The Diff - Longreads + Open Thread

Longreads

• Steven Levy writes in Wired about the origins of "Attention is All You Need," and what's happened to the writers since. This story has been told in bits and pieces elsewhere, but it's nice to see it all at once. Two notable themes: first, even the people involved think of LLMs as being somewhat magical (and this has logical implications: Noam Shazeer is described as a "wizard.") Second, this piece is the strongest indirect endorsement of return-to-office that I've ever read—it's full of references to accidentally-overheard conversations, chats in the hallway, and the like.
• There is nothing new under the sun: Fortune profiles a high-growth tech company with great unit economics—long-term contracts with a high upfront cost, and rising margins thereafter, which has investors excited about growth but worried about valuation. The year is 1968, and the company is Ross Perot's EDS. Even the minor details are right: thanks to equity compensation, "Several of the top executives are multimillionaires now, and some of the lower-rank employees, including computer programmers in their twenties, are worth six figures."
• Dune is a fun story about what would happen if the main character in a novel read that novel's Wikipedia entry and decided that they hated the ending. The ecological and political world-building is great, but John Nye questions the economics: spice is simply too valuable to be treated so casually! On the other hand, we do have real-world examples of exactly this phenomenon—discovering oil is probably net bad for the median country, and only a survivable curse if the country already has significant non-oil economic activity (the US and UK) or copies the homework of a different country that discovered oil, like Norway ($, FT).
• Barry Ritholz interviews David Snyderman of Magnetar Capital. Given their size, Magnetar doesn't get that much media attention, but it's a fascinating company. Every financial services company tries to be a tech company, but Snyderman is the first one I've read who waxes eloquently about data pipelines, and how fun it is to ingest 40GB of transaction data in four minutes.
• John Cochrane has a wonderful piece from a few years ago trying to reconcile financial theory with observed behavior. One of the running themes in this piece is that academic finance ends up making some odd assumptions about market participants: we have to assume that people are seeking to generate excess returns by taking on certain risks, but we also assume that there are eager counterparties willing to take the other side of the bet. We also tend to treat paper gains as less meaningful than other kinds of income, and to worry about paper losses, even though these are mostly driven by valuation rather than by dividends. It's an admirable piece especially because it elegantly threads the needle between disregarding theory entirely and just assuming that everyone else is being unreasonable.
• In Capital Gains this week, we covered the cadence of alpha, and how long-term compounding is often the result of slowly deploying and refining a one-time insight.
• And in The Riff: e-commerce rollups, platform incentives, and why content moderation never ends. Listen with Twitter/Spotify/Apple/Substack.

Books

The Trading Game: A Confession: Trading memoirs are hit-or-miss. There are two main themes you can write about: "Here's how I made so much money," or "Here's how I ruined my life." The latter usually takes the form of substance abuse, but The Trading Game spends a lot more time witnessing that kind of behavior than talking about it. Instead, the author's life-ruining is a more introspective one—you can think of it as the story of a man figuring out that his mental downward spiral started much earlier than he realized.

But first, the optimistic part: Gary Stevenson is born into a poor family, runs into a lot of trouble at school—his first experience making a market involves drugs, and authorities are not pleased. But he's smart, tests well, gets into a better school and then wins a trading competition that results in an internship, which later turns into a full-time job.

Wall Street types often get called arrogant, and often for good reason. But given this context, it's astounding how humble he is: in his twenties, he has interactions with his boss that basically take this form: "Thank you for all of your hard work over the last twelve months. Here's a bonus that is more money than your parents have ever earned in their entire lifetimes. But don't worry—it's just because you're extremely smart compared to everybody else!" It takes a lot of mental fortitude to stay even remotely normal in a situation like that.

Stevenson's money partly comes from the usual right-place-right-time sort of thing; he was working for a US investment bank in Europe during the financial crisis, and was basically getting paid to provide dollar liquidity to the European financial system, something that Europe was willing to pay dearly for. That's a brief explanation of why a trader in his seat would have done well at the time, but Stevenson did quite a bit better. His main job was making a market, i.e. buying from whoever wanted to sell and selling to whoever wanted to buy. But he could keep some positions (especially if he made money at it). Stevenson's macro observation was that after the financial crisis, the rich recovered faster than the rest—but rich people have a low marginal propensity to spend, so that didn’t translate into higher consumption or faster growth. As a result, the market persistently assumed that growth would pick up and rates would as a consequence. As long as he kept some bets that interest rates wouldn't rise, and they didn't, he'd prosper. And, in fact, he prospered mightily.

Which led to what happened next. On average, it feels better to make money from good things than bad things. Short sellers in equities can take comfort in the fact that they're making the market more efficient, drawing attention to scams, etc. But in macro, if you make your money betting that rates will stay low forever, you are quite directly profiting from human misery. You might be contributing, in some tiny marginal way, to providing price signals that tell policymakers that whatever they're doing isn't working, but in one of the world's most liquid markets, those signals will be invisible. (The LIBOR scandal was part of a deliberate effort to manipulate rates, but these manipulations usually meant moving rates by a basis point, for a day, and not necessarily successfully. And that’s for a survey-based metric, not a market price-based one.)

So, he spends the last half or so of the book in an increasingly bad mental state—unable or unwilling to work, similarly unable to enjoy his money (he talks about living in an apartment that wasn't just unfurnished but unfinished, with bare concrete floors. (Further reducing the median rich person's marginal propensity to consume!) By the end, it's unclear whether he truly quits or gets fired—both sides seem to reverse their decisions a lot, revealing that there is a mindset that can make one person make big decisions with the same pace and inconsistency of a big bureaucracy.

Stylistically, it's quite good: I've never seen someone described as "a huge, lumbering presence with a body the shape of a muffin, who walked like a farmer falling over." At one point, he muses that the site of his last job assignment "must be the room that you wait in when you're going to hell but there's been some sort of administrative delay." So, as a trading book, a book about the world's most brutal case of burnout, and a well-written look at what life on a trading floor was like in the early 2010s, it's excellent. (I just want to know if he kept the "low rates forever" trade on.)

Open Thread

Diff Jobs

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

• A seed-stage startup is using blockchains to enforce commitments and is in need of a fullstack developer with Solidity experience. (Remote)
• A CRM-ingesting startup is on-boarding customers to its LLM-powered sales software, and is in need of a product engineer with a track record of building on their own. (NYC)
• A concentrated crossover fund is looking for an experienced full stack software engineer to help develop and maintain internal applications to improve investment decision-making and external applications to enable portfolio companies. (SF)
• A private credit fund denominated in Bitcoin needs a credit analyst that can negotiate derivatives pricing. Experience with low-risk crypto lending preferred (i.e. to large miners, prop-trading firms in safe jurisdictions). (Remote)
• A seed-stage startup is helping homebuyers assume the homeseller’s low-rate mortgage, and is in need of a product designer and a product manager. (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.