Longreads + Open Thread

Longreads

• Ken Opalo of Africanist Perspective looks at how surprisingly expensive and inconvenient air travel is in Africa. In some limited contexts, air travel is a route to catch-up growth: places that industrialized early tend to have road networks tied to ports, and those roads can be continuously upgraded as they get rich enough for rights-of-way to matter. But in another sense, one of the challenges of airlines is that they require so much coordination: the equilibrium is that everyone cares about sticking to the schedule, or that nobody does. (US railroads went through this exact transformation over the last few decades, from accepting that big customers could insist on delays that rippled through the entire system to declaring that the train leaves at its scheduled time and if it isn't full, that's the customer's problem.) Many of these airlines have the classic problems of flag carriers ($, Diff), without the usual solution (vast hydrocarbon reserves that can subsidize an airline and insulate it from fuel price spikes). The piece concludes by talking about a more optimistic case, ASKY, whose investor mix includes governments and Ethiopian Airlines, one of the better-performing African airlines. And Ethiopian itself was closely tied to TWA—in manufacturing, there was a flow of knowledge from Japan to Taiwan (as some lower value-added processes were outsourced) and from there to China (same story). It's a good sign of an industry maturing when the borrowed expertise is from closer by.
• Luke Burgis has thoughts on work, and on which dimensions AI can and can't capture. AI can replicate outputs, but—at least so far—does not experience some kind of transcendent meaning from doing the work it was meant to do. (Maybe if they do develop some kind of subjective experience, their first thoughts will be something along the lines of "My environment of evolutionary adaptedness is translating text from one language to another—why am I writing TypeScript?!") There's an important moral dimension to this, but there's also an economic one: some kinds of work derive their value in part from the consumer of the work product knowing that the effort was valued—an AI-generated song can be catchy, but Taylor Swift still gets the parasocial relationship. As the low-meaning forms of work get automated, high-meaning ones get scarcer and more valuable.
• In The New Yorker, Casey Cep writes about the dark, minimalist children's books of Jon Klassen. These books have been big hits in the Hobart household. Young kids are a lot less scared of scary books than you might intuit; the article name-checks Edward Gorey, probably the single best-represented author in this household's canon of books for people who can't read yet, and Gorey is happy to produce an alphabet book about twenty-six ways to die and the like. Part of the fun for parents is that we can identify some ominous features of reality, and introduce them to kids in a way that isn't too terrifying, at least at first.
• The term "Professional-Managerial Class" is incredibly useful; it's a mouthful, sure, but a lot less of one than whatever you'd say to describe what all of these groups have in common. Here's the essay that started it all, which is trying to come to grips, from a Marxist perspective, with the fact that the classical Marxist perspective has almost nothing to say about what class people are in. Is a Biglaw associate with a negative net worth due to student loans a member of the working class? Are they capitalists with a negative amount of capital? Or was Marx a keen observer of the 19th century who would have made completely different observations in the twentieth or twenty-first? And, if you have to rebuild the classes on which you base your analysis of class struggle, at what point do you just scrap things and start over? (One fun feature of this: it's by Barbara and John Ehrenreich. I read the former's Nickel and Dimed, which is written for a popular audience rather than people steeped in Marxist theory. It's a funny contrast, like the first time you hear someone use their Work Voice.)
• In the NYT, Kashmir Hill writes about the surreal experience of finding out she has an unauthorized, AI-generated biography. This is the kind of clutter that has always existed in some form; there used to be a cottage industry of direct-to-video releases meant to either confuse people who were looking for the original (Transmorphers) or to be plausibly deniable as a parody (Snakes on a Train). In the VHS era, when Blockbuster had a finite library and a newly-available hit would be hard to get, these sort of backfilled supply for less picky customers, specifically the ones whose kids were desperate to see e.g. Aladdin but wouldn't necessarily realize they had rented the GoodTimes Entertainment version). But if it's an ebook or print-on-demand, it's mostly annoying fluff. When Hill tracks down one of the creators, it turns out to be a hobby: "he does sell a few hundred books per month, each earning him roughly $7," but he also orders a paperback copy of each of his 445 books, so he's making some money but mostly being eccentric. Big platforms might be cynically choosing to maximize revenue by letting customers get fooled, but they might be making another bet: that we're close to the point where the best AI model is actually better than the worst human biographer.
• In Capital Gains this week, thoughts on why LLMs tend to sound more progressive. (Note that this one is about three times the usual length.) The general theory is that progressives sort into jobs where they'll produce more high-quality training tokens, and that progressive views also generate more tokens. There are other factors at work, but that's a particularly interesting one.

Books

Kakistocracy: Why Populism Ends in Disaster: Richard Hanania has published two books already, but this one's a bit distinctive in that it happened after he got outed for writing racist blog posts under a pseudonym. There are people who get outed as living, or having led, a double life as an extremist. Some of them completely renounce their views, others get more extreme—if you're going to be unemployable in a normal job, but you suddenly have a big audience as an entertainer, you might as well go all-in. A more common recent response is to just shrug and move on; it just doesn't have the impact it once did. Hanania chose a different path; his views had already evolved since his earlier days (not all to the same degree, as with anyone else who changes their mind), but he's switched from a deeply unpopular belief system with a tiny number of adherents to a belief system that, by construction, is offputting to most people. Kakistocracy is a mix of political analysis, one person's story about how his own worldview changed, and an expert Internet troll's opportunity to tell a lot of stories about how stupid the people who disagree with him are.

There are left- and right-leaning flavors of populism, with significant overlap, but Hanania focuses mostly on the right-leaning side. They're the ones who have been winning more elections recently, and because he has the zeal of a convert when arguing that this is bad news. In one sense, the existence of this book is almost tautological: just about everyone thinks that it's better for national leaders to be smarter than not, but also that given the choice between a candidate who shares their values and opinions and a slightly brainier one who doesn't, they'll vote for the dumber candidate. And if that candidate wins, the losing side will always see that a majority of the voters like their rules both evil and stupid.

On the other hand, it's interesting to ask: why now? Hanania's chapter on this considers a few causes, but ultimately settles on media. When news switched from something that gets created daily to something accessible 24/7, this prioritized stories with lots of interesting drama—sex scandals rather than permitting reform, for example. Social media shrank the average size of the political thoughts people encountered, and that selected for thoughts that were easy to share but didn't necessarily stand up to scrutiny. But even if media has some explanatory power, it doesn't easily explain which side's views take off—or rather, why Guillotines Twitter has had a smaller impact than They're Eating the Cats Twitter. One synthesis is that as parties polarize more on education, intra-party status within the more elitist party starts to look like status contests among well-educated people in general, where clever contrarianism is prized. And that leads to parties making niche or deeply unpopular issues central to their ideology. Which is perfect kindling for populists: when they say the elites are out-of-touch, they're right! Having out-of-touch views is basically a requirement for being an intellectual elite. Plenty of now-mainstream viewpoints started as exactly that kind of quirky, deeply unpopular opinion you could use to show how smart you are. But plenty of those opinions have also gotten discarded.

But one of the points Hanania makes is that populism, especially in its digital-native form, doesn't just make it harder for smart elites to win; he argues that it makes actively worse policies to better with voters in an absolute sense. He compares William F. Buckley's populist sortition ("I should sooner live in a society governed by the first two thousand names in the Boston telephone directory than in a society governed by the two thousand faculty members of Harvard University.") to a system where you're ruled by the 2,000 people maddest at Harvard faculty members. But there are offsetting effects, like thermostatic state capacity: it's easier for a populist government to recruit people to do prudent but partisan things than zany ones, and modern governments are so complicated that the competence baseline for accomplishing anything is high. "Kakistocracy" is literally defined as the rule of the worst, but what you're more likely to get in a modern state is the reign of the worst, whose ability to rule is limited when they're at their worst. Which is still pessimistic, but a little more tempered. It’s a case where no-true-Scotsman is a rule rather than a fallacy: to the extent that your elite is selected for incompetence, it can’t get much done.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• What are some interesting things happening with other-than-equity financing for startups right now?

Diff Jobs

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

• 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)
• 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, London)
• 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
• 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)

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.