Longreads + Open Thread

Longreads

• Jane Psmith reviews Cambridge Latin Course Unit 1, in what is really a piece about learning Latin: should you? Why? And how? Latin is a fascinating language, in part because its historical role was like English, only more so: the majority of English speakers learned it as a second language, and for any given pair of people who don't share a common first language, English is the most likely default. At some points in European history, nobody spoke Latin as their first language, but it was a good default second language. Which means that Latin punches above its weight in two ways: first, as the language in which a surprising share of important documents were first written, and second, as a language that was preserved in the form of something that would be taught in a classroom. The book being reviewed teaches Latin through immersion, instead of by memorizing declensions and the like. As the piece points out, because Latin has existed as a second language for so long, we literally have workbooks for grade school-aged children from two millennia ago.
• Cliff Asness has a complaint about critiques of market timing, specifically the claim that if you missed the best N days or months of returns over some long period, your overall returns would be materially lower. It's correct, but as Asness notes, this really means "if you had an infallible short signal that perfectly predicted extreme events, you wouldn't want to run the inverse." It's definitely striking that the S&P's return since 1970 drops from 9.3% to 4.1% if you miss the twenty best days over that period, but also, very few people would. (Though one minor defense of this thinking—which doesn't apply to the target audience!—is that if you adjust your net exposure based on volatility, you will probably be underweight equities on the best days to be in the market, because most of the single best days in market history occur in the worst months and years.
• Dwarkesh Patel on how AI will take longer than people think to have an impact, because while models are smart, they don't learn on the job. This actually makes AI models quite similar to other technologies, prosaic and not: if you buy a dishwasher, that dishwasher is never going to get any more efficient, and will probably get a little bit worse over time. However, you will get more efficient at figuring out the right cadence for loading it and running it, and at optimizing the layout of dishes. So when you buy some new piece of capital that's a complement to your labor, part of what you get is some compounding in the value of those labor-hours, which may or may not top out eventually. That still leaves a lot of room for improvement, and a lot less room for economic disruption: if the one-off improvement comes from using a new tool and then continued improvements come from using it better, human labor gets more valuable for a while yet.
• In Vulture, a fun story that feels like the premise for a modern Oscar Wilde play, on how everyone in movies is using AI, but no one wants to admit it. This is another portrait of possible AI-induced post-scarcity—what if you eliminate 80% of your work hours, and spend the new free time browsing or watching surreal Veo3 videos, but your boss is smart enough to prompt his BossBot Agent not to ask whether or not you're using AI. It's basically keeping nominal wages and nominal-in-a-different-sense work hours sticky, with the increased productivity absorbed by leisure time. The best parts of this piece are the efforts to get away with it: "One animator who asked to remain anonymous described a costume designer generating concept images with AI, then hiring an illustrator to redraw them—cleaning the fingerprints, so to speak." At last, computers have freed us from the drudgery of creative work, leaving only intrinsically worthwhile tasks like copying somebody else's drawing onto a different piece of paper.
• In Conspicuous Cognition, which remains the single best publication on the topic of misinformation, thoughts on the crisis of expertise as a status problem. If you take the claim "experts have specialized knowledge which they share with the rest of us in order to help us make better decisions," you can rephrase it without changing its substance but radically changing its appeal by saying "the experts are a special class of people whose job is to tell you that you're wrong and correct your behavior." That's what they do! That's what they're for! You don't need dentists or physicists or infosec specialists if everybody naturally knows how to maintain dental hygiene, model the quantum interactions that lead to semiconductivity, or not click on shady links. But that also means that one way to signal expertise is to look down on people, which leads to some perverse incentives on both sides.
• In Capital Gains this week, some thoughts on the market for pre-IPO stock: why it used to not exist, the one-time force that made it happen, and why it persists today.
• Last week's newsletter linked to ReadHaus (disclosure: I'm an advisor), but the link was to a specific instance of a chat, rather than a general one, so I accidentally invited everyone to what amounted to a groupchat with a virtual me. Anyway, here's a link you can use to start a chat session with the Diff/Riff/Capital Gains/Medium etc. corpus

Open Thread

• Drop in any links or comments of interest to Diff readers.
• Which book should we do for FinTwit book club after The Snowball?

Diff Jobs

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

• A well-funded startup ($16M raised) that’s building the universal electronic cash system by taking stablecoin adoption from edge cases to the mainstream is looking for a senior full-stack engineer. (NYC, Singapore)
• YC-backed startup using AI to transform how companies quantify and optimize engineering productivity is hiring formidable full-stack and AI engineers. Experience with React + Typescript, Go, or Python on the ML side a plus. (SF)
• Well-funded, fast-moving team is looking for a full-stack engineer to help build the best AI powered video editor for marketers. Tackle advanced media pipelines, LLM applications, and more. TypeScript/React expertise required. (Austin, Remote)
• A Google Ventures-backed startup founded by SpaceX engineers that’s building data infrastructure and tooling for hardware companies is looking for a product manager with 3+ years experience building product at high-growth enterprise SaaS businesses. Technical background preferred. (LA, Hybrid)
• 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)

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.