Back to Discover

Longreads + Open Thread

February 22, 2026
•The diff
Longreads + Open Thread

Here's something that might surprise you — social groups and insider networks are more about conferring trust and reputation than just exclusive access. Byrne Hobart highlights how in fields like tech or finance, forming tight-knit ingroups offers early career boosts, but over time, that can turn into harmful gatekeeping. Interestingly, Hobart points out that some groups, like Quakers in early industry, gained influence because their reputation for honesty created a high-fidelity communication network. Meanwhile, on the tech side, Jeff Dean from Google explains how scaling systems and reducing latency isn’t just about speed — it's about efficiency and user delight. And for professionals, Hobart notes that industries with high regulation or labor intensity tend to have prices that rise over time — think healthcare or education — because they’re hard to export or automate. So what does this all mean? Building genuine trust and reputation might be your best asset — more than any secret insider tip.

Screenshot-2022-02-17-at-12.51.50-1.png

Longreads

  • Zoë Bernard has a piece in Wired about the gay tech mafia. Or, rather, she has a piece about how there are various jokes and rumors about the existence of a cohort of influential gay founders and funders who help one another out, and, through extensive research and dozens of interviews, she was able to piece together—lots of cases where people confirm that yeah, there are definitely jokes and rumors about this! The piece is reasonably fun as a modern comedy of manners (the stories about straight founders and investors pretending to be gay to get deal flow, for example). But it misses an important point: any kind of social capital gets more valuable as coordination gets harder, and one of the best ways to have group cohesion is for group membership to confer underdog status. Turkey is an Islamic country where secular revolutionaries organized through masonic lodges, and Italy, perhaps the most Catholic country not ruled by a Pope, had a masonic lodge that appears to exist entirely for conspiring. One reason this is such a common pattern is that when a field gets important or lucrative, it starts to attract sociopaths and grifters, which means that field can't be as open and friendly as it used to be. But if you have some kind of in with the people you're trying to reach—you have mutual friends, you went to the same school, you're mutually romantically interested, etc.—there's a way to get in. That kind of leg up can be helpful early in a career, but giving someone an unfair leg-up is harmful later in one's career. So this kind of thing is only sustainable if it leads to higher trust, lower transaction costs, and net value creation. Let a thousand elusive ingroups bloom.
  • Not that every ingroup gets ahead purely because it's an ingroup—sometimes, group membership confers direct reputational benefits. Will Manidis and Nabeel Qureshi write about why Quakers were so prominent in business during the early years of the industrial revolution. Following a minority religion, especially in a country with an established church, marks you as a deviant, but if the particular faith you follow forbids you from lying, then it's a deviation in a prosocial direction. A group of people who are known to be unusually honest are basically a high-fidelity communications network, and when bandwidth is constrained, the things that still get communicated are the valuable ones.
  • Jeff Dean interviewed on the Latent Space podcast. Google (disclosure: long) is an amazing story of scaling; the latency for an AI-generated search result is lower today than the latency for a non-AI search result was early in their existence (in early 1999, the average search took 3.5 seconds). And scaling means identifying bottlenecks in advance, and building accordingly. So this interview features things like a discussion of how batching some processes saves you picojoules at a time, which really adds up (especially if that scaling means that you can offer the same energy-constrained service to more users at a lower cost—you're going to be the one getting all that usage!). Energy is the asymptote at the end, but the thing that governs how quickly you get there is latency: " I’m a big believer in pushing on latency because I think being able to have really low latency interactions with a system you’re using is just much more delightful than something that is, you know, 10 times as slow or 20 times as slow." And that implicitly hits a third variable: scaling sometimes means following the aesthetic imperative that inefficiency is ugly and speed is beautiful; ugly systems have more opportunities to break.
  • Chris Guo has a fun piece on calibrating compensation. One of the reasons the financial services sector works so well is that it can pay people ludicrous sums, and has coevolved with some effective money sinks—cocaine, alimony, Effective Altruism—to the point that people who've earned more by 30 than most people will in a lifetime still feel like they can't afford to quit. It's a good tour of some misaligned compensation schemes that gave people enough money to quit working for their employer and promptly go into the business of competing with that employer. But this is really an ecosystem problem, not an incentive problem: from the perspective of any given worker in a mature industry, it's fine if people are quitting to do their own thing so long as a) they need some capital to do it, b) they'll give up equity in exchange for that capital, and c) their former boss and colleagues are in an unusually good position to underwrite that bet. There are people who've had long and prosperous careers in tech where most of their money came from the right seed-stage investment, even though they've had other successes along the way. This is bad for big companies, but high-variance bets are the kinds of bets big companies aren't best positioned to take.
  • Benjamin Breen asks what is happening to writing now that LLMs can do such a mass-adequate job? In some cases, writing is a pretty impersonal force: if I want to answer some specific question about my taxes, I'm not using this as a pretext to develop some sort of personal connection with whoever is going to explain the nuances of the wash-sale rule or whatever. But in other categories—ones with a substantial overlap with what we think of as "writing for a living"—that parasocial relationship is important. Reading a narrative means looking at the world through someone else's eyes, and knowing something about the person whose eyes you're borrowing is important. So it shifts the bundle, utterly destroys the economics of part of it, but might end up improving some of what's left.
  • A Read.Haus user asks about AEI's famous chart of the century, showing different rates of inflation for different products—in categories like consumer electronics and clothes, prices are down; for housing, healthcare, and education, they're way up. You can look at this as a commentary on regulated versus unregulated industries, but it's also a look at which industries are a) labor-intensive, and b) can't be easily exported. These industries often end up being highly regulated industries in part because both sides of the transaction are in the same country, operating under the same legal system, and thus interested in lobbying in their interests. It's a noisy story; if virtual reality and remote work get good enough, perhaps there won't be a big difference between living 45 minutes away from the nearest big city and living four hours away instead; it's possible to learn a lot online, but when it becomes possible to earn a credential as impressive as a bachelor's in CS from MIT, the price of education will decline. This makes the graph tricky from a political perspective: what you'd want, from a points-scoring perspective, is to cut some big cost by a few percentage points next year. But what you can get, in practice, is an increase in the odds that the price level will suddenly drop by half at some point in the next two decades.
  • In Capital Gains this week, a look at how to structure compensation. Even in cases where you could pay some a cut of their P&L, you run into the problem that this is tantamount to arguing that all earnings deserve the same multiple.

You're on the free list for The Diff! This week, paying subscribers read about making markets in death ($), how the SaaS-pocalypse is self-referential ($), and why software engineers are earning more per year but spending more of their time unemployed ($). Upgrade today for full access.

Upgrade Today

Books

Chicagonomics: The Evolution of Chicago Free Market Economics: There are lots of biographies, and lots of books about institutions, but relatively fewer books about specific parts of institutions. There are a few, like the MITI book or the one about Xerox PARC, but in most cases writing about a small part of a big organization means periodically throwing in some remarks about an exogenous issue like a funding cut that randomly complicates things. The Chicago School is partly an ideological cluster, and partly an approach to economics that happens to produce answers that favor one ideology over the other.

The author is upfront about his biases as a kind of classical liberalism hipster: the Chicago greats do a fine job when describing the world, but go a bit too far when they suggest cutting taxes or relaxing some regulations. By far the best tell for this viewpoint is saying that you love Adam Smith, and then swerving over to talking about the Theory of Moral Sentiments rather than the Wealth of Nations. Which is fair enough, but it's like saying that you're a huge Hedy Lamarr fan and then revealing that it's because of your more general interest in encrypted communications over publicly-accessible networks.

The early U Chicago economics story is roughly what you'd expect from a school that's well-funded but hasn't established its reputation yet, and is trying to find its footing. Postwar, they get a flood of new students, and also start moving in a much more libertarian direction. Which the author finds confusing, but which from a libertarian perspective makes plenty of sense: by the end of the Second World War, there had been more than a decade and a half of extraordinary government intervention in the economy, and postwar defense spending and other policies implied more of the same. To someone who'd seen government spending rise from 3% of GDP in the 20s to 17% in the 50s, it must have felt like the world was ending and the only question was how fast each country would go socialist. Of course they were alarmed! It's as unsurprising as people getting more worried about high wealth inequality and rising corporate profits today—the numbers are at historically high levels and that makes them salient.

If you're writing an intellectual history of people you agree with, you don't have to think about causation. All you have to do is tell the story of a group of clever people who all agree with you because all of you are smart enough to figure out the truth. But from this book's perspective, the Chicago School started out being reasonable moderate classical liberals, and eventually went over to the shallower bits of off-the-deep-end. Here, it's a more personal story: at one point, the book notes that Chicago is somewhat unique in that teacheres tended to live within walking distance, and that some of them (Milton Friedman and Aaron Director in particular) were exceptional interpersonal communicators who could accrete a coherent ideology around themselves.

Open Thread

  • Drop in any links or comments of interest to Diff readers.
  • What are some other good history-of-a-school-of-thought books? It's always fun to compare how a group originally conceived of their ideology to the more marketable version they eventually get associated with.

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

  • High-growth startup building dev tools for wrangling and debugging complex codebases is looking for someone who can personally execute the SaaS bear case: review the third-party software they use and figure out what to keep, what to drop, and what to implement in-house. (SF, DC)
  • Ex-Bridgewater, Worldcoin founders using LLMs to generate investment signals, systematize fundamental analysis, and power the superintelligence for investing are looking for machine learning and full-stack software engineers (Typescript/React + Python) who want to build highly-scalable infrastructure that enables previously impossible machine learning results. Experience with large scale data pipelines, applied machine learning, etc. preferred. If you’re a sharp generalist with strong technical skills, please reach out
  • Ex-Citadel/D.E. Shaw team building AI-native infrastructure that turns lots of insurance data—structured and unstructured—into decision-grade plumbing that helps casualty risk and insurance liabilities move is looking for forward deployed data scientists to help clients optimize/underwrite/price their portfolios. Experience in consulting, banking, PE, etc. with a technical academic background (CS, Applied Math, Statistics) a plus. Traditional data scientists with a commercial bent also encouraged. (NYC)
  • Series A startup that powers 2 of the 3 frontier labs’ coding agents with the highest quality SFT and RLVR data pipelines is looking for growth/ops folks to help customers improve the underlying intelligence and usefulness of their models by scaling data quality and quantity. If you read axRiv, but also love playing strategy games, this one is for you. (SF)
  • A startup is automating the highest tier of scientific evidence and building the HuggingFace for humans + machines to read/write scientific research to. They’re hiring engineers and academics to help index the world’s scientific corpus, design interfaces at the right level of abstraction for users to verify results, and launch new initiatives to grow into academia and the pharma industry. A background in systematic reviews or medicine/biology is a plus, along with a strong interest in LLMs, EU4, Factorio, and the humanities.

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

Audio Transcript

Screenshot-2022-02-17-at-12.51.50-1.png

Longreads

  • Zoë Bernard has a piece in Wired about the gay tech mafia. Or, rather, she has a piece about how there are various jokes and rumors about the existence of a cohort of influential gay founders and funders who help one another out, and, through extensive research and dozens of interviews, she was able to piece together—lots of cases where people confirm that yeah, there are definitely jokes and rumors about this! The piece is reasonably fun as a modern comedy of manners (the stories about straight founders and investors pretending to be gay to get deal flow, for example). But it misses an important point: any kind of social capital gets more valuable as coordination gets harder, and one of the best ways to have group cohesion is for group membership to confer underdog status. Turkey is an Islamic country where secular revolutionaries organized through masonic lodges, and Italy, perhaps the most Catholic country not ruled by a Pope, had a masonic lodge that appears to exist entirely for conspiring. One reason this is such a common pattern is that when a field gets important or lucrative, it starts to attract sociopaths and grifters, which means that field can't be as open and friendly as it used to be. But if you have some kind of in with the people you're trying to reach—you have mutual friends, you went to the same school, you're mutually romantically interested, etc.—there's a way to get in. That kind of leg up can be helpful early in a career, but giving someone an unfair leg-up is harmful later in one's career. So this kind of thing is only sustainable if it leads to higher trust, lower transaction costs, and net value creation. Let a thousand elusive ingroups bloom.
  • Not that every ingroup gets ahead purely because it's an ingroup—sometimes, group membership confers direct reputational benefits. Will Manidis and Nabeel Qureshi write about why Quakers were so prominent in business during the early years of the industrial revolution. Following a minority religion, especially in a country with an established church, marks you as a deviant, but if the particular faith you follow forbids you from lying, then it's a deviation in a prosocial direction. A group of people who are known to be unusually honest are basically a high-fidelity communications network, and when bandwidth is constrained, the things that still get communicated are the valuable ones.
  • Jeff Dean interviewed on the Latent Space podcast. Google (disclosure: long) is an amazing story of scaling; the latency for an AI-generated search result is lower today than the latency for a non-AI search result was early in their existence (in early 1999, the average search took 3.5 seconds). And scaling means identifying bottlenecks in advance, and building accordingly. So this interview features things like a discussion of how batching some processes saves you picojoules at a time, which really adds up (especially if that scaling means that you can offer the same energy-constrained service to more users at a lower cost—you're going to be the one getting all that usage!). Energy is the asymptote at the end, but the thing that governs how quickly you get there is latency: " I’m a big believer in pushing on latency because I think being able to have really low latency interactions with a system you’re using is just much more delightful than something that is, you know, 10 times as slow or 20 times as slow." And that implicitly hits a third variable: scaling sometimes means following the aesthetic imperative that inefficiency is ugly and speed is beautiful; ugly systems have more opportunities to break.
  • Chris Guo has a fun piece on calibrating compensation. One of the reasons the financial services sector works so well is that it can pay people ludicrous sums, and has coevolved with some effective money sinks—cocaine, alimony, Effective Altruism—to the point that people who've earned more by 30 than most people will in a lifetime still feel like they can't afford to quit. It's a good tour of some misaligned compensation schemes that gave people enough money to quit working for their employer and promptly go into the business of competing with that employer. But this is really an ecosystem problem, not an incentive problem: from the perspective of any given worker in a mature industry, it's fine if people are quitting to do their own thing so long as a) they need some capital to do it, b) they'll give up equity in exchange for that capital, and c) their former boss and colleagues are in an unusually good position to underwrite that bet. There are people who've had long and prosperous careers in tech where most of their money came from the right seed-stage investment, even though they've had other successes along the way. This is bad for big companies, but high-variance bets are the kinds of bets big companies aren't best positioned to take.
  • Benjamin Breen asks what is happening to writing now that LLMs can do such a mass-adequate job? In some cases, writing is a pretty impersonal force: if I want to answer some specific question about my taxes, I'm not using this as a pretext to develop some sort of personal connection with whoever is going to explain the nuances of the wash-sale rule or whatever. But in other categories—ones with a substantial overlap with what we think of as "writing for a living"—that parasocial relationship is important. Reading a narrative means looking at the world through someone else's eyes, and knowing something about the person whose eyes you're borrowing is important. So it shifts the bundle, utterly destroys the economics of part of it, but might end up improving some of what's left.
  • A Read.Haus user asks about AEI's famous chart of the century, showing different rates of inflation for different products—in categories like consumer electronics and clothes, prices are down; for housing, healthcare, and education, they're way up. You can look at this as a commentary on regulated versus unregulated industries, but it's also a look at which industries are a) labor-intensive, and b) can't be easily exported. These industries often end up being highly regulated industries in part because both sides of the transaction are in the same country, operating under the same legal system, and thus interested in lobbying in their interests. It's a noisy story; if virtual reality and remote work get good enough, perhaps there won't be a big difference between living 45 minutes away from the nearest big city and living four hours away instead; it's possible to learn a lot online, but when it becomes possible to earn a credential as impressive as a bachelor's in CS from MIT, the price of education will decline. This makes the graph tricky from a political perspective: what you'd want, from a points-scoring perspective, is to cut some big cost by a few percentage points next year. But what you can get, in practice, is an increase in the odds that the price level will suddenly drop by half at some point in the next two decades.
  • In Capital Gains this week, a look at how to structure compensation. Even in cases where you could pay some a cut of their P&L, you run into the problem that this is tantamount to arguing that all earnings deserve the same multiple.

You're on the free list for The Diff! This week, paying subscribers read about making markets in death ($), how the SaaS-pocalypse is self-referential ($), and why software engineers are earning more per year but spending more of their time unemployed ($). Upgrade today for full access.

Upgrade Today

Books

Chicagonomics: The Evolution of Chicago Free Market Economics: There are lots of biographies, and lots of books about institutions, but relatively fewer books about specific parts of institutions. There are a few, like the MITI book or the one about Xerox PARC, but in most cases writing about a small part of a big organization means periodically throwing in some remarks about an exogenous issue like a funding cut that randomly complicates things. The Chicago School is partly an ideological cluster, and partly an approach to economics that happens to produce answers that favor one ideology over the other.

The author is upfront about his biases as a kind of classical liberalism hipster: the Chicago greats do a fine job when describing the world, but go a bit too far when they suggest cutting taxes or relaxing some regulations. By far the best tell for this viewpoint is saying that you love Adam Smith, and then swerving over to talking about the Theory of Moral Sentiments rather than the Wealth of Nations. Which is fair enough, but it's like saying that you're a huge Hedy Lamarr fan and then revealing that it's because of your more general interest in encrypted communications over publicly-accessible networks.

The early U Chicago economics story is roughly what you'd expect from a school that's well-funded but hasn't established its reputation yet, and is trying to find its footing. Postwar, they get a flood of new students, and also start moving in a much more libertarian direction. Which the author finds confusing, but which from a libertarian perspective makes plenty of sense: by the end of the Second World War, there had been more than a decade and a half of extraordinary government intervention in the economy, and postwar defense spending and other policies implied more of the same. To someone who'd seen government spending rise from 3% of GDP in the 20s to 17% in the 50s, it must have felt like the world was ending and the only question was how fast each country would go socialist. Of course they were alarmed! It's as unsurprising as people getting more worried about high wealth inequality and rising corporate profits today—the numbers are at historically high levels and that makes them salient.

If you're writing an intellectual history of people you agree with, you don't have to think about causation. All you have to do is tell the story of a group of clever people who all agree with you because all of you are smart enough to figure out the truth. But from this book's perspective, the Chicago School started out being reasonable moderate classical liberals, and eventually went over to the shallower bits of off-the-deep-end. Here, it's a more personal story: at one point, the book notes that Chicago is somewhat unique in that teacheres tended to live within walking distance, and that some of them (Milton Friedman and Aaron Director in particular) were exceptional interpersonal communicators who could accrete a coherent ideology around themselves.

Open Thread

  • Drop in any links or comments of interest to Diff readers.
  • What are some other good history-of-a-school-of-thought books? It's always fun to compare how a group originally conceived of their ideology to the more marketable version they eventually get associated with.

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

  • High-growth startup building dev tools for wrangling and debugging complex codebases is looking for someone who can personally execute the SaaS bear case: review the third-party software they use and figure out what to keep, what to drop, and what to implement in-house. (SF, DC)
  • Ex-Bridgewater, Worldcoin founders using LLMs to generate investment signals, systematize fundamental analysis, and power the superintelligence for investing are looking for machine learning and full-stack software engineers (Typescript/React + Python) who want to build highly-scalable infrastructure that enables previously impossible machine learning results. Experience with large scale data pipelines, applied machine learning, etc. preferred. If you’re a sharp generalist with strong technical skills, please reach out
  • Ex-Citadel/D.E. Shaw team building AI-native infrastructure that turns lots of insurance data—structured and unstructured—into decision-grade plumbing that helps casualty risk and insurance liabilities move is looking for forward deployed data scientists to help clients optimize/underwrite/price their portfolios. Experience in consulting, banking, PE, etc. with a technical academic background (CS, Applied Math, Statistics) a plus. Traditional data scientists with a commercial bent also encouraged. (NYC)
  • Series A startup that powers 2 of the 3 frontier labs’ coding agents with the highest quality SFT and RLVR data pipelines is looking for growth/ops folks to help customers improve the underlying intelligence and usefulness of their models by scaling data quality and quantity. If you read axRiv, but also love playing strategy games, this one is for you. (SF)
  • A startup is automating the highest tier of scientific evidence and building the HuggingFace for humans + machines to read/write scientific research to. They’re hiring engineers and academics to help index the world’s scientific corpus, design interfaces at the right level of abstraction for users to verify results, and launch new initiatives to grow into academia and the pharma industry. A background in systematic reviews or medicine/biology is a plus, along with a strong interest in LLMs, EU4, Factorio, and the humanities.

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

View original article
0:00/0:00