Longreads + Open Thread

Longreads

• A few years back, a friend observed how peculiar it was that Oliver Sacks, an incredible writer, also happened to meet so many fascinating patients. What are the odds? Anyway, Rachel Aviv's New Yorker profile of Sacks makes use of previously-unreleased letters and journal entries to reveal that some of the details in Sacks' stories were made up. This is really unfortunate, because he was a fantastic writer. It's too bad the stories were basically Robin Cook novels for the London Review of Books set. In retrospect, some of the stories were sketchy: Sacks tells one story about a pair of autistic twins who were obsessed with prime numbers (a plausible comorbidity), but also that they were capable of calculating twenty-digit prime numbers. (Apparently checking this will Miller-Rabin would involve 1-2k individual calculations.) It's not so much that this is completely impossible—Ramanujan was capable of some absolutely crazy calculations—just that it's surprising that the first person to notice this would happen to be a bestselling author. It's kind of a Two Cultures probem—these liberal arts bros really need to take a STEM course!
• Casey Handmer has some updated long-term predictions about energy. Much to think about here: the cool thing about energy is that it's a universal complement&dmash;when it gets cheaper, everything gets cheaper, but things vary in their energy intensity and this makes the effects nonlinear.
• Erik Hoel has a nice model of stuck culture as overfitting. The more accurately you can run a principal component analysis of what people find appealing, the more you can over-optimize for it, and at some point you run into an uncanny valley effect. This holds true in plenty of human interactions: cosmetics, bodybuilding, pick-up artistry, populism, snack foods—at some point, it's so over-optimized that you can't ignore it. (P.J. O'Rourke once joked that he'd freebased cocaine, and suddenly understood why the lab rats would press the lever for it until they starved to death, so he didn't do it again.) This thesis still raises the question of why nobody breaks the equilibrium. If you're being properly epsilon-greedy ($, Diff), you still need to throw some randomness into the mix from time to time. Hopefully enough stagnation creates the demand for a renaissance.
• Sam Kriss on how AI writes. This is a fun preview of the future of language: a synthesis between the linguistic norms of American English (who have posted the most online, and infected the writing style of people who learn English as a second language) and Nigerian English (because the models' outputs are rated by whichever part of the global labor market is Anglophone, has Internet access, and lives in a country with very low prevailing wages). The piece is also a good source of AI tells in fiction: genre fiction has had infinitely fractal micro-genres for a while now (the Cambrian explosion of romance novel subtypes predates LLMs, though it was probably helped by whatever earlier generation of ML recommendation systems told people which Kindle book to buy next). But now a substantial fraction of that fiction seems to be LLM-assisted ("There are now hundreds of self-published books on Amazon featuring Elara Voss or Elena Voss; before 2023, there was not a single one.")
• Tanner Greer on the first "tech-right" political coalition of the late 19th century. This is a very fun piece both for its framing of US political history and for the direct analogy it makes. When the scope of government expands, it tends ot get more technocratic, because the technocrats are the only ones who know what's going on. Greer points to the Civil War as the origin of this: soldiers and generals got firsthand experience being part of large organizations that had to coordinate many unrelated tasks to get anything done—pretty good practice for running a steel mill, railroad, or oil refinery! And the postwar Republican political monopoly meant that there wasn't as much pressure to enact popular policies—elites could go ahead and do what they felt was in the national interest (or in their own). And what replaced it wasn't a reduction in the scope of the government; right-technocrat Hoover was replaced by FDR, who promptly staffed his administration with left technocrats.
• In this week's Capital Gains, we consider why it's hard to structure executive pay in a way that aligns incentives and doesn't give away shareholders' money. Fortunately, there are some non-financial forces that keep incentives aligned.
• A user shared a fun question on the Diff Read.Haus chat: why was the yield curve inverted in the 19th century? It's another case where a sufficiently long time series of interest rate data is compariing completely different kinds of rates. 19th century rates look weird to us, because the main long-term borrowers were big and fairly creditworthy—countries, municipalities, and transportation companies with monopolistic economics. Short-term debt was for smaller, less established companies, who were riskier credits—if they needed capital, it was either because they were shrinking and losing money, or because they were growing and had to borrow to fund that growth. Either way, lenders were taking risks. This was also a very local market; when transportation is slow and communication is expensive, each city will have a slightly different credit market, and money can be tight in one place and available somewhere else. This also tends to distort rates upward, since lenders can switch to longer-term bonds if nobody wants to borrow short-term locally, but borrowers don't have that option. And that leads to the last reason the yield curve was inverted: under a deflationary gold standard, there's no inflation risk and little duration risk, so borrowers don't have to pay lenders as much to lock up the money.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• Are there examples similar to the rates one where we have a long time series that technically tracks the same variable, but is actually measuring something completely different at different points in the series? 

Diff Jobs

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

• A well-funded, Series C startup building the platform and agent primitives to drive operational transformation at large, complex institutions (starting with higher education) is hiring platform engineers. The work spans distributed systems, applied AI, and full-stack infrastructure, focused on deploying reliable agents that meaningfully bend institutional cost curves. (Remote)
• A first of its kind PE firm focused on buying businesses burdened by high financial friction (cross-border payments, FX considerations, working capital drag, etc.) and building/applying blockchain technologies that drive operational efficiencies and unlock durable opportunities for growth is looking for Investment Associates excited to identify, diligence, acquire, and transform businesses with decentralized primitives. Experience in private equity/banking and an intuition for how decentralized technologies can drive operational efficiencies/create long-term value required. (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 Associates, VPs, and Principals to lead AI transformations at portfolio companies starting from investment underwriting through AI deployment. If you’re a generalist with deal/client-facing experience in top-tier consulting, product management, PE, IB, etc. and a technical degree (e.g., CS/EE/Engineering/Math) or comparable experience this is for you. (Remote)
• A hyper-growth startup that’s turning the fastest growing unicorns’ sales and marketing data into revenue (driven $XXXM incremental customer revenue the last year alone) is looking for a senior/staff-level software engineer with a track record of building large, performant distributed systems and owning customer delivery at high velocity. Experience with AI agents, orchestration frameworks, and contributing to open source AI a plus. (NYC)
• A growing pod at a multi-manager platform is looking for new quantitative researchers, no prior finance experience necessary, exceptional coding and stats skills required. 250k+ (NYC)
• Series-A defense tech company that’s redefining logistics superiority with AI is looking for a MLE to build and deploy models that eliminate weeks of Excel work for the Special Forces. If you want to turn complex logistics systems into parametric models, fit them using Bayesian inference, and optimize logistics decision-making with gradient descent, this is for you. Python, PyTorch/TensorFlow, MLOps (Kubernetes, MLflow), and cloud infrastructure experience preferred. (NYC, Boston, SLC)

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.