Longreads + Open Thread

Longreads

• Editors are an invisibly influential job. If they do their work well, and they're working with good people, readers don't give them any thought—you might like Raymond Carver or Donna Tartt or Cormac McCarthy, but you probably don't identify as a fan of the late Gary Fisketjon, who edited all of them, and many of the other big names in literary fiction in the 80s through 2000s. Being an editor might be the best way for someone who doesn't want the downsides of fame to capture some of the upside—he spent a lot of time hobnobbing with fascinating people, but was rarely the most famous person in the room, so he didn't face the information asymmetry downside of fame, which is that strangers who strike up a conversation have an imagined one-way relationship with you whereas you have no idea who they are. One notable thing about Fisketjon is that while he made his mark in a creative career, part of why he could do so was material constraints—high inflation in the 70s made hardcovers more expensive, which made them a riskier way to package a first-time author's book. Mass-market paperbacks were focusing more on, well, mass-market products. But trade paperbacks were cheap enough that they were worth using for more experimental fare. And the fact that that was the medium in which some of the greats first published meant that the connotation of that form factor was that the book might be really good, and you might be the first of your friends to realize this. Some talents are not to be denied, and surely some great authors and editors would do well even if books were harder to produce in various ways. But when it gets cheaper to experiment in some medium, more people try it, and some of them turn out to be great at it.
• Richard Radford was studying economics at Cambridge at the start of the Second World War, and promptly enlisted. In 1942, he was captured, and spent the rest of the war in prisoners of war camps. Which led to this wonderful paper on the economics of P.O.W. camps. It features everything you'd want to see from an economic petri dish where everyone gets the same Red Cross rations and there's little opportunity for productive labor: choosing a standard currency as a game theory exercise in reducing transaction costs (they used cigarettes, of course), barriers to trade impeding the optimal allocation of resources (one Urdu-speaking prisoner made money buying meat from Sikh prisoners in exchange for butter and jam from his fellow English-speakers), booms, and busts. The result was a surprisingly complex trade system considering how few people were involved and how long they had to perfect it. But that's also because getting smarter about trade was the only way anyone had to improve their standard of living! In a way, the modern economy is like this, in that it takes a tiny fraction of the available labor-hours to produce enough to meet everyone's basic wants, and as we go beyond that, the big problems are allocation problems solved by a mix of ads, financial markets, and central planning within large organizations.
• In 2001, Larissa MacFarquhar profiled Richard Posner in The New Yorker. This article was a big influence on me personally, though it would be a long time before I read any of Posner's work. Posner loved to throw out wild assertions, or to state as ideals what other people usually confine to thought experiments. He was publishing in pretty prestigious venues—a judge is a kind of writer who gets to opine about how things should be and thus make them so—but he retained the soul of a comments-section troll. Couple that with a high-output mind ("[In 2001] alone, while working full time as a judge and teaching at the University of Chicago Law School, he published “Breaking the Deadlock,” a book about the Bush-Gore election; a second, updated edition of his 1976 book, “Antitrust Law”; and two collections of essays. He also wrote “Public Intellectuals,” a four-hundred-page diatribe against the species, and “Law, Pragmatism, and Democracy"[.]"), and you have someone who was basically a walking outrage machine. But also an incredibly thoughtful guy, and one who was willing to change his mind—the late 60s turned him conservative rather than progressive, 2008 pushed him back in the other direction. Posner made his fans smarter by giving them good ideas, and made his detractors smarter by putting them through the brutal intellectual workout of keeping up with him. (And, on a personal note: I was once on a date, found out that my date wasn't familiar with Posner, and insisted that she read this exact article posthaste. Our twelfth anniversary is next month.)
• High performers are often very weird; if some traits and habits have tradeoffs, then someone who is at the extremes on some output will also have some other extremes. Terence Tao is one of the most normal, even-keeled people who's also far on the tail of an important distribution, as this Tyler Cowen interview shows. One theory I subscribe to is that the happiness of a smart person is a function of how old they are when they discover they aren't the smartest person, and if someone might actually be the smartest person alive right now (however you'd measure that), the only way to engineer this is to get them in an extremely high-performing peer group at an age when a) they're impressionable, and b) they still have a lot to learn. He had a sufficiently accelerated educational experience that he actually managed to fail some classes, though these were college classes on quantum mechanics and Fortran programming and he was in his early to mid teens. (Also, he notes that in both cases he failed partly because he refused to do some of the work. Also a good lesson to internalize at an early age!) One of his fairly friendly and egalitarian views is that math should have prized exposition more. That's a lot more valuable in what he refers to as a time of "proof abundance," when AI can tell us amazing mathematical facts that would no doubt be fascinating if they gave us any inkling whatsoever of how they got them. (Genius claims that, late in his career, Richard Feynman basically did this: he loved figuring things out, hated the rigmarole of publishing things, so when his grad students would approach him with some new insight he'd basically say "Oh, yeah, I figured that out in 1973, probably should have told someone." This is probably why we have different words for cruelty and sadism—he didn't mean anything by it! Just wanted to mention that he'd doodled the core insight of that dissertation on a napkin one time and then forgotten about it.)
• In Foreign Affairs, Gordon Liao, Eswar Prasad, and Tony Zhang write about the paradoxes of dollar dominance. A reserve currency needs to be both trusted and ubiquitous. The dollar is not as trusted as it used to be, given the US's penchant for freezing the funds of people or countries it dislikes—for the purposes of reserve status, it doesn't matter whether or not America is right to do this; what matters is whether bad people want to stash their ill-gotten gains in some other medium. A chronic deficit has a paradoxical effect, where it makes the dollar less trustworthy because it has lower purchasing power, but more ubiquitous since there's more dollar-denominated debt. And as long as the dollar is the global default, stablecoins also reduce the US's control over how many dollars are in circulation, but cement the dollar's role as the default medium for transactions.
• A Read.Haus reader asks if smart contracts are dead, at least in the sense of providing a bull case for Ethereum and Solana. Crypto tends to overrate how important customization is and underrate network effects (they occasionally swing around in the opposite direction and try to bootstrap a network by giving away lots of tokens—as Antonio García Martínez has pointed out a few times, if your growth strategy is to give away tokens, you have no idea what your customer acquisition cost is. A side effect of this is that projects fight for share within crypto, rather than facing the brutal reality of how low the ROI would be on non-crypto-native user acquisition). I wouldn't write smart contracts off, but there needs to be some reason that people don't just opt in to standard contracts, or use the legal system's rails for customized ones. Crypto is in the middle of those two extremes, and the middle is always a tough place to be.
• In this week's Capital Gains: I'm baffled by how many people claim that Medicaid is somehow a subsidy to big companies like Walmart. First of all, if they're talking about Walmart specifically, they're using the worst possible example. Second, the phenomenon they think it's an example of doesn't make any sense, except in a case where the average worker is so close to subsistence that without redistribution they'll literally die. This doesn't describe anyone in America in the last half-century. It just doesn't make any sense once you think about it.

Books

Evolution in Four Dimensions, revised edition: Genetic, Epigenetic, Behavioral, and Symbolic Variation in the History of Life: Darwinian evolution is a ridiculously powerful model. If you're sufficiently tolerant of tautologies, it explains everything: we were selected to do it, thus we do it thus you (don't get a second date/get served with divorce papers/get great material for your book about the relevance of evolutionary psychology). Evolutionary psychology is a weird branch of science, in that it functions well as both a bland explanation of human behavior and as a bracing or thrilling answer to the fundamental question of human psychology—what does tonight have in store?

The book proposes that instead of the one-dimensional view of genes, we consider three more: which genes get expressed, behaviors we learn through observation and imitation, and even more abstract changes passed down through symbolic communication. The most fun and unpredictable of these is the first new dimension; it turns out that you can have lots of natural selection and phenotypic diversity with a fixed genome, if the genes that are expressed in parents tend to be expressed in offspring. That moves things in a more Lamarckian direction, but it's Lamarck bounded by Darwin—genes can be toggled on and off, but they have to exist first.

But the book is weaker when it tries to attack the Darwinian outlook. For example, it has an extended riff on how an alien anthropologist might assume that humans have some sort of reading gene, and don't need some external influence to make us literate. Which is, of course, not true. On the other hand: I have ended up in various situations where it's helpful to have a piece of paper signed by some medical authority saying that a kid of mine has ADHD. And one surprise about this is that the ADHD-diagnosed kids absolutely love to read. Since sitting still and doing one task for a long time is, at least in theory, a way you'd identify someone as not having ADHD, this seems confusing. But humans are novelty-seeking, the ones with ADHD especially so, and a book is basically a compressed stream of novelty that tries to strip out all of the boring stuff. Of course they love it! It's one new thing after another! Novelty-seeking and compression-seeking could be somewhat hard-wired, and while you can't predict the means by which they're expressed, you can still be unsurprised that they get expressed.

It is useful to consider how these other kinds of variation affect natural selection. The ability to be violent, or to withstand violence, doesn't make nearly as big a difference as it once did in people's evolutionary outcomes. But all of these non-genetic factors tend to move faster than genes. Genes can work surprisingly quickly, as the silver fox experiment shows, but they're still marginal changes over many generations. Whereas literacy, or inherited switching on or off of particular genes, happens faster. But that also means that these features tend to describe a brief period of change, not an indefinite trend: plenty of species have some mix of epigenetic, behavioral, and symbolic adaptation (e.g. eusocial insects can communicate the location of food sources with smells or dances) but humans are able to accomplish unique things that those species can't. So, keeping all of those other means of changing behavior and physical characteristics in mind arguably makes genes even more important, since they're the canvas for the rest.

But you can argue that, at some point, a species can hit escape velocity. Symbolic communication is an incredibly powerful tool, and has compounding returns. A Darwinian alien looking at humans half a million years ago might have been able to guess the magnitude of our species' impact, but the specific things we do are too path-dependent.

Open Thread

• Drop in any links or comments of interest to Diff readers.

Diff Jobs

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

• Ex-Palantir team re-routing $500bn of home & auto insurance premium through AI assistants is hiring two roles: an orchestration engineer to coordinate agents across thousands of discrete problems, and a growth engineer to make them the default tool for insurance. If you’ve been building tools and infra for other companies but now want to be in the driving seat, this is for you. (NYC)
• Ex-Anduril, Ex-Abnormal Security, Ex-Bridgewater, fast growing startup building frontier agentic cybersecurity to make the long tail of businesses defense dominant via MSPs is looking for a senior machine learning / quantitative researcher. Working in low data environments or working in low signal-to-noise environments e.g, mid-frequency quant finance is a big plus. Evidence of data rigor required. (SF)
• 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)
• 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)
• 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. (Remote) 

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.