Longreads + Open Thread

Longreads

• Hunter Hopcroft has an admiring but pessimistic review of The Land Trap. One way to think about economic growth's impact on prices is to think in terms of ratios: if the economy grows faster than the population, everything that requires human effort gets more expensive because there's a rising ratio of output to people. And since it's hard to create new land, the price of land goes up relative to everything else, too. Meanwhile, land is expensive and immobile enough that it represents good collateral for loans, which means financial systems are often built around turning it into money. But once land is a big part of people's balance sheets, and those people influence policy, things get wonky; they don't want its price to go down, and there are many tempting policy choices to keep its price high. That's true of your political system whether it's an oligarchy controlled by a tiny minority of landowners, or a democracy where most voters are in home-owning households; either way, the political system is skewed in the interests of real estate holders, in a way that produces a wide variety of exciting economic catastrophes.
• Nicholas Decker has a good review of the economic literature on how unions affect wages, the main takeaway being that every measure we have is confounded one way or another. Even control groups aren't really well-controlled; he cites the example of comparing workplaces that voted to unionize by a slim margin, compared to the ones that voted against the union by a similarly small margin, where close elections tend to get contested and redone, making them sensitive to who'd adjudicating the vote-counting process. The piece is definitely opinionated, and Decker does a nice job of saying something inflammatory but uncontroversial at the beginning (yes, unions are a subset of cartels, but in some industries a cartel can be a more beneficial arrangement!) and then at the very end. So, whether you hate-skim it or hate-read it, you get some good material.
• Nate Silver has some reflections on the demise of FiveThirtyEight, catalyzed by Disney deleting the archives. He's understandably miffed. Aside from being a good short history of the business of data journalism, which Silver played a big part in inventing, it's also a good cautionary tale about a particular kind of economic arrangement. Sometimes, creative types muse that the ideal job is some kind of patronage system where the job exists to subsidize their work, which has either a nebulous and indirect benefit or is just personally appealing to whoever pays them. And this model can work. But the problem, especially if it happens under a big company, is that you're being paid by a specific fan, but the money actually comes from the parent company. So it's basically being in a business with a single customer who could, for inscrutable reasons, either change their mind or disappear. It would be nice if the market for statistically-informed journalism cleared better, but there's a constant problem with adverse selection—if you're really good at telling stories with data, you make a lot more money if that story has a ticker and a price target and your audience is someone managing a portfolio. Some of the best data-driven investigative journalism is found in public short reports, and perhaps soon, some of the best will be published by those who monetize by betting on their findings in prediction markets. 
• Bill Gurley has thoughts on open source, particularly as a defensive move. One model here is that creating a defensive open-source standard means collectively creating the economies of scale a monopolist would have, but not monetizing them like a monopolist. (Though, of course, it can be a great business to have a monopoly adjacent to such a standard.) 
• Bastianello and Fontanier: Partial Equilibrium Thinking, Extrapolation, and Bubbles. A reader sent this piece in response to The Diff's own agent-based market simulation. This one models a different part of the market, and it's specifying market participants' behavior without simulating them, but does produce an interesting, testable hypothesis: momentum starts to dominate the market when fundamentals are uncertain, perhaps because of some technology shock. That translates quite well, both at a macro scale (what 2030 will be like is less clear from the perspective of 2026 than it was from that of 2022) and in specific industries (like figuring out which quantum computing company has something real going on, if any). It's another showcase of the strengths and weaknesses of agent-based modeling: you can confirm that a plausible description of reality produces results that correspond with reality, but it's hard to be sure you've identified the true mechanism.
• On Read.Haus, a reader asks about generating alpha at the end of history, when the economy is perfectly modeled. I agree with the bot's answer, that this will never happen, and that better tools for modeling the economy actually complicate it faster than they solve it. You could have cleaned up if you were the only guy under the Buttonwood tree, but once every market participant has computers, having one yourself is insufficient to model them. But, taking the premise seriously, an economy with perfect forecasting and complete optimization is probably post-alpha entirely, and might not have any risk premia at all! Everything would just return the risk-free rate, and the amount invested rather than consumed would just reflect the average person's preference for more lifetime consumption or an early retirement leading to more lifetime leisure. It's a scary thought!
• In Capital Gains this week, we're looking at financial technologies that were forgotten and rediscovered: original-issue junk bonds, buybacks, and oil futures all feel like innovations that date back to the 70s and 80s, but they actually existed generations before and then disappeared.

Books

House of Fidelity: The Rise of the Johnson Dynasty and the Company That Changed American Investing: A stock picker is in the business of picking stocks, but an asset manager is in the business of finding the economic arrangement that best allows them to capture the upside from finding someone who is good at picking stocks or making other financial decisions. For many strategies, the answer to that question is either to run a hedge fund that charges performance fees, or to build a portfolio of capacity-constrained strategies without taking outside money. But before the market got so ruthlessly efficient, there's a good chance that the best way to maximize fees from a smart manager was to have them run a mutual fund.

This book does a good job of toggling back and forth between those questions. Fidelity has, a few times, been home to the most famous money-manager around. In the 60s, it was Gerry Tsai, who was considered a growth investor at the time but seems to have identified momentum as a good strategy (if you're trying to minimize commissions and tax impact, momentum in growth stocks is more plausible than momentum in value stocks, because the higher a company's P/E ratio gets, the harder it is to argue that it can't be even higher). In the 80s, they had Peter Lynch, who did a little bit of everything—he made money calling the cycle in autos, had a great eye for growth stocks, did a lot with small-cap banks, and somehow retained mindshare as a folksy stock-picker despite Buffett's efforts to corner that market.

One reason for Lynch's celebrity is that Fidelity went to great lengths to make it easy for investors to give them money. When the book isn't talking about finding and cultivating investment talent, a lot of it's about operations: what to bring in-house, what to automate when, how to jump on regulatory changes like the 401(k), etc. If the company made a macro bet, it was that one way or another, the American public would keep getting more interested in participating in the market, and that it ought to be as simple as possible for them to do so.

The book touches on a few Fidelity scandals, but they mostly fall into the everyone-else-was-doing-it category. There was a little front-running, back when front-running a fund you managed was seen as a slightly dubious perk, like being flexible with expense accounts. And one Fidelity manager bought high-yield bonds for a fund she managed, but kept the warrants sold alongside those bonds in a personal account in the name of her husband. But in that particular deal, Storer Communications, that seems to be exactly what the bond underwriters intended, and something that many other managers did. That's just the cycle of financial regulation: something is either allowed, or banned but indifferently enforced, and it turns into too big a grift to ignore and gets shut down. An asset manager would ideally avoid those problems entirely, especially with a name like "Fidelity," but for an organization that big, operating that long, it's actually a pretty solid track record.

Asset management goes through constant evolution in terms of what can be charged for and how much people will pay for it. As index funds have gotten more popular, it's harder for a diversified long-only strategy to be worth the extra fees. But that popularization can only happen if, beforehand, people got used to the idea of investing in equities, and perhaps paying someone else to handle all the details. Fidelity is in the position of anyone else who finds a good source of alpha: their own actions erode that alpha over time, and they have to keep looking for the next thing.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• What's happening in data labeling these days? It's been a high-growth space, but do we hit an equilibrium where a lot of the workforce is helping to train and evaluate models, or do we finish most of the easy stuff and turn this into a more specialized job?

Diff Jobs

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

• Well-funded, frontier AI neolab working on video pretraining and computer action models as the path to general intelligence is looking for researchers who are excited about creating machines that learn from experience, not text. Ideally you have zero-to-one pre-training experience and/or are a high-slope generalist who’s frustrated that the big labs aren't doing this. (SF)
• Lightspeed-backed team building the engineering services firm of the future is looking for founding members of technical staff excited about working alongside civil engineers to translate their domain expertise into the operating system that powers the next era of great American infrastructure. If you’re an engineer with strong product intuition, who's energized by access to users, and excited by the prospect of transforming how we design and construct our built world with frontier AI, this is for you. (NYC, SF or Remote) 
• Ex-Anduril, Ex-Abnormal Security, Ex-Bridgewater, fast growing startup bringing agentic cybersecurity to 99% of businesses via MSPs is looking for platform and machine learning engineers. Startup experience preferred; what matters most is that you've grown in scope and handled ambiguity over the last few years. (SF)
• Series A startup building multi-agent simulations to predict the behavior of hard to sample human populations is looking for a founding recruiter who’s able to attract and close the best research and engineering talent in the world. Experience building high-quality teams as a former founder, VC, or operator a plus. No formal experience in a “recruiting” function required. If you have experience communicating and persuading smart, disagreeable counterparties of your vision, this is for you. (NYC)
• AI Transformation firm with an ambition to build an economic world model to run swathes of the private, unstructured economy is looking for Systems Engineers, Platform Engineers, and business generalists who understand how to solve problems.  

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.

And: we're now actively deploying capital into early-stage companies through Anomaly. Our focus is on defense, logistics, robotics, and energy. If you'd like to chat, please reach out.