Longreads + Open Thread

Longreads

• Adam Mastroianni on how excellence in almost any field requires an unusual bundle of traits, which we're bad at modeling. It's pretty straightforward to make this argument by working backwards: there are many jobs that lots of people would like to do because they pay well, are high-status, or both. But most people don't have such jobs! Clearly, there is some kind of filtering mechanism that leads people to give up on them. (You could say that they get excluded from them, but if you really dream of being a professional musician, nobody but you can stop yourself from using all of your free time to practice and taking every gig you can get. Bob Dylan wrote in his autobiography that he played “morning, afternoon and night often falling asleep with his guitar in his hands.” He would play in coffee houses and pizza parlors, not exactly dream venues, for $3 to $5 a day. Similarly, you might find yourself off the track for being CEO of Goldman Sachs, but if you start your own thing you're CEO by default.) It's an interesting heuristic to work backwards from what you find less intolerable than anyone else into what your actual dream job ought to be.
• Zach Perkel tracks how much Hacker News has been dominated by discussions of AI, and how sentiment has changed. The most surprising result is that aside from one outlier, sentiment hasn't really shown much of a trend. Somehow, the progression from BERT and GPT-2 to Gemini 2.5, GPT-5, and Claude Opus 4.1 has not at all affected the tendency for 60-80% of Hacker News comments about AI to be roughly positive. It's a good look at sentiment; some people are blind early adopters who will try everything but only consistently use the same things everyone else does. Some people are naturally grumpy, and often perform the incredibly helpful function of sketching out where the newest and shiniest thing will often fail. (That, or sentiment detection is hard and there's some unknowable trend in actual belief that's roughly offset by a trend in the use of irony and sarcasm that LLMs can't pick up on.)
• Agustin Lebron argues against prediction markets, on the grounds that they don't function well as markets but do work as a de facto bounty on antisocial behavior. At some quoted odds, it's financially worth it to do all sorts of unpleasant and dangerous things specifically to get a payout. It's a thoughtful piece, and The Diff agrees with some of its points. But it's worth interrogating two questions more thoroughly. First, how many liquidity providers will take the other side of someone who wants to buy the YES contract for "At tonight's performance of Don Giovanni, will an audience member stand up and yell 'YEE-HAW' in the middle of the Catalogue Aria?" Anyone with a strong opinion on this contract's value is probably planning to influence it, so sellers want to price accordingly. But second, might contracts be a way to financially incentivize people to escape a local minimum? Consider a contract on Cuomo dropping out of the New York mayor's race. Low odds are a subsidy for people to insider-trade on their own efforts to persuade him, but as the odds get higher, some markets spawn prop bets and conditional markets, like "if he did drop out, where would his voters split?" and the like. These can sometimes reveal cases where a large number of people have slight confidence in one direction and a smaller set have extreme conviction in another one. There just aren't better ways to get something approximating "expert opinion, weighted by expertise" than a prediction market, though in the short term people can of course burn money to influence markets. In the long run, having more markets means having more venues where people get paid to be right, and the externalities of antisocial bets and gambling being a low-return vice might be worth that price.
• Jack Despain Zhou on which moment of the distribution traditionalists want to conserve: you can identify with your ancestors because they lived somewhere for a very long time, or because they refused to stay put and kept moving on to the next place. American culture in general seems to prize that kind of traditionalism, at least in part because we don't have very much of the "It's been this way for a thousand-plus years" variety. But that should inform US policy. It's not traditionalist to keep US Steel under US control in 2025, because the US Steel of today is simply a less important enterprise than the US Steel of the early 20th century, the time when that name became evocative of something. Matching the pattern more abstractly, a traditionalist view of US Steel is that a cutting-edge tech company founded by a first-generation immigrant could sell his company to a second-generation PE fund manager who flipped it to the public for a quick buck.
• Antti Ilmanen published a piece recently about how equity investors extrapolate growth and fixed income investors predict mean-reversion. It's a very interesting piece, in part because it drives home how historically unusual the US's stock market performance has been recently: US stocks' multiple of long-term earnings went from half the developed world average in 1990 to twice that today. The US probably does lead the world in the elasticity of actual value creation with respect to market cap creation—if you want to convert open market transactions into real-world outcomes, the US is just the best place to do it. So a period of US outperformance can last for a long time. But it's a very asymmetric bet; can 4% of the world's population produce the majority of the world's risk-adjusted equity market returns indefinitely? That is, implicitly, what the marginal price-setter in equities is betting on.
• In Capital Gains this week: which money managers disclose how much, and why? Some investors are terse to the point of rudeness, some engage in extensive creative writing. Most of this can be explained by their incentives, to protect the secrets behind repeatable trades and to spread the thesis behind an n-of-1 reflexive bet. But some of the differences come down to historical contingencies.

Books

Breakneck: China's Quest to Engineer the Future : Dan Wang used to write excellent annual letters about China, usually with a focus on political and economic issues but plenty of time for cuisine and the arts. It's a format that translates quite well to a book, because things can be organized more topically.

Wang's broad summary of his thesis is that America is run by lawyers and China by engineers, which is both a literal description of which academic/career background is most overrepresented in government, and a great summary of each country's pathologies—in the time it takes for an American real estate developer to get permission to build a single apartment building, a Chinese city could construct a dozen of them in a city that already has a double-digit vacancy rate and declining population! This model is obviously easy to exaggerate, but it's a great framing for many of the details—China has a great high-speed rail network, but the book mentions in an aside that there still aren't any places where it's completely safe to drink the tap water.

The book's strongest chapter covers China's Covid response. You'll read some detail, think to yourself, "that's the most cyberpunk-dystopian thing that has or will ever happen. It simply can't be topped." And then a page later, you'll read about people in Shanghai, stuck in their apartments while a loudspeaker-equipped drone outside chants the slogan "Repress your soul's yearning for freedom."

This book does a great job of highlighting some of the things that China's already ahead of the US on, and some of the weird brittleness of their system. It would be interesting to look at the US from the same lens, probably with a focus on social trust—that you can raise money from someone you've never met in-person, but if you're at a pharmacy in a big city you need to talk to someone to buy toothpaste. The Chinese system is great at building things and relatively worse at figuring out which things people want built. The demand problem is easy to ignore for a small and desperately poor country, but it's a much more important one for a country that makes a growing share of just about everything.

Open Thread

• Drop in any links or comments of interest to Diff readers?
• If there were a version of Breakneck about the modern US, who would write it?

Diff Jobs

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

• Thiel fellow founder (series A) building full-stack software, hardware, and chemistry to end water scarcity, is looking for an experienced software engineer to help build the core abstractions that enable global cloud seeding operations - from mission planning to post-flight analysis. If you have 5+ years building production systems with complex integration requirements, please reach out. 
• An OpenAI backed startup that’s applying advanced reasoning techniques to reinvent investment analysis from first principles and build the IDE for financial research is looking for software engineers and a fundamental analyst. Experience at a Tiger Cub a plus. (NYC)
• A leading AI transformation & PE investment firm (think private equity meets Palantir) that’s been focused on investing in and transforming businesses with AI long before ChatGPT (100+ successful portfolio company AI transformations since 2019) is hiring experienced forward deployed AI engineers to design, implement, test, and maintain cutting edge AI products that solve complex problems in a variety of sector areas. If you have 3+ years of experience across the development lifecycle and enjoy working with clients to solve concrete problems please reach out. Experience managing engineering teams is a plus. (Remote)
• Well funded, Ex-Stripe founders are building the agentic back-office automation platform that turns business processes into self-directed, self-improving workflows which know when to ask humans for input. They are initially focused on making ERP workflows (invoice management, accounting, financial close, etc.) in the enterprise more accurate/complete and are looking for FDEs and Platform Engineers. If you enjoy working with the C-suite at some of the largest enterprises to drive operational efficiency with AI and have 3+ YOE as a SWE, this is for you. (Remote)
• Ex-Citadel/D.E. Shaw team building AI-native infrastructure to turn lots of insurance data—structured and unstructured—into decision-grade plumbing that helps casualty risk and insurance liabilities move is looking for a data scientist with classical and generative/agentic ML experience. You will develop, refine, and productionize the company’s core models. (NYC, Boston)

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.