Byrne Hobart writes The Diff, a newsletter about inflections in finance and technology with 24,000+ subscribers. His central preoccupation is understanding how people and institutions coordinate to solve complicated problems, and why that coordination has been breaking down in ways that contribute to productivity stagnation across both technology and social institutions.
The coordination problem as the root of stagnation
Many important problems cannot be solved by lone geniuses; they require institutional coordination, but institutions often have conflicting mandates—a public-facing mission and an internal, unstated goal. For example, SpaceX publicly aims to make humans interplanetary, but internally some may see that as a recruitment and fundraising tool while the real goal is long-term shareholder value maximization.
This double-talk is not necessarily sinister: when groups have intractable and divergent goals, some degree of misalignment between stated and actual objectives is inherent. Even within a single institution, different people may genuinely believe different things about what the organization is for.
A key driver of stagnation is that institutions often perpetuate themselves rather than solve the problems they were created for. If the problem keeps growing, the people working on it can stay important indefinitely—even if they fail.
Productivity stagnation and the regulatory bottleneck
Measured total factor productivity growth has been roughly 1% per year lower than its mid-20th century trend. Compounded over 50 years, this implies GDP per capita could be 60–70% higher today. The gap is not just an abstract number—it represents specific innovations that never got developed and deployed.
The book Where’s My Flying Car argues that flying cars are technically feasible and that regulatory tightening—not any specific ban—has made it prohibitively difficult to build and fly novel aircraft. The early history of steam engines shows that new technologies often involve dangerous failures at the start, but these are the learning costs that lead to mature, safe systems later.
The US is one of the few countries wealthy enough and with developed capital markets enough to support such experimentation, so strict regulation in the US and a handful of other rich countries matters disproportionately. Aviation in particular has extremely high barriers to entry: only Boeing and Airbus make wide-body commercial planes competitively, and China has struggled to catch up despite decades of trying.
Even when Western companies face regulatory barriers, they generally do not relocate experiments to less-regulated countries. There is a coordination between media and regulators: if you are seen deliberately circumventing FDA rules, the FDA will make it very difficult to ever sell the resulting product in the US. This enforcement incentive applies to large private institutions as well—Disney aggressively pursues knockoff characters, and game publishers pursue piracy.
One specific technology where regulatory barriers matter is embryo selection using polygenic scores. Testing for genetic diseases is relatively uncontroversial, but selecting for traits like extreme height would be very hard to get approved by the FDA. China’s government appears more open to genetic research and selection, so if such practices become legal anywhere, China is the likely location.
Patient longtermism and why compounding wealth across centuries rarely works
Robin Hanson has proposed “patient longtermism”: park money in a compounding fund and over centuries become the wealthiest entity in the world. In practice this almost never happens at scale.
The main reason is catastrophic drawdowns: war, famine, pestilence, and revolution periodically destroy wealth, especially financial assets that are easy to seize. Even in the US—which has been unusually free of such events—long-term trusts have failed. Adam Smith (the financial commentator) wrote about a 19th-century investor who bought reliable blue-chip stocks and locked them in a trust for generations; by the time heirs accessed it, the companies (including American Alarm Clock Manufacturing) had been bankrupt for decades.
There is some persistence of status across generations: a study of 15th-century Florentine wills found that last names associated with certain professions then are still associated with those professions today. But this persistence of status is not the same as persistence of compounded financial wealth.
Secret societies and hidden institutional power
Byrne is skeptical that many secret societies with over $1 billion in assets exist. Any organization that persists for a long time needs a legal structure to hold assets and a system for selecting leadership, both of which create paper trails.
Universities may be the closest modern analogue to secret societies: they were often founded by people with peculiar ideas about how to shape young, impressionable minds, and they contain internal secret societies (like Skull and Bones) that have become largely networking pipelines to investment banks.
Secret societies face a structural tension: they must keep secrets to maintain cohesion, but they must also reveal secrets to recruit and retain members. If the revealed truth differs from the recruitment narrative, the society can fracture. Al Qaeda is an example in reverse: many early members thought Bin Laden was a loser and were just scamming him for money, yet he eventually organized effective attacks.
Families are a more viable model for multi-generational coordination because parents can be honest with children and gradually reveal long-term plans. The Kennedy family is an example: Joseph Kennedy’s goal was to get one of his children elected president, and they accomplished it.
The broader evidence against powerful hidden forces is that world events rarely look like the product of clever coordination—they look more like random, incompetent forces with nobody in charge.
Optionality, conformity, and elite education
Byrne’s essay “Optionality for Cowards” argues that people increasingly treat optionality—keeping all paths open—as a goal in itself, but this is often a form of risk aversion disguised as rational planning. In financial markets, selling optionality is systematically more profitable than buying it, suggesting that in real life, decisions that forego options are undervalued because so many people are competing for maximum flexibility.
Elite higher education selects for risk aversion: getting into Harvard or Yale is so competitive and random at the margin that attendees feel pressure to stay on the safe path (consulting, banking, big tech) rather than risk a deviation. Students who get into a better-than-expected school feel they must stay on that exact path to maintain their perceived percentile ranking.
However, Byrne still views elite education as a good deal for individuals—just a bad deal for society because too many people pursue it. The real question is whether to drop out when a specific, time-sensitive opportunity arises. Both Zuckerberg and Gates are technically on hiatus from Harvard and could re-enroll, illustrating how far the preference for optionality extends.
Tyler Cowan’s The Complacent Class can be read as describing a society whose preference for optionality has grown: people delay starting businesses, delay having families, and generally avoid commitments that close off future paths.
The wealth effect explains much of this: in a subsistence society, you either work or starve, so there is little optionality. As societies get richer, the range of viable life paths expands, and people are wired to feel more satisfied by pursuing specific goals than by keeping options open—yet many still chase optionality at the fractal level (keeping multiple backup plans for every plan).
Conscientiousness versus introversion in the modern economy
Byrne disagrees with Tyler Cowan’s prediction that conscientiousness will become more valuable. He defines conscientiousness as the ability to fulfill externally-set goals, and argues that as communication costs fall and more of the economy is mediated through software, there will be more specialized jobs where people are intrinsically motivated—they would do the work for free.
In such environments, conscientiousness matters less because the person is self-motivated by genuine interest. The counter-argument is that entertainment is becoming more addictive and better at marketing itself, so the ability to resist distraction (a form of conscientiousness) is becoming more valuable.
Cowan also argues that in a world of global supply chains, an unreliable or volatile person can cause far more damage than in the past. Byrne is skeptical of this, noting that people who manage complex supply chains (like Apple’s response to Shenzhen shutdowns during COVID) often find the work intrinsically gratifying—similar to the video game Factorio, which is entirely about managing complex supply chains through crises.
Rationality, prediction, and influence
Byrne describes investing as a “rationality dojo” because it requires modeling other people’s behavior and identifying why the other side of a trade is wrong—often due to institutional or regulatory constraints rather than factual errors. For example, small-cap stocks are less efficiently priced because many institutions have size cutoffs below which they will not invest analyst time.
The rationalist subculture skews more toward programming than finance, possibly because the rationalist mindset—identifying flaws one by one and iterating—aligns more naturally with debugging code. Finance is the “extrovert rationality dojo” (modeling other people), while debugging is the “introvert rationality dojo” (outsmarting yourself).
Rationalists predicted COVID-19 early because they took ideas seriously and were willing to extrapolate from early data (R-naught above 2%, case fatality rate around 2%, Chinese New Year travel patterns). They were not making big inferential leaps—just compounding numbers.
However, the same willingness to take ideas seriously means rationalists also hold beliefs that sound absurd to outsiders (simulation theory, cryonics, Bitcoin saving the world). This creates a credibility problem: anyone warning about COVID could be immediately associated with other “crazy” rationalist beliefs, making it harder to earn political capital and influence outcomes. Rationalists are good at predicting the future but comparatively worse at changing it because earning influence often requires counter-rational social decisions.
Reading, knowledge, and the convexity of learning
Byrne argues that reading is underappreciated relative to video because books have a lower production cost, which means there is a long tail of niche content that is worth writing as a book but not worth producing as a documentary. A book requires one person typing; a documentary requires a crew, equipment, travel, and editing, all compressed into an hour of content.
The writing process for a book happens primarily during reading: the author reads extensively, highlights connections, and then applies a coherent narrative when typing. In a documentary, the production process can take more time than the underlying research.
Podcasts are a different medium entirely—they are dialogues rather than lectures, creating broader connections between ideas through back-and-forth, but sacrificing depth compared to books.
Knowledge has increasing returns (convexity): each new thing you learn makes more sense in the context of what you already know, and it makes the existing knowledge more memorable. Paul Graham described learning history as initially being a series of disconnected dates and names, but as you read multiple books on the same period (political history, biography, technological history), connections emerge and everything becomes more memorable.
The convex returns to knowledge mean that an expert reading a new book may derive more total value from it than a novice, even though the novice has more to gain in absolute terms. The expert can have “dialogues” with authors—noticing when an author missed something the expert knows, which makes the entire reading experience more memorable.
The middle income trap and career strategy
Byrne’s essay on the middle income trap compares countries that get stuck at middle-income levels to careers that plateau. Countries escape by either building branded products that command premium prices or developing indigenous technologies that cannot be matched elsewhere.
Japan’s escape involved both: temporary competitive advantages in automobiles (where Japanese manufacturers were superior to American ones for a period) and persistent monopolies on intermediate industrial goods in semiconductor supply chains—specific steps in the chip fabrication process that only one or two Japanese companies can perform.
The career analogue is to develop skills adjacent to your current work that you are initially bad at, so you can intelligently manage people who do those tasks and ask better questions. The goal is to build a set of skills that nobody can match, allowing you to compete on something other than price.
Competing on price works when you are young and your cost of living is low, but you cannot get rich by being cheaper than average forever. At some point you must differentiate.
Advice for young people
Because books are cheap and the internet is a powerful distribution mechanism, a determined person can become close to one of the world’s leading experts on a narrow topic in a fairly short time. The topic must be narrow enough that no one has written a dissertation on it, or alternatively, you can become the world’s public-facing expert by reading academic papers, following their citations, and synthesizing the material into an accessible format.
This exercise is valuable both as a demonstrable body of work (showing you can learn, synthesize, and teach) and as a way to internalize how complicated the world really is. The only way to truly learn this is to repeatedly discover that you were wrong about something because you oversimplified it—and ideally to find that the first source you read had a theory that was later overturned.
When studying a topic, read a smattering of different books taking different angles: contemporary accounts, biographies, fiction from the period, economic history, and technical analyses. Fiction in particular captures tangible details that non-fiction misses and adds texture.
History is easier to study because outcomes are known, but contemporary topics have more data. The right attitude toward current events is to read with a pseudo-historian’s perspective: ask what people will think about this in 50 or 100 years, what the main schools of thought will be, and what their strengths and weaknesses will look like.
