Longreads + Open Thread

Longreads

• Felix Stocker has a primer on ad economics in mobile games, and a fun analogy for how trends in venture create tailwinds for different sectors. Ad-supported games are a very interesting form of time arbitrage: they're short-term addictions that market themselves through the longer-term addiction of social media feeds. So they're basically converting "time spent x ad load x CPM" equation into another. One fun takeaway from this is that the better social media sites get at targeting ads, the worse mobile gaming companies will perform: they'll be competing more for the attention of prospective customers, and then competing with the same ads they use in order to monetize those users. It's surprising that the business has survived as long as it has.
• Astral Codex Ten on Hugo Chavez's rule-through-reality-TV regime. The book is full of surprises: Chavez certainly violated international norms, but he didn't steal elections much; voters seemed to genuinely like his bizarre antics. He did run the country terribly, but had a cheat code: Venezuela's vast oil wealth meant that there was plenty of subsidy money to go around. But, as many other governments have found, it's actually harder to build a competitive economy when there's oil money to be made—it sucks up the talent, pushes up the currency, and generally complicates policy enough that it's hard to get other things done. So Venezuela is a good case study in the resource curse, and in the general utility of being able to hit rock bottom: having money means not having serious consequences for your mistakes, but given enough time and effort, mistakes can cancel out even the most extreme natural resource wealth.
• Tyler Cowen writes about a change in how we think about the global economy: there won't be global growth trends in a de-globalizing world. When economies were tightly linked, growth in the US led to more demand for manufactured goods from China, which created more demand for raw materials from other parts of the developing world. But if that link is weaker, it's entirely possible for there to be a boom in some places and a bust elsewhere. That probably increases the personal returns from global macro investing while decreasing its social return: when the world is closely-linked, there are massive positive externalities in predicting recessions, because there are so few places to hide. It's comparatively less essential for the world to know that German is slowing down but growth in Indonesia is picking up, but it also means that macro questions are more tractable.
• Ben Thompson has a really good piece on OpenAI's keynote earlier this week. LLMs keep expanding the set of tasks for which they're the right tool for the job, and that's going to have interesting downstream effects—they're both a tool itself and an interface for other tools, so one way to model API companies is that they're a task-specific System Two whose biggest use case, long-term, will be as a fallback accessed via the System One of LLM chat.
• A Letter a Day has a transcript of a fantastic interview with investor Reece Duca, conducted by the CFO of Tegus. (Disclosure: Tegus advertises in The Diff, and I'm a user of the product—readers can get a trial.) Duca started with $2,000 in 1962 and has a lot more now (his stake in AppFolio is worth over $1bn), without raising outside capital other than some borrowing in the 70s. His strategy is the kind that's hard to execute with outside investors: concentrated long-term holdings, focused for the last few decades on just the software industry. One surprise is that many investors start out with extremely narrow specialties, and broaden their interests over time; Duca was originally a generalist, but decided that software had such good economics that it made sense to specialize in it exclusively. In the interview, he spends less time evangelizing his specific style and more time arguing that there are many, many ways to run a profitable investment business, only some of which require outside capital.
• And in this week's Capital Gains, we examine the economics of paying someone to tell you what you want to hear: some parts of consulting, accounting, and investment banking start with an expected answer and then pay an expert to repeat that answer back with an impressive brand name and detailed powerpoint behind it. But this can still be useful, in a surprising number of cases.

Books

• Henry Kissinger and American Power: A Political Biography: I'm of a generation that grew up hearing about Henry Kissinger as somewhere between a complicated figure and a war criminal, so the biggest surprise in this book was how universally popular he once was—Kissinger had an 80% approval rating, and a running joke (rumor?) in the DC media world was that he was the "White House swinger." The book also highlights the cozy nature of the mid-century elite. When Kissinger threw his first big DC party, around 1971, it was at the home of a journalist. A very pre-Watergate event.
• The Fund: Ray Dalio, Bridgewater Associates, and the Unraveling of a Wall Street Legend: If you've ever read a profile of an investor and thought "this was fun, but I'd like fewer descriptions of trades and more anecdotes about people being berated in meetings until they're literally sobbing," then you're in luck. Crying was not explicitly part of the Bridgewater culture, but it ends up being the result of a system of aggressively interrogating every decision and aiming for perfect knowledge of people's actions and tendencies. There's a lot less about Bridgewater's track record, which is too bad; it's only in the closing pages of the book that the author notes that Bridgewater is, by one measure, the most profitable hedge fund of all time. Still a good look at how a corporate culture can evolve from distinctive to pathological. For much more, there's a full Diff review for paying subscribers ($).
• Eighteen Days in October: The Yom Kippur War and How It Created the Modern Middle East: recent events have reminded me that I don't have enough context to understand what's going on in the Middle East. Even asking how far back in history you reach for context is a contentious, politicized question! So I picked a more recent conflict at random. The war described is a conventional war, i.e. there are uniformed combatants, both sides have planes and tanks, etc. And it's an absolutely astounding one, full of people overcoming hundred-to-one odds, getting astounding lucky breaks, sometimes making catastrophic mistakes despite their lucky breaks. (For example, Israel had an informant, Ashraf Marwan, who was a close aide to Egyptian President Anwar Sadat. Despite very high-quality intelligence indicating an imminent attack, Israel was unprepared, partly from bad luck and partly because of uncooperative intelligence services.) There are some vivid, too-implausible-for-an-action-movie-to-get-away with combat scenes, and a look at some high-level discussions. Among other things, the book set and then beat the record for highest one-day nicotine consumption I've ever seen referenced: Nasser smoked six packs of cigarettes a day, and Golda Meir's peak was nine.

Diff Jobs

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

• The leading provider of advanced options analytics — “the ASML of options trading” — is growing rapidly, very profitable, and looking for a generalist who can excel in chief of staff and business development functions. A trading, quant, or similarly technical background is a big plus. (Connecticut, NYC)
• A startup building a new financial market within a multi-trillion dollar asset class is looking for generalists with banking and legal experience. (US, Remote)
• A data consultancy is looking for a senior data scientist with prior experience in marketing data science and e-commerce. (NYC)
• A crypto proprietary trading firm is actively seeking systematic-oriented traders with crypto experience—ideally someone with experience across a variety of exchanges and tokens. (Remote)
• A fund that backs growth companies in public and private markets  is looking for a senior engineer with experience building data pipelines. Experience with data science and/or experience developing frontends to display the results of models is a plus (SF).

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.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• From Twitter: what’s the most recent date for which there’s no historical record of anything happening. If you go back far enough, we just don’t know much; unless there was an interesting astronomical event on November 11th, 2023 B.C., we probably have no record whatsoever of anything that happened on that date. But there are plenty of things we know happened today! What’s the cutoff? (I’m counting any document with a date or a reference to one, so letters, diaries, financial records, etc. all count.)