Longreads + Open Thread

Longreads

• The brilliant science fiction author Ted Chiang recently wrote a not-so-brilliant piece arguing against AI consciousness. Henry Shevlin has a thoughtful rebuttal. Science fiction seems like good practice for encountering alien forms of intelligence that may or may not be conscious, but writing fiction also gives authors a nice trapdoor: if there's some problem they either struggle with or don't find very interesting, they can simply skip it and write about something else. But as models get smarter, we have to think more rigorously about what makes us smart in the first place. There are still plenty of unanswered questions, and the smart-money bet is that current models are not conscious, and certainly aren't in any way that we'd recognize. But the more complex they get, the more we have to avoid some kind of substrate-based bigotry that supposes that what happens in carbon can never happen in silicon; so far, all the silicon-based intelligence we have is the result of carbon-based intelligence figuring things out. So we're brute-forcing our way towards either identifying the things artificial intelligence can't possibly do or finding out that there aren't any.
• Sebastian Garren on how the medieval church slowly decided to accept earning interest. Banning usury is one of those ideas that you can only endorse in an insulting way: it was probably a pretty good idea for illiterate medieval peasants to be unable to agree to whatever terms lenders offered them, because those peasants didn't have much collateral or context. If the market only clears in pathological ways, it's not a great market. In the case of lending, that's a function of how many economic opportunities there are, and how stable property rights are. This is one of those stories that combines broad economic trends (gradual technological improvement, a shortage of labor relative to capital after the bubonic plague) and more contingent factors (Pope Leo X, who wrote a papal bull allowing interest in some cases, was a also a member of the Medici banking family).
• Tyler Cowen interviews Roblox cofounder and CEO Dave Baszucki. Sometimes, when software nerds dream of the ideal company, they imagine a cool platform where they'd handle all of the abstract problems and then give users the tools to make whatever concrete thing they wanted. It turns out that there's almost never demand for this, but Roblox is an exception: the company is a constantly, relentless effort to stay out of the way of users who create their own unique experiences on the site, which Roblox monetizes by charging money to convert their internal currency back into the kind of money you can use to buy things in the real world and pay taxes. He has some interesting thoughts on how to build a system where users can communicate with one another, but where young users  aren't being constantly pestered or preyed upon, which is a tricky problem to solve. But if you're going to run a platform where you can sit back and let other people figure out how to monetize it, you'll have to put in a lot of work.
• Gary Sernovitz is a fun and pugnacious financial writer whose readership overlaps heavily with that of The Diff (I have received a literal slap on the wrist for not reading The Counting House). His latest is a profile of Citadel's Ken Griffin. Griffin is kind of hard to write about because Citadel is one of those organizations that isn't special for one particular thing, but for doing dozens or hundreds of things better than anyone else. One of the best little tidbits is that Griffin has been like that for a long time: he's one of the very rare people who used AP credits to graduate from Harvard in three years, which is the kind of bloodless rationality you'd expect. There isn't really a grand theory to Citadel, other than that the more uncorrelated return streams you put together, the bigger your asset and fee base can be. If someone's a master of details, but all of those details are proprietary, there just isn't much for him to say in an interview.
• Matt Welch in Reason on how most of the founding fathers died worried that America was falling apart. There's a cycle here: a sufficiently big crisis (the Revolutionary War, the Civil War, the Great Depression and Second World War) creates a cohort of elites who all trust each other and have some common idea of what success looks like because they've lived through it. But many of them will live long enough to see that social trust fray, and for the grubbier world of politics to intrude. Given how costly those episodes of high trust are, it is, in a way, a privilege to end your life thinking that everything you worked for is falling apart—it means there hasn't been a serious threat to the social order you and your peers helped establish.
• In Capital Gains this week: the most important thing missing from wealth inequality discussions is volatility drag, the phenomenon where concentrated portfolios with high arithmetic-average returns have poor compounded returns and, over time, tend to go to zero. If you look at the mean net worth of the mean rich person, you see a relentless advance, but it's caused by the same force that means that if you look at the median net worth of someone who was rich a decade ago, you see a surprising number of people who missed the bull market or even went broke.
• A Read.Haus user asks why prediction market people don't talk about catastrophe bonds. Cat bonds are a real market, and a bigger one than prediction markets, but they cater to a completely different kind of investor, in a revealing way. The counterparties for a cat bond are, typically, an insurer who wants to hedge out the risk of some singularly bad natural disaster, and a buyer who wants to add one more source of uncorrelated returns. Natural disasters tend to have weak macro effects, so they tend to be a source of uncorrelated returns. But they represent an instance of hedging rather than gambling, and, for now, prediction markets are small enough that it would be hard for someone to arbitrage between them. As prediction markets grow, we might reach the point where people hedge cat bond positions by making Kalshi bets about hurricanes, but for now the issuers of cat bonds are getting the hedge they want, and the buyers are getting the idiosyncratic return they want, and neither side needs the extra benefits prediction markets could offer.

Books

Genie out of the Bottle: World Oil since 1970: Global oil production is conventionally measured in barrels per day; conveniently for anyone who wants to do mental math, if you add in condensates, natural gas liquids, and biofuels, you get to about 100m barrels per day, 80-85m of which are crude. Many barrels per day of ink are spilled trying to figure out what oil is worth, both in practical terms of being long or short, and in terms of understanding how resource economies work.

Morris Adelman, author of this book (which is, in effect, a history of world oil from 1970 to 1995), had a bold proposition: what if oil is just like every other resource, and every other product, and we're all wildly overthinking it? Adelman lived long enough that there were multiple wars over oil during his lifetime, OPEC induced economic chaos among oil-consuming nations, and numerous booms and busts.

One natural way to think about oil history through the 1970s is that for a while, mostly-American oil monopolies abused their position, particularly at the expense of developing countries. They paid artificially low prices, earned profits as a cozy weak cartel, and that led to overconsumption, which ultimately made the world's oil consumers vulnerable to oil producing nations raising the price to something that reflected economic reality, which they ultimately did.

Adelman's view is: this is nonsense! Aramco was earning insane returns on equity, and the Arab world's oil production as a share of reserves was far below what it was elsewhere. The long era of oil prices being flat in nominal terms and declining in real terms was actually an era of overpriced oil, not because oil companies were running their own imperialist foreign policy, but because of prosaic domestic policy concerns: The American oil producers had high costs compared to Aramco, but those domestic producers also represented more votes—and were very generous political donors besides. In this view, OPEC's price increases were a demonstration that short-term oil prices could be increased, but they didn't say anything about the long term. Eventually, if a minority of producers withhold production in order to raise the price, they're encouraging the rest of the world's oil producers to step in and fill the gap, and they're also risking defection from their own cartel. And anyway, as Adelman notes, some of the oil price increases during OPEC's production cuts were the result of oil buyers accumulating more oil as insurance against future price increases, and paying temporarily high prices to maintain relationships that they expected to monetize at lower prices. The economic substance of these transactions was not that oil had been underpriced by an order of magnitude, but that even at elevated prices, certainty of supply was worth paying for.

The idea that oil was a normal commodity whose production would return to some equilibrium level was a lonely position when Adelman was writing in the mid-90s. Even though oil prices were ticking down, we were still in a post-OPEC regime. But he held on; after 65 years with MIT, he died in June 2014—a month in which oil production was 60% higher than it had been a decade earlier, due to fracking. He was, eventually, correct: there are a lot of ways that one country or one bad actor can make oil prices higher tomorrow, and very few ways that anyone can do anything that affects what oil will be worth in a decade.

Open Thread

• Drop in any links or comments of interest to Diff readers.
• Are there any other good contrarian-but-vindicated books like this?

Diff Jobs

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

• Ex-Palantir, Citadel founders building the meta-harness (the system that knows what hills to climb, and what the right loss functions are) for all the lucrative agents need full-stack engineers that understand that turning AI into economically valuable solutions means a system that includes deterministic infrastructure. If you’re curious about building a system that finds what the efficient frontier between determinism and stochasticity actually is, this is for you.  (NYC)
• Well-funded, frontier AI neolab working on video pretraining and computer action models as the path to general intelligence is looking for researchers who are excited about creating machines that learn from experience, not text. Ideally you have zero-to-one pre-training experience and/or are a high-slope generalist who’s frustrated that the big labs aren't doing this. (SF)
• Series A startup building multi-agent simulations to predict the behavior of hard to sample human populations is looking for researchers and engineers (ML, platform, infrastructure, etc.) to improve simulation fidelity and scale the platform to hundreds of millions of simulation requests. Problem-solving and genuine interest in simulation matter more than pedigree. Experience working with languages with an algebraic type system is a plus. (NYC)
• A Fortune 500 cybersecurity company with decades of proprietary security data is running an internal incubation with a pre-seed startup mentality and a mandate to build something new in AI. They are looking for a founding engineer who can ship fast, an engineer with a security background who’d be excited to contribute to OpenClaw’s security efforts, an AI researcher, and a generalist (ex-banking/consulting/PE background preferred) who wants to wear a bunch of different hats. Comp is FAANG+ and cash heavy. If you want to build something new in AI, but also need runway, this is for you. (SF/Peninsula) 
• High-growth startup building dev tools that help highly technical organizations autonomously test and debug complex codebases is looking for senior product managers who enjoy defining developer-facing APIs and abstractions. Experience with fuzzing or property-based testing a plus! (London, D.C.)
• 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. 

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.