obviously that has to be fairly new because it's not like anyone who was born before 1970 just had this constant yearning to program computers and could never satisfy it. Those kids found something else to do. Maybe a generation before, it was repairing transistor radios like mine did when he was a kid and maybe a century before that it was experimenting by building little internal combustion engines and seeing whether or not they explode like Henry Ford did with his friends at school. The earlier you get, the harder it gets to really map these activities to anything concrete that we understand and can relate to but there's probably some extent to which you can direct kids into whatever the modern instantiation of this long-term enduring tendency is. 

One interesting example of that is — I've been reading the Robert Caro LBJ biography and there's this bit towards the end of the first volume where LBJ is put in charge of this fundraising organization for Democrats in Congress and when you read about it he sounds like a traitor. He sounds like someone who was just born to be slinging currency derivatives because he is constantly on the phone, constantly picking up rumors, constantly sending money here and there and everywhere else. He's always sending money overnight and then sending someone a telegram the day before saying “You're going to get a package from Lyndon Baines Johnson and you're welcome.” He's doing this thing where he's constantly relentlessly optimizing every little tiny detail of some very complicated process which clearly requires enormous working memory and requires a very strong poker face. He has to be able to differentiate between someone who is begging for money because they're pulling at 49% and with a little bit more money for newspaper ads they get to 50.1% versus someone who just wants the money or is just constantly freaking out by their nature. So it requires a lot of the same character traits but the 1930s were just not a great time to go to Wall Street. Maybe if LBJ had been born at a slightly different time that's just what he would have done and he would have been a very successful private equity executive or something but sometimes these general skills can translate into a lot of different areas and they get honed into very specific skills through deliberate practice in those areas if you have that combination of natural tendency and some level of motivation, which in LBJ's case, his dad was also a politician so he had this example of — this is part of the life script, you can do it. But his dad was broke after a while and so he also had this example of what not to do and ended up making good money for himself, in addition to his political career.

Dwarkesh Patel 0:30:35

I'm glad you brought up the biography, I'm reading it right now as well. In the last episode, we go deep into the other biography by Robert Caro, The Power Broker, and talk about why it might be inaccurate in certain respects. But what it is accurate in and what Caro has a genius in is talking about the personalities of these great men, the people who have really shaped their cities or their countries for decades and centuries. There's many places where if you understand the economics of an issue he's talking about there's a lot to be left to Caro’s explanation but the actual breakdown of the personalities is just so fascinating and worth reading Caro for. 

Being a politician is about building a network, building know-how, building this sort of inarticulable knowledge from an early age. It might be the case that in those situations just having connections and having parental involvement gets you far but if it's like becoming a programmer sure, you'll have done data structures by the time you're 16 but eventually you'll get to the point where everybody knows the basics and now you actually how to do interesting and cool things in computer science and now your 95th percentile of spatial reasoning IQ is not going to get you that far. 

Let me ask you about the Caro biography because you had a really interesting comment in your review of the book that I've been wanting to talk to you about. You said — “It's worth speculating on how many LBJ-level figures exist today, perhaps in domains outside of politics, and how many Caro-level biographers there are who could do them justice.” 

Do you have some idea of who these figures are? Or if not that, at least what areas you'd expect them to be in?

Byrne Hobart 0:32:34

I think a lot of people who are close to that tier and have some of the same personality types are in sales and corporate development and stuff like that where they're building a big network, they are constantly building out this giant levered balance sheet of favors. Favors owed to them, favors they owe to other people, and like all forms of leverage it does allow you to grow a lot faster but you occasionally get these big big blowups. So that's one place I would look. 

I think if you try to look at the more pure executive founder types then it gets harder to find someone who would have exactly that kind of personality. Part of what made LBJ’s methods work was that he was adjacent to a bunch of these really big institutions and he could siphon off some of the power that these institutions had and in some cases, make them more powerful. I'm about a third of the way through Master of the Senate right now so it's just getting to the point where he's making the senate more effective than it used to be and also making it an organization where it's less seniority based. 

You need to be attached to something much bigger than yourself for that particular skill set to work really well. That said, you could have a really big impact because it's another form of leverage. At that point the senate was 96 senators and if you're able to exert a lot more influence and be the equivalent to 40 senators for example, then you can get a whole lot done because it's the US senate. But if you have that same kind of skill set and you're the CEO of your company, well you're already incharge of the company. There's only so much extra force you can exert. 

So you see a figure with exactly that kind of personality trait in a case where there are big institutions that have slowed down somewhat. Another interesting point that is raised early in Master of the Senate is that the senate was getting old and if you look at these long-term charts of average age of politicians, we're definitely in a bull market for extremely, extremely old politicians in the U.S right now but we've gone through cycles before. One of the things that tends to cause a reset is war. Wars, among other things, cause a huge reset in social capital. The people who made mistakes in the early stages all get discredited and then the social bonds that people forge from actually fighting alongside one another and the prestige you get from being part of the winning side. That is very hard to replicate so you end up with much younger people in positions of a lot more power. 

In the 1930s, there just weren't a lot of organizations that were hiring heavily and looking for really ambitious young people who are going to shake things up but the U.S government was. That's how LBJ got in and started on his path. The new deal created these big programs like the national youth administration and they needed people like Johnson to run them. 

When you look at an industry that is aging it's usually an industry that ambitious people stay away from. They recognize it's becoming more seniority focused and there's just less going on but then it becomes this huge opportunity when the aging stops because a bunch of people either retire or they get discredited and have to leave and suddenly the average age of the industry ratchets down and you can basically look at the set of opportunities that were missed over the previous decade because the industry was too risk averse and you get to take all of those opportunities at once so you have tons and tons of low-hanging fruit when that shift happens. 

So I think that's the other thing to look for. Look for cases where there's some institution or some part of the economy or society that has just been slowing down for a long time and is clearly getting to the limit of whatever its current operating model is and hasn't found a new model and there's someone young and disruptive who's just entering it. 

Maybe the place to look for the next LBJ is someone doing independent films and someone who looks at the top box office results and sees that everything is a spin-off of a spin-off of a spin-off and it's 50% marvel and says “This is disgusting. We have to destroy it and I'm going to build something completely different” Maybe that person is actually the kind of LBJ archetype. 

Now the other half of this question is the Caro archetype and part of what I found fun about this was that I felt like Caro was disgusted with himself when he realized how similar some of his methods were to LBJ’s. He's writing this story about this guy who's will do anything to make a friendship but it's really a fake friendship just to accomplish his goals and he's constantly doing doing the reading that other people aren't doing, and doing the work, and making the calls, and reiterating and reiterating and reiterating, just endless patience. 

And then you read about how Caro works and he does things like — moves to DC for a while, talks to everyone in DC and befriends people. Goes to texas, moves to the hill country and gets to know people. The book sort of has these hints of gonzo journalism where sometimes Caro will just go from — here's what happened in 1946 to here's what happened to me in the 70s while I was talking to this guy about what he did in 1946. And sometimes he will basically come out and say, “I waited until the person who paid this bribe had Alzheimer's and then I asked him if he remembered paying the bribe and he remembered that he did it and didn't remember he wasn't supposed to say it. So that's how I know.” There's this line that Caro keeps quoting from LBJ which I think was from LBJ's debate team coach days where his line was — “If you do everything you will win.” and Caro does everything. 

I think the population of Caros is smaller than the population of LBJ's because the people who have that skill set probably have ambitions other than writing two canonical works on two important people. Maybe a lot of those people are just doing things other than typing.

Dwarkesh Patel 0:39:03

Man, there's so many threads there that I'm tempted to spend the rest of the episode just digesting and talking about. There's so many interesting things about Caro's story and the impact he’s had. 

In circles like effective altruism there's been this focus in terms of thinking about impact, of trying to crunch the numbers, and there's no reasonable crunching into the numbers you could have come up with before The Power Broker was written to justify it. By the way, he tries to downplay his accomplishments as a journalist before he wrote The Power Broker but he was nominated for the Pulitzer Prize for his journalism before he wrote the book. So he's a top level investigative journalist and then says — “Here's how I'm going to spend my talents. I'm going to spend eight years looking into and researching every conceivable person who has even potentially been in the same room as or been impacted by Robert Moses and I'm going to document all this. I'm going to write a book of like a million words or something.” 

He probably didn't think about it this way, right? But look at the result. That book probably changed how many of the most influential people who came up through politics think about politics. It probably changed how urban governance is done, how we think about accountability and transparency, for good or ill, depending on your perspective. 

That example alone really makes me suspect of the number crunching way of thinking about what to do and rather just be like, from Caro’s perspective, “I gotta understand how this guy accumulated this power.” He does it and it completely transforms how urban governance is done.

Byrne Hobart 0:40:41

Yeah. Looping back to the parental influence thing, I think part of what happened was that the more Caro dug into it, the more he realized this is actually a big and compelling project. 

There's this fun phenomenon that you can get when you're researching something where you've read enough that when you read something new and you see that there's a footnote, you actually know what is going to be cited in that footnote and maybe you've also read the thing about how the thing in that footnote is wrong and here's why and you're picking up information a lot faster. You get that nice convexity where you can skim through the stuff you know and everything you read is new information and challenges something about what you previously knew and that's just a really intoxicating feeling. I can imagine that it's even more fun if you're actually digging up the primary sources. 

So if you're Caro you've gone through The New York Times archives, you've read through all of the external coverage of what people said about Moses’ time, and then you start talking to people and you realize there are things that we got completely wrong. We thought Moses didn't want X to happen and it turns out that he kept scheming and plotting to make X happen and just wanted to pretend that it wasn't his doing. 

But what happens is you build this ongoing motivation and then you can make something that you just wouldn't be able to make before. And if you start out saying “I'm going to write a million words about how cities are run.” you will probably fail but if you keep writing another 500 words a day about how Robert Moses operated and what he did, and then you have some reflections throughout that on what that means for cities then maybe you actually get there. 

Maybe some of this is like you want to have an adversary. Caro’s books do partly seem to be this cross-examination of who he's writing about and often he seems to have very mixed feelings. One of the really interesting things in The Years of Lyndon Johnson is Caro's description of Coke Stevenson and how he contrasts him with LBJ because it's really clear that Caro's politics are completely opposed to Stevenson's. When Caro's writing about LBJ, there's the good stuff he did which is the great society and his participation in New Zealand, and there's the bad stuff, which is anything that wasn't bad. So he clearly likes what LBJ accomplished and despises the person and then really likes the person of Coke Stevenson and wishes him well but also doesn't actually want people like that to be in charge of anything. 

It’s partly Caro debating with his subject and interrogating his subject and partly debating with himself and asking these very long-standing questions about whether or not the means justify the ends and would it be worth it to not have a great society in exchange for not letting LBJ steal an election in 1948. If he's good at his writing he shouldn't be coming to firm conclusions on that and he should be presenting this very, very mixed picture where you really only get the things you really want if you also accept that there are some very bad things that come along with that. As long as the things you want come from powerful ambitious people who will do anything to win.

Dwarkesh Patel 0:44:14

It's worth remembering that it takes him a decade to write each of those volumes and in the course of a decade just imagine how many times you would change your mind on a given subject. You really notice this when you read different paragraphs of The Power Broker where early on, if you just read the first third or the first half The Power Broker, you're like, clearly Caro is writing about Robert Moses the way he writes about Lyndon Johnson where it's like, “Yeah, this guy had some flaws but look at the awesome stuff he did for New York.” and then the tone completely changes. But you gotta remember that he's writing this over so many years.

I do want to talk about war being a catalyst for young people entering an arena. I did an interview with Alexander Mikaberidze. He wrote a really interesting book about the Napoleonic wars and this is actually one of the things we talked about. There's a line from war and peace where one of the Russian aristocrats is mad that his son is joining the war and he goes, “It's that man Napoleon. You've all seen him and now you all want to go off to war.” 

I’m curious. Filmmaking doesn't seem like a place where super quantitative and super smart and super competent people go. Somebody who has thumos and the desire to dominate and the desire to achieve recognition, do you really think he's making films? Where is he really? Is he still trying to start a startup or is that now a decade too old and now he's trying to dominate some other arena? 

Byrne Hobart 0:46:31

Maybe the lame answer is we don't actually know. Paul Graham has that essay about the trope   of startups starting in garages. I think it's called The Power of the Marginal and it's all about how the really interesting projects are the ones that can barely get off the ground because they're so weird and so out there that there is no infrastructure to support them. And what that ends up doing is selecting for people who are extremely passionate about that project and who are extremely willful and will get impossible things done. 

It's hard to just rattle off a bunch of examples of that because your hit rate would be low. 99 things out of 100 are just things you read one fun blog post speculating about and they're actually never going to happen and then one of them maybe you're right, but it's very hard to tell which one it is. If it were very easy venture capital would not have such skewed returns.

Maybe it is harder to optimize for what area do you look for. Maybe it's actually easier to do the meta optimization of identifying the things you quit podcasting and go work on given the opportunity. It's good to have that sort of dread list. 

If someone at Spotify pinged me and they need a product manager who can help us display classical music such that we don't list tons of redundant information in the first 50 characters of the track name and the actual incremental useful information in the 10 characters you have to wait for it to scroll through unless it doesn't actually scroll through. If someone pinged me and was like, “We really need someone to fix that. Can you come and do this?” I'd be sorely tempted. I feel the same way about Google Finance. If someone emails me and says you have a mandate to make Google Finance good, I'd be tempted. 

Thinking of what industries would have that kind of pull for you and then what can you do to really dig into those industries. You probably find the proto successful people in spaces like that versus trying to optimize in advance for — if I were someone who thinks like nobody else thinks and were a true natural contrarian, and also had spent several years learning about different opportunities, which one would I have ended up picking? 

Because then you're sort of magicking away all of the things that actually make the person you're looking for it looking for. So yeah, it can't quite be done that way. 

Dwarkesh Patel 0:49:05

I want to go back to that thing you said a moment ago about how you couldn't have written a million words that were as impactful about how cities work but if you just wrote 500 words at a time about how Robert Moses accumulated power, did the things he did, you can actually have a really interesting and influential piece of work. 

Is that how you see The Diff? That you can't write one million words at a time about where technology is going, what's happening with the productivity slowdown, what's happening with all these emerging industries. But if you just write 2,000 words a day about what's happening with any particular company or industry then you can compile this really interesting overall worldview about finance and tech.

Byrne Hobart 0:49:44

That's the hope. I might be projecting things about my own attention span onto Caro when I say you can't just set out to do a million words on topic X and then do it. But I hope that I am by increments producing something that is a lot more than the sum of a bunch of business profiles and a bunch of strategy breakdown things and things like that. That's one of the reasons that I spent time on things like reading Machiavelli and thinking about how Machiavelli's thoughts. Not just the totally cynical amoral stuff but the other stuff he wrote at the same time about how to build a sustainable and good republic rather than how to be a completely amoral monarch.I try to read that kind of thing because I do think that it's valuable to have that more rounded view of the human condition and I think that it contributes a lot to to writing about these individual companies. 

Technology changes a lot. Humans change very slowly. So if you want to understand technology you do have to study this specific object level case of — What is this thing? What does it do differently? What is it a substitute for? What are the compliments to it? Etc. But if you're trying to understand things like — Why did this company do X? Why did they fire this person and not that person? Why did they choose to acquire this other business? Why is the CEO dumping tons of money into this thing that seems like it doesn't make much sense? Well, you can find lots of historical examples of people in power making these decisions that just get continuously worse and continuously more costly and they refuse to back down. Sometimes they turn out to be right. Sometimes they turn out to be very very wrong. But you'll find more examples of that if you go back further in history and they're often just a lot more fun to read about. Whereas, you can read about things like Ford spending too much money on the Edsel and it not working out or, IBM investing a ton in the 360 and that working out very nicely. But you can also go back to the Iliad and read another case where sunk cost fallacy dominated strictly rational decision making and only divine intervention could ultimately lead to a good outcome for the attacker and even then maybe not such a great outcome all things considered.

Dwarkesh Patel 0:52:04

That particular question about trying to predict if somebody is overstepping or if they're making the best bet of their life is something that I've been trying to think about and I really have no reasonable method. 

If you think about what Elon Musk is doing with Twitter is this like Napoleon trying to conquer Russia and it's this super ego filled, pride filled, completely illogical bet from somebody who has just had 20 consecutive wins in a row and he thinks he's invincible? Or is it like Elon Musk 20 years ago where he's like, “Yeah. I did PayPal and now let's build some rockets and let's build some electric vehicles.” In each of these cases there's so many analogies to complete bust and there's so many analogies to — “Oh. This is just part one of this grand plan.” 

How do you figure out which one is happening? How do you distinguish the visionary from the collapsing star? 

Byrne Hobart 0:53:18

The cynical answer is you wait about 200 years and then you write about how it was obvious all along. There are a lot of cases that are actually still ambiguous. Alexander conquered most of the world that people knew of around where he grew up and then just goes to Babylon and drinks himself to death and that's the end. 

There could have been an alternate story where he gets his life together a little bit and runs a giant sprawling empire. On the other hand, reading the story battle to battle, a lot of it actually is basically this Ponzi scheme where every time he conquers a city he gets enough loot to pay off the people he hired to help him conquer the city and then has to move to the next city because they want to get paid again. So he sort of was being chased by his obligations the entire way through until he finally got just ahead of them enough to get a lot of loot and a lot of land that he could give people instead of just giving money. 

Even in that story it's very hard to say that he rolled the dice a bunch of times and won every time so clearly he was just one of those people who's born to win. Maybe he actually backed himself into a bunch of corners over and over and over again and then desperately fought his way out every single time and then was just completely sick of it and burnt out by the time he was in his early 30s.

In terms of how you would figure it out in advance, I think some of it comes down to getting a sense of whether they're responding to circumstances or whether they actually have a long-term plan. But then there's probably nothing more dangerous than a long-term plan that someone actually has the means to execute. Five-year plan does not have a good connotation, Stalin had some of those and they didn't turn out well for a lot of people. 

Even within that there's some difficulty in evaluating. There's kind of that meta-cynical layer where if they don't know what they're doing then probably it's dumb luck that they keep succeeding. On the other hand if they do know what they're doing then maybe you hope that the world is lucky enough that they get unlucky and can't actually pull off whatever it is that they're planning to do. 

I guess another thing would be — Is there an end state that they can get to? Because Alexander basically just kept going until he couldn't go any farther, until his troops were basically on the point of mutiny and then just turned around and went to the nicest place halfway home and hung out there and partied. 

If the story is less about conquest and more about reconquest and restoration of something then there are these natural limits. You can say, “You go this far and you don't go any farther.” because you've actually finished your task. For example, the Generals who chased Napoleon out of Russia. For them the master plan was not — “We're going to conquer all of Europe. The master plan was — “We're getting our country back and then we're going to chase him far enough that he doesn't feel like he can just wait a year and do this again when it's not winter.” 

So maybe that's another way to constrain it but then you end up naturally selecting for less ambitious people. One way to have these guardrails on your behavior is just don't have very big ambitions. In that case those people are also stuck responding to circumstances. 

Maybe you just end up with many different iterations of the same thing on different scales where everyone is stuck in certain historical circumstances they have. They have their skills, they have their opportunities, they can they can go after some things, maybe they achieve great things maybe they fail but either way eventually their luck runs out or they run out of ideas and then there's nothing to do except go home or just keep trying to keep being bolder until you eventually fail. 

On Musk particularly, I don't really understand it. I think there's a remote possibility that he actually has a bunch of specific concrete ideas for how to increase Twitter's free cash flow and how to pay down the debt and make it a more profitable company maybe he just had that sense that it was overstaffed and that it should survive with a smaller headcount and if you cut headcount enough then you you end up with with a profitable business. it could also just have been fun. Seems fun so far. 

The pursuit of fun is not to be discounted. If you're super rich you can afford to do all sorts of things to varying levels of entertainment. It may be that the only thing that is actually truly novel and a thrill seeking fun opportunity is to do something like buy Twitter and then turn it into what it is. 

I think [unclear] had this point about how the nature of Twitter's legitimacy has changed and that now it is under the rule of a single monarch instead of ruled by these faceless bureaucracies. Now if Twitter does something you don't like, there's actually a specific person you can blame and because you have Twitter you can actually yell at that person and potentially get an answer. Whereas if Twitter bans you because you made a joke and the joke looked like it was serious, there's no recourse. There's nothing lower status than arguing with someone in authority about how seriously they should take your jokes. And it works both ways. 

I started noticing this years ago because there are these _txt twitter accounts where they're just posting out of context comments from some niche community and the comments always sound deranged. In a lot of cases to me the comments read as someone who is doing a bit. They're playing a role, they know it's funny, they're exaggerating for their friends and then you take it out of context and read it as totally serious and then you get to say these people are all like this, they're all crazy. It is a marker of high status to be able to not get jokes and be able to be righteously angry at someone because they made a joke. If they were serious that would have been an appalling thing to say but they obviously weren't. If you can get away with saying “No, I actually don't think it was a joke at all. These people are humorless and they must have been totally serious.” then that's cool. That's high status and makes you impressive. 

Musk’s rule as this personal monarch speaks to this question of legitimacy — why do people trust moderation and why do they trust sites to operate in the way that they do? You can either say these are really high quality institutions. You can take the [unclear] approach and say we built these systems such that anyone can be dropped in and can do a reasonably good job. It's very hard for bad people to do a very bad job because there are so many checks and balances. Or you could say no, we actually trust this one person to do a really exceptional job that nobody else could do and we don't want institutional constraints on them. 

Those philosophies go in and out of fashion even within systems that nominally don't change. The US was a lot closer to that kind of centralized system with personal legitimacy invested in one person under FDR than it was under Calvin Coolidge. Under Coolidge it was a lot more of — There's this institution. There are a bunch of rules. People follow the rules. You have this nice New England guy who gives an annual update of the State of the Union but it's just written down and then he has a clerk read it to Congress. You're not betting on charisma. You're not betting on judgment. You're just betting that the rules are pretty good and as long as things keep working according to the rules they'll keep on working. 

Dwarkesh Patel 1:01:28

The Musk example is similar to how some people will load up a horribly broken game of Civ where their civilization is losing. They’ve gotten so good at the game that they just need some noob to send them their save file which is complete carnage and they're losing their cities and then the fun is in loading it up and trying to win anyways. 

One thing you've written about that I find really interesting is — we're both fans of Fukuyama's book The End of History and the last quarter of that book completely contradicts the first three quarters of the book where he's just saying that these men at the end of history are pathetic last men who have no desire for recognition. They just want to be comfortable. You've made the comparison with that and big tech, at least before the crash. 

One of the things Fukuyama talks about in the book is — Once there is a great war, once there is a struggle that requires the first men of history who can withstand adversity and can accomplish great things, you won't have them around by the time when things have gotten comfortable for a while.

Are there enough first men left in companies like Twitter and Facebook now that they do face adversity? That they can just reboot and go into wartime again?

Byrne Hobart 1:02:56

Yeah, I suspect there are. I think Tolkien gets it right that even if someone is born a Hobbit and they live in Hobbiton and have a nice comfortable life, they still have that capacity and yearning for adventure. And that in the right circumstances they will rise to the occasion and go ahead and do it. 

This seems to happen with a lot of countries when they face these great stresses. Sometimes a civilization just can't withstand it and it collapses and the sea people just take everything and then you have no civilization left and you're all just back to subsistence farming. But in a lot of other cases, even if they ultimately don't survive, they go through a very long decline because they do fight to maintain what they have for an extended period. 

It's hard to think of a mechanism by which you can eradicate that thirst for glory and that ability to rise to the occasion, unless it's like microplastics or something. Maybe that does constitute the end of history in which case hopefully we exported up enough microplastics to make sure that we don't have any last pockets of thumos. Sort of like the Scott Alexander riff about the steppe nomad invasion risk where it's an existential risk that comes along every couple hundred years. Yeah, you want to avoid that. 

I think part of having that kind of thumos and thirst for glory should be that you can't be so habituated to a life of ease and comfort and lack of difficulty that you just won't actually respond appropriately when there's an external threat that you need to respond to. Maybe you weren't first man material after all if you can't and you just want to stay on your couch. 

I'm sure we can sort of deplete that reserve. Post World War II U.S. was definitely a country where there were a lot more people who had taken very serious risks they've gone through a lot of hardship. 

I recently read a book The Economics of World War II which was comparing a bunch of countries and how their economies performed in World War II and one of the things that stood out about the U.S was that in a lot of terms of material consumption, the U.S didn't really look like a country going through a war. 

In most other countries you saw this decline in literally how much food people had to eat. Especially how much protein and fat they had to eat. Calorie intake had dropped by like a third in a lot of places by the end of the war but in the U.S calorie consumption actually went up. So on the home front, the U.S was inconvenienced by the war and things like gas and tires were hard to get, but people were still eating well whereas in a lot of other parts of the world people were literally going hungry so that their country could continue to fight the war.

Maybe there's some level of hermetic response where you you suffer a bit because your country is contributing to this and then you're heartier for it and the country has accumulated a lot of social capital and you had to get really good at organizing and building things and then maybe there is some level of suffering from conflict where you've totally had enough and you're never doing anything like that ever again and you're just too done.

I think one of the interesting things to consider is the extent to which different countries fit into that model. I'm very interested in Japan and Japanese industrial policy and how the Japanese post-war recovery went and one of the annoying things about that is that I thought the question was — How did Japan have this wonderful post-war recovery? But when you look at a lot of the institutions involved, they don't start in 1945 or 1951 or whatever. They actually started before World War II. So you can actually sort of see World War II as part of this arc of the same historical process that continued post-war which is Japan wanted to be economically self-sufficient and independent and a country that could determine its own fate and during the last gasp of imperialism, one way to do that was invade countries with lots of natural resources, take those resources and then manufacture things at home. 

But when that became untenable then the next best option was be within the sphere of influence of the most powerful military in the world and be very closely tied to their import and export markets and then import everything you need under the protection of the US military and then export things to the US in order to pay for those imports. Basically run the same strategy just with someone else during the military part. In one sense that was a total defeat of the imperialist model and in another sense it was this strategic realignment but actually basically the same end goal. Very different external facing view of that goal but same ultimate idea.

There's a book called Princess of the Yen which is mostly about Japanese central banking policy but it has some early bits about how the structure of Japan's economy works. The way the author describes it is that post-war Japan had a war economy in peacetime with lots of centralized control, suppressed consumption, and lots of heavy industry, heavy manufacturing. The modern structure of a lot of companies in Japan dates back to the wartime period, sometimes the post-war period. This includes the biggest advertising agency in Japan that was apparently a wartime, or immediately pre-war, attempt to agglomerate all the smaller ad companies into one big more efficient company that would free up resources that could be used for building battleships and other stuff like that. 

So yeah, it's kind of the same story just being expressed in a bunch of different ways. I think you can look at other countries and try to see what threads of continuity there are between the pre-war and post-war order. Tony Judt's postwar book is a really phenomenal look at that question. In a lot of cases there's like there's a surprising level of continuity. There are some things that totally broke and had to be totally reformed and there are some things that just kept going exactly the way they've been going before.

Dwarkesh Patel 1:09:33

Everybody wants to be a first man but nobody wants to go on a diet. [Laughter] 

You've mentioned this line before but there's a line from How Asia Works where they're talking about the reparations that Korea got after World War Two from Japan and how they're using that to build up their industrial capacity and there's a line from one of the line managers in the factory who goes, “Listen you guys have to work 14 hours a day, seven days a week and the reason is this money is blood money. It's our blood. This money was gotten from ripping your mother and killing your father and if you can't use that money to rebuild our country, what good are you? You might as well just kill yourself.”

I was reading about lean production before I interviewed Austin Vernon and in all these books, they're talking about how America was never able to replicate the productivity of Japanese lean production and it's just because you're talking about Americans who are trying to save up their pensions and working eight hours a day and have hour-long lunches and then you have these hardcore Japanese who just got through World War Two and barely survived. The thumos is completely different. You just can't replicate that in America.

Byrne Hobart 1:10:55

There's this book called the reckoning about the US and how it dealt with the rise of Japanese exports in the 70s and 80s. When it's talking about the post-war recovery in Japan there's this bit where the character they're following works for a bank. And in the bank's office in some city in Japan, there’s a room with a fire and a big pot of stew and the stew is the food that the employees will eat at the bank as part of their benefits package. He likes to work in the room where there's a fire because that's the only room in the bank that is warm enough that you can actually work. 

That is a level of material austerity that is inconceivable to me. I don't know anyone who grew up in such poor circumstances. I can't imagine it. And then it did not take very long at all for the country to significantly recover from that. And there are countries that also had this massive catch-up growth where they went from very poor subsistence level or even below, in some cases, to actually being middle income or even fairly rich countries.

Dwarkesh Patel 1:12:10

I want to touch on the fact that Sam Bankman Fried was an effective altruist and that he was a strong proponent of risk neutrality. We were talking many months ago and he made this really interesting comment that in many belief systems they have a way of segregating and limiting the impact of the most hardcore believers. If you're a Christian, you can just make the people who take it the most seriously monks so they don't cause that much damage to the rest of the world. EAs don't have that. So if you're a hardcore risk-neutral utilitarian you're out in the world making billion dollar crypto companies. 

As a side note it's interesting that a year ago the meme was — “Oh look at these useless rationalists. They're just reading blogs all day and they have all these mind palaces and whatever. What good are they?” And now everybody's like, “Oh these neutral utilitarians are gonna wager our entire civilization in 51-49 schemes.” [Laughter] I just want to get your commentary on all this.

Byrne Hobart 1:13:17

Yeah, I think it's a useful pattern to observe because it goes back to the point that human nature just doesn't change all that fast even to the extent that it ever does. Different civilizations have had this problem of — we've got some rules and we've got these beliefs and they're they're generally going to guide people to behave the right way but they're going to guide people to be the right kind of normal person and not to be someone whose life is entirely defined by this incredibly strict rigid moral code and by whatever you get if you take the premises of that and just extrapolate them linearly as far as they can go.

I think that gets especially dangerous with really smart people because you can give them a set of first principles and they can ask really really interesting questions and come up with edge cases. Some people encounter these edge cases in their first philosophy class where they just reject it as stupid and say things like “If you shove the fat man in front of the trolley why do you think the trolley would stop? The trolley would just kill him too. This is dumb.” 

I think it is useful to keep in mind that the thought experiments are designed to be implausible and they are supposed to be the intuition pumps but the more you get this complicated highly abstract economy where an increasing share of it is software interacting software well, software doesn't have that common sense break on behavior and if you have this very composable economy you can find cases where first principles thinking actually is action guiding and can can guide you to extreme behaviors. 

Unfortunately those extreme behaviors are things like trading cryptocurrencies with lots and lots  of leverage and it's maybe merciful that the atoms to bits interface has not been fully completed while we still have time to deal with malevolent unfriendly EA. [Laughter] 

You see this problem a lot and societies tend to have some kind of safety valve. If you really think that praying all day is the thing you should do, you should go do it somewhere else and you shouldn't really be part of what we're doing. I think that's healthy and in some cases it's a temporary thing. You do that and get it out of your system and you either come back as this totally cynical person who doesn't believe in any of it or you come back as someone who is still deeply religious and is willing to integrate with society in a productive way.

Even within the monastic system you have different levels of engagement and interaction with the outside world. I think that's something that EA should take seriously as an observation, as a design pattern for societies, that you typically don't want the people in charge to be the most fanatical people and because EA beliefs tend to correlate with being a very effective shape rotator or a very symbol effective manipulator and those skills are very lucrative and money does have some exchange with power, you basically have a system where very smart people can become very powerful and if very smart people can also become very crazy then you tend to increase the correlation between power and craziness. 

It doesn't take very long clicking through Wikipedia articles on various leaders in world history to say that you ideally do not want your powerful people to be all that crazy or your crazy people to be all that powerful.

As far as what to actually do about that, one model is that smart people should be advisors but not in an executive capacity. They shouldn't be executives. You don't want the smartest person in the organization also being the person who makes the final decisions for various reasons. But you do want them around. You want the person making final decisions to be reasonably smart, smart enough they understand what the smart person is telling them and why that might be wrong or what the flaws might be. That might be one model where the EAs are dispersed throughout different organizations of the world as someone working with non-EAs and nudging them in an EA friendly direction by giving them helpful advice but not actually being the executive. 

One possibility is that every other society got it wrong and that the monastic tradition was stupid and it has been independently discovered by numerous stupid civilizations that have all been around for much longer than effective altruism. You can't discount it but I think if you run the probabilities that’s probably not the case.

Dwarkesh Patel 1:17:54

In general, the leaders who take ideas seriously don't necessarily have a great track record. Stalin apparently had a library of 20,000 books. If you listen to Putin's speech on Ukraine it's laden with all kinds of historical references. Obviously there's many ways you can disagree with that but he’s like a man of ideas. And it’s not clear that you want a man of ideas in charge of important institutions.

Byrne Hobart 1:18:20

Well, a lot of the founding fathers were wordsmiths and we basically had a whole collection of anons flaming each other through pamphlets. So in one sense, it was a nation of nerds. On the other hand, as far as I know, Washington did not have huge contributions to that literary corpus. So maybe that is actually the model — You want the nerds. You want them to debate things. You want the debates to either reach interesting conclusions or at least tell you where the fault lines are, what are the things nobody can actually come to a good agreement on? And then you want someone who is not quite that smart and not really into playing wars to actually make the final call.

Dwarkesh Patel 1:19:07

Yeah, that's a really good point. Forget about Jefferson, think if Thomas Paine was made president of the United States, that would be very bad news. 

Byrne Hobart 1:19:17

Yeah. It's important to note that it's better to have some level of fanaticism than no fanaticism. There's an optimal amount and there's an optimal place for it. But yeah, from a totally cynical perspective maybe your most demonic people are at the front lines doing things and taking risks but also not making the decisions about who goes to the front lines. Or making sure that the person deciding where the front lines is saying, “The front line is we keep France safe from the invaders.” and not “The front line is Moscow. Get to Moscow and burn it down.” 

There's a recent Napoleon biography that I'm also in the middle of. It's been a good year for  reading about power tripping people. It points out that technically Napoleon had more countries declare war on him than he declared war on, so on average France was fighting defensive wars during the Napoleonic era. It's just that they kept defending farther and farther from France.

Dwarkesh Patel 1:20:24

[Laughter] Defense often requires some strange kinds of offense. 

One meta question I've had is, in all other kinds of discourse there's this question about whether you're trying to figure out which charities do the most good, whether you’re trying to figure out which policies are best, whether you're trying to figure out how you should promote leaders, anything. 

There's two kinds of discourse. One that's like, “We've got these few dozen RCTs. Let's see how we can extrapolate the data from these in the least theory-laden way.” And there's another where it's like, “I've just read a shit ton of classics and I'm a thinking person. I think a lot about culture and philosophy and here's my big intricate worldview about how these things are going to shape out.” 

Investing is an interesting realm because there's both kinds of people and you can see the track records over long periods of time. Having seen this track record, is there any indication to you where this sort of microeconomic approach actually leads to better concrete results than somebody like Thiel or Soros who are motivated by an intricate worldview that's based on philosophy or something? Which one actually makes better concrete predictions that are actionable.

Byrne Hobart 1:21:41

Typically the greats have some synthesis of the two and it probably leans more towards a big worldview than towards micro-level observations. One way to divide things is to say that the quants are all these micro level observations. You could be a quant who does not actually know what the numbers mean, doesn't know what the product is, doesn't know the time scale, and is just looking for patterns and finds them. People have done it that way but it seems like quantitative strategies get more successful when you find some anomaly and then you find an explanation for the anomaly. The explanation might be some psychological factor you've identified and maybe you find studies indicating that loss aversion is real and this affects how fast stocks go down versus how fast they should often go down and that gives you a trading strategy. Maybe it's something more more mundane like maybe there is some large investor who has some policy of rebalancing between stocks and bonds on the first day of every quarter, and if you know that the investors who have that policy control X trillion dollars assets and you know how they'll rebalance, then every at the end of every quarter you know money is sloshing between stocks and bonds and that's predictable. 

A lot of the quantitative strategies that have those theories behind them tend to blow up more rarely because they sort of know why the strategy works and then they know why it'll stop working. Data mining is always a risk. If you find an anomaly and there's no explanation to keep repeating itself, one of the explanations is that other people found the anomaly too and they are exacerbating it by trading it. 

On the other end, if you have these totally abstract theory driven views usually what kills them is time. A lot of the best abstract theories are where you look at some part of the economy and say, “This is obviously unsustainable.” The problem is that you can say that at any point during its arc and it can look sustainable to other people for a very long time. 

One of the favorite examples of this is this snappy one-liner that is looking at housing and saying a subprime borrower who didn't put down a down payment is basically just a renter with upside. It's some line about the economics of it and that these people are not actually safe borrowers. But the paper that it came from came out in 2001. So if you had read that and been like “The American housing market is broken. People are massively overpaying for houses. They're all over leveraged.” You could have shorted housing stocks and then lost 400% of your money as home prices soared over the next four or five years. 

One of the ways you get around that is you have the high-level theory you say, “Okay, here's what's actually going on in the world. Here's what people don't understand.” but you also have to have this lower level theory of — “Here's what they think is going on. Here's why things keep ratifying the theory that they have.” and then the next step is “Okay, what actually breaks down that causes perception to collide with reality?” And then the other question is, can perception actually undershoot in the same direction?

A lot of money was made by people who looked at the tech bubble in the late 90s said “This is gonna blow up and we're gonna figure out when to short it and then we'll short it. We know it'll just keep going down for a while.” It wasn't that the dot-coms were 20% too expensive, it's that most of them are worth zero. Those people made a lot of money by shorting after things started declining knowing they would keep on declining. But a lot more money was made by people in 2003-2004 saying, this didn't actually discredit the internet. It's still a good technology and it's not like we fundamentally can't make money online, it's that you can't make money online if you don't know what you're doing and you massively overspend and there just aren't that many people online and no one's spending money online yet. So a lot more money was made by people who were able to take advantage of the overshooting in the opposite direction rather than the people who figured out from first principles that the bubble had bubble characteristics and was eventually gonna pop. 

That just requires a lot more of this micro level analysis. A lot of macro people are looking at what individual companies are doing and what they're saying and how consumer sentiment is changing month to month and all these other very low level indicators where the indicators are not a thesis but the indicators tell you something about when your thesis will become true.

Dwarkesh Patel 1:26:17

That's really interesting. 

I'm just gonna do some rapid fire questions in the final few minutes. First, how can somebody be long AI but hedge for the possibility that Taiwan will be invaded? So, I don't know if I should put money into TSMC but I know that GPUs are going to be the next big thing or are going to be very important in the future. How do you make that position concrete?

Byrne Hobart 1:26:38

Oh man, that's really hard. I guess the next best thing would be looking at the Korean fabs because they're close but they don't produce the same chips. I guess the bet would be that Korea is the country most likely to catch up to Taiwan today if Taiwan is no longer an option. But if you're trying to make the AI bet conditional on Taiwan bet, I think a lot of what you want to do is actually think about how you underwrite the Taiwan invasion bet because that's probably the thing with the bigger long term impact. Maybe not. It gets tricky. 

Sometimes geopolitical changes can just lead to these permanent inflections. There's some kind of emissions that you can measure that are the result of copper mining and so we can see how much copper mining changes year to year throughout history and we do actually see it rising during the Roman Empire and then peaked and then went down and then didn't come back for a millennium. Sometimes there is a really unfortunate geopolitical inflection in underlying technology. But if you think about what that means in the real world, your big concern is not your portfolio in a world where we were an invasion is the reason AI does not happen. 

I would separate the invasion bet from the AI bet. And then I guess the sad answer is Intel and TSM are sort of America's last hopes on this. You can you can tell a story where invasion becomes more and more likely and the US does a sort of operation paperclip with no connotation about the the political views of the engineers of old but you know, operation paperclips all the best TSM engineers out to Arizona and has them all work on building those chips in the US in which case TSM is still you know, still a play although it's certainly lost some valuable assets. 

But that's a very tricky question. And it may be one of those things where it's kind of kind of hard to hit like. There are sufficiently bad things you know, there's not really a good meteor hedge in the minutes before the meteor hits us, maybe treasuries do outperform equities, but you don't really care. [Laughter]

Dwarkesh Patel 1:29:10

I'll definitely have to have you on again in a few weeks because we've gone through only a quarter of the questions. But it was really interesting. This is probably the most fun episode I've done so far. 

Byrne Hobart 1:29:31


Dwarkesh Patel 1:29:35

Just another plug. It's thediff.co. What is your Twitter handle?

Byrne Hobart 1:29:48

It's my full name @ByrneHobart

Dwarkesh Patel 1:29:51

I highly recommend it for the most schizophrenic and galaxy-brained takes. Awesome. Thanks, Byrne.

Thanks for reading The Lunar Society! Subscribe for free to receive new posts and support my work.