THE SCENARICA SUNDAY

In 2000, a neuroscientist named Eleanor Maguire slid London taxi drivers into an MRI scanner at University College London and found something that should have changed how we think about knowledge. The drivers who had passed “The Knowledge”, the legendary exam requiring memorisation of 25,000 streets, had measurably larger posterior hippocampi than a matched control group. The difference was not metaphorical. It was anatomical. The longer a driver had spent navigating London from memory, the larger the relevant brain region had grown. The tissue had physically rebuilt itself to hold what the driver had learned.

The Knowledge is not an ordinary exam. Candidates spend three to four years riding motorcycles through the city, drilling themselves on the fastest route between any two of 25,000 points. The examiner can ask any pair. The candidate must answer without hesitation. Most who begin do not finish. What Maguire’s scans revealed was that the survivors of this process had not merely memorised a city. They had remodelled a brain. The act of effortful navigation was not just using the hippocampus. It was growing it.

Then GPS arrived. And the studies that followed found exactly what Maguire’s work predicted. Habitual GPS users showed measurably worse spatial memory during self-guided navigation. In a three-year follow-up, heavier GPS use predicted steeper decline in hippocampal-dependent spatial memory over time. The drivers who used satellite navigation could still get from A to B. Their competence was intact. But the competence was borrowed. It lived in the device, not the brain. Remove the device, and the knowledge had never been there.

In March 2026, the workplace analytics firm ActivTrak published an analysis of 443 million hours of work activity across 1,111 companies. The finding contradicted the dominant narrative about artificial intelligence so directly that Fortune ran it as a corrective: AI users were not working less. They were working more.

Among 10,584 workers tracked for 180 days before and after AI adoption, time spent on email had risen 104 percent. Messaging had increased 145 percent. Business management tasks had climbed 94 percent. No activity category decreased after adoption. The tools designed to reduce workload had, by every measurable indicator, increased it. But the nature of the increase mattered more than its size. Workers were not spending more time writing. They were spending more time reviewing, editing, and managing the output their tools had drafted. The shift was categorical: from production to supervision, from generating to selecting, from building to curating.

Separately, Microsoft’s Work Trend Index reported that 58 percent of AI users now produce work they could not have completed a year earlier. The number sounds like progress. Read it again and it sounds like something else. Fifty-eight percent of workers regularly deliver output that exceeds their own understanding. The gap between what they produce and what they comprehend is not evidence of growing skill. It is a measure of growing dependency.

Between 2017 and 2025, the number of AI researchers choosing to move to the United States fell 89 percent. The decline accelerated sharply, dropping 80 percent in the most recent twelve-month period alone. Stanford’s AI Index published the figure in April 2026 alongside a finding that would have been unthinkable five years earlier: China now produces more highly cited AI research papers than the United States, 20.6 percent of the global total to America’s 12.6 percent.

The talent exodus matters because of what it reveals about the relationship between capability and comprehension at the national scale. The United States remains the world’s largest deployer of AI systems. American companies lead in investment, infrastructure, and commercial application. But the researchers who build the foundational models, who understand the mathematics beneath the interface, who can diagnose a novel failure rather than follow a governance checklist, are increasingly choosing not to come. A country can be the world’s largest consumer of a technology and simultaneously lose the ability to understand what it consumes.

These three patterns, a taxi driver’s shrinking hippocampus, a knowledge worker’s expanding inbox, a nation’s emptying research labs, look like different problems in different domains. They are the same problem. The mechanism that connects them has been documented in cognitive science for nearly fifty years. It has a name.

In 1978, the psychologists Norman Slamecka and Peter Graf ran a series of experiments in which subjects either generated words themselves or simply read them. The finding was stark and has survived decades of replication across hundreds of studies. Self-generated information is retained dramatically better than passively received information. The act of producing an answer is not merely evidence that understanding exists. It is the process by which understanding is created. Effortful generation builds neural pathways that passive reception does not. This is why writing an essay teaches you more than reading one. Why solving a problem teaches you more than reviewing the solution. Why navigating by memory grows the hippocampus in ways that following GPS cannot.

Slamecka and Graf called it the generation effect. And when AI generates and humans curate, the generation effect transfers from mind to machine. The human retains the ability to recognise good output without retaining the ability to produce it. The skill of judgment remains. The skill of creation quietly atrophies. Each delegation is individually rational, individually efficient, individually productive. The aggregate is a system that performs beyond its own comprehension.

There is a word for a system that performs beyond its own comprehension: brittle.

The brittleness is now measurable. A meta-analysis published in Nature Human Behaviour in 2024 examined 106 experimental studies involving human-AI collaboration. The finding cut against the industry’s founding promise: on average, human-AI combinations performed significantly worse than the best of either working alone. When AI outperformed humans, adding a human to the process degraded the result. We have built the most powerful cognitive tools in history and arranged our collaboration with them so that they make us collectively less capable than either party working independently.

In education, where the generation effect matters most because the entire purpose of learning is the effortful creation of understanding, the transfer is nearly complete. Ninety-four percent of UK undergraduates now use generative AI for assessed academic work. The proportion directly including AI-generated text in their assessments has quadrupled in two years, from 3 percent in 2024 to 12 percent in 2026. But the statistic understates the structural shift. The assessment was never the product. The thinking was the product. The assessment was the evidence that the thinking had occurred. A student who has never written an essay without AI assistance has never experienced the cognitive process that essay-writing was designed to produce. The student who generates the argument, struggles with the structure, revises the phrasing, and resolves the contradictions understands the material. The student who prompts, selects, and polishes has learned to recognise quality. Not to produce it.

Institutions are beginning to sense the corruption without being able to name it. Gartner predicted that by the end of 2026, half of all organisations globally would require AI-free skills assessments. The prediction is a policy admission that the signal has been corrupted. When every employee produces competent work with AI assistance, competence ceases to distinguish between people who understand what they produce and people who merely recognise quality when a machine produces it. The organisation can no longer tell what it actually knows.

But the institutional response reveals its own paradox. Companies with active AI governance frameworks move twelve times more AI projects from pilot to production than their ungoverned peers, according to Databricks’ analysis of over 20,000 organisations. Governance accelerates deployment. It does not prevent the competence trap. It deepens it. A well-governed organisation deploys more AI, more quickly, across more workflows, into more decisions. Each deployment shifts another cognitive task from generation to curation. The governance framework manages the failure modes it was designed to anticipate. The competence trap’s danger is the failure mode nobody anticipated, the one that requires the human understanding that twelve times more deployment has been quietly eroding.

Aviation found this trap in wreckage. When Air France Flight 447 fell into the Atlantic in June 2009, killing all 228 people aboard, investigators discovered that the pilots, confronted with a sudden failure of the automated flight system, had responded with inputs that deepened the stall rather than correcting it. The automation had performed the task for so long that the humans no longer possessed the skill the automation was supposed to supplement. The FAA now urges pilots to periodically hand-fly entire departure and arrival routes, a deliberate return to effortful generation designed to maintain the very capability that the system’s efficiency erodes.

The pattern is identical at every scale. A hippocampus shrinks when GPS does the navigating. A worker’s comprehension atrophies when AI does the generating. A nation’s strategic understanding hollows out when the researchers who build foundations leave and the engineers who deploy products stay. Rising output. Declining understanding. The generation effect does not care whether the unit of analysis is a brain, an organisation, or a country.

The trap is invisible because the system works. Glass does not look brittle. It holds weight beautifully, performs under every load it has been tested against, and gives no warning before it shatters. A workforce that produces excellent output through AI augmentation looks like a high-performing workforce. It will continue to look like one until the tool is unavailable, until the failure mode is novel, until the situation demands the understanding that the tool was supposed to supplement but instead replaced.

The question is not whether AI makes us more productive. It does. The competence trap is not a story about a technology that fails. It is a story about a technology that succeeds so completely that the humans who use it are gradually absolved of the cognitive effort that made them worth augmenting in the first place. And the question nobody can yet answer, not the governance frameworks, not the workplace analytics, not the AI indices, is whether productivity purchased at the cost of comprehension is a bargain we can afford, or a debt we are running up against a future that will, as futures always do, eventually present the bill.

Sources:

Maguire et al., Navigation-Related Structural Change in the Hippocampi of Cab Drivers, Proceedings of the National Academy of Sciences, 2000

Slamecka and Graf, The Generation Effect: Delineation of a Phenomenon, Journal of Experimental Psychology: Human Learning and Memory, 1978

Vaccaro, Almaatouq, and Malone, When Combinations of Humans and AI Are Useful: A Systematic Review and Meta-Analysis, Nature Human Behaviour, 2024

ActivTrak, 2026 State of the Workplace, March 2026

Microsoft, Work Trend Index 2026

Stanford HAI, 2026 AI Index Report, April 2026

HEPI, Student Generative AI Survey 2026

Gartner, Top Predictions for IT Organizations and Users in 2026 and Beyond, October 2025

Databricks, 2026 State of AI Agents

Bureau d’Enquetes et d’Analyses, Final Report on the Accident on 1st June 2009 to the Airbus A330-203, Flight AF 447, 2012

“The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com”

“Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.”

THE SCENARICA SUNDAY

In January, a secondhand electronics marketplace called Back Market reported that sales of discontinued iPods had jumped forty-eight percent in a single year. The buyers were not nostalgic middle-aged collectors. They were twenty-somethings paying premium prices for devices Apple had killed in 2022, devices that do less than any smartphone on the market. The iPod cannot stream. It cannot scroll. It cannot notify. That is the entire point. What these buyers wanted was a machine whose only function was to play music they had chosen in advance, on a screen too small to distract them, with no algorithm suggesting what to listen to next. They wanted friction. And they were willing to pay for it.

This is the defining consumer instinct of 2026, and it has a name. In January, the columnist Kathryn Jezer-Morton published a piece in The Cut titled “In 2026, We Are Friction-Maxxing.” The essay argued that technology companies had systematically trained us to experience ordinary human moments, boredom, awkward silences, the effort of remembering a fact we could look up, as problems requiring solutions. Friction-maxxing is the deliberate reversal: choosing the slower, harder, less optimised path. Flip phones instead of iPhones. Handwritten letters instead of texts. Cooking from scratch instead of ordering through an app. Wired earphones instead of AirPods. The essay went viral. Within weeks, the Financial Times, Harper’s Bazaar, Architectural Digest, Forbes, and The Guardian had all published their own friction-maxxing pieces. Wikipedia now has a dedicated entry. The movement has a name, a following, and, increasingly, a market.

The case for it is compelling, and it is not trivial. When Jezer-Morton defined friction-maxxing, she was careful to frame it not as anti-technology puritanism but as a recalibration: “building up tolerance for inconvenience (which is usually not inconvenience at all but just the vagaries of being a person living with other people in spaces that are impossible to completely control).” There is real evidence behind this impulse. A Microsoft and Carnegie Mellon University study found that workers who trusted generative AI tools exercised less independent critical thinking. The psychological literature on the IKEA effect demonstrates that people value outcomes more when they have invested effort in producing them. Gianpiero Petriglieri, a professor of organisational behaviour at INSEAD, told the Financial Times that his students worry convenience will prevent them developing judgment: “A little bit like a controlling parent, it makes life easier for you, then you don’t know how to manage life on your own.” These are not the anxieties of people who misunderstand technology. They are the anxieties of people who understand exactly what they are gaining and suspect what they are losing.

But there is a question that nobody promoting friction-maxxing has paused long enough to answer: who, exactly, can afford to make their life harder?

Andre Spicer, the executive dean of Bayes Business School, raised this point with the Financial Times in language that was polite enough to be devastating. “We often find people use friction as a way of increasing the difficulty and inconvenience of a task,” he said, “to create status around it.” The observation sits uncomfortably next to the movement’s self-image. The person who swaps their iPhone for a flip phone is almost certainly someone whose job does not require them to be reachable by app at all times. The person who cooks every meal from scratch is almost certainly someone whose schedule contains the hours that cooking demands. The person who writes letters by hand is almost certainly someone who does not depend on instant communication for their livelihood. Friction-maxxing, as described by its proponents, is the rejection of convenience. But it is a rejection available only to people for whom convenience was never a necessity.

This objection, however strong, is also too easy. It is the criticism that greets every countercultural movement at birth, and history suggests that it is usually both correct and beside the point.

Consider mindfulness. In 1979, Jon Kabat-Zinn, a molecular biologist at the University of Massachusetts, adapted Buddhist meditation techniques into an eight-week clinical programme he called Mindfulness-Based Stress Reduction. For nearly three decades, it remained a medical intervention for chronic pain patients and a niche practice in academic psychology. Then, in 2007, Chade-Meng Tan, one of Google’s earliest engineers, built a course called Search Inside Yourself that taught mindfulness to Google employees as a tool for emotional intelligence and professional performance. The programme was so popular that by 2012 it had spun off into an independent nonprofit serving corporations worldwide. From Google, mindfulness spread to Goldman Sachs, to General Mills, to the United States Marine Corps, and eventually to a billion free YouTube videos accessible to anyone with a phone. The criticism of corporate mindfulness, that it was a luxury for people wealthy enough to hire meditation coaches rather than confront the structural conditions making them stressed, was entirely valid. It was also insufficient. The practice did reach beyond its original wealthy audience. It did help people who had no connection to Silicon Valley. The starting point was elite. The trajectory was not.

But the history of countercultural movements also reveals something less comforting than the mindfulness arc suggests. The trajectory from elite hobby to democratic practice is not the only pattern. There is another pattern, older and more persistent, in which the movement remains a class marker long after it has been rebranded as a universal value.

The Luddites are the textbook case. The popular version of the story has them as desperate workers smashing machines they did not understand. E.P. Thompson’s “The Making of the English Working Class,” published in 1963, demolished this caricature. Thompson showed that the Luddites were not unskilled labourers but highly skilled artisans: stockingers, croppers, and weavers whose work was so specialised and well remunerated that they owned their own tools and frames. They were not rebelling against technology itself. They were defending a specific economic arrangement in which skill and craft commanded respect and income. Their rebellion was funded by their prosperity. The workers who had no tools, no savings, no skills worth defending, were not Luddites. They were the people the Luddites were trying not to become.

The organic food movement follows the same arc. Erewhon, the health food store that has become a symbol of affluent wellness culture in Los Angeles, was founded in Boston in 1966 by Michio and Aveline Kushi, macrobiotic diet devotees who opened a tiny shop on Newbury Street selling whole grains and brown rice. When the brand moved to Los Angeles in 1969, it found its natural audience: health-conscious consumers with the income to pay more for food that promised purity. For decades, the organic label functioned as a class marker. Study after study confirmed what anyone who had compared the price of organic strawberries to conventional ones already knew: higher income and higher education predicted organic purchasing. The food itself may have been better. The access to it was not equal.

The communes of the 1960s tell the same story from a different angle. A Vermont Historical Society survey of more than seven hundred participants in the 1970s counterculture found that seventy-five percent held college degrees and many had postdoctoral education. These were not factory workers fleeing poverty. They were the educated children of the middle class, choosing a harder life precisely because they had the softer one waiting if the experiment failed. The commune was not an escape from privilege. It was an exercise of it: the freedom to opt out, temporarily, secured by the knowledge that opting back in remained an option.

And yet. The ideas that emerged from those communes, about ecological sustainability, communal responsibility, living deliberately rather than automatically, proved more durable than the communes themselves. The people who lived on them were privileged. The insights they generated were real. Dismissing the insight because of who generated it is a satisfying form of criticism that leaves you with nothing.

This is the tension at the heart of friction-maxxing in 2026. The movement may be telling the truth. Convenience may be eroding cognitive independence, emotional resilience, and the embodied experience of being a person in the physical world. The evidence from the Microsoft and Carnegie Mellon study, from the IKEA effect literature, from the testimony of INSEAD students who worry about their own deskilling, is not trivial. Something is being lost. The people who notice first are the people with the leisure to notice. That has always been true. It does not make the observation wrong.

What it makes the observation is expensive. And this is where the pattern stops being merely interesting and becomes something Thorstein Veblen would have recognised on sight.

In 1899, Veblen published “The Theory of the Leisure Class,” the book that gave us the phrase “conspicuous consumption.” Veblen’s central insight was not merely that rich people buy expensive things. It was that the function of expensive things is to demonstrate the buyer’s distance from necessity. The leisure class proved its status not through productivity but through visible waste: waste of money, waste of time, waste of effort. The more impractical the display, the more potent the signal. A manicured lawn signals status not because it is beautiful but because it is useless: land that could be productive, deliberately kept unproductive, for the sole purpose of proving that the owner can afford to waste it.

Now look at friction-maxxing through Veblen’s lens, and the category snaps into focus. The flip phone is not a rejection of the iPhone. It is a manicured lawn that looks like a rejection of the iPhone. The person carrying it broadcasts a specific signal: I am so securely positioned in the knowledge economy that I do not need to be reachable. I can afford the career risk of being uncontactable. I can afford the social risk of not being on the group chat. The device costs less than an iPhone. The lifestyle it requires costs considerably more.

What 2026 has produced is not the rejection of conspicuous consumption. It is its successor: conspicuous inefficiency. The visible, deliberate, curated embrace of difficulty as a status marker. Cooking from scratch when you could order in. Writing by hand when you could type. Reading a physical book on the train when you could read on a Kindle. Each choice broadcasts the same message: I have the time, the security, and the cultural capital to make my life harder, and I want you to see me doing it.

The person who cooks from scratch can order takeout tomorrow. The letter writer’s iPhone is charging in the kitchen. The flip phone carrier has a smartphone in the drawer. In every case, the exit exists. The hard life is a performance precisely because it has an exit, and having an exit is itself a form of capital as tangible as any salary. You do not want the hard life. You want the hard life with a door marked RETURN TO COMFORT whenever the friction stops being interesting and starts being real.

And this is where the pattern turns cruel. The single parent working two jobs does not friction-max. They are already living with more friction than any lifestyle influencer could manufacture. The gig worker who cannot afford to miss a notification does not carry a flip phone by choice. The family buying conventional strawberries at the budget supermarket is not making a statement about industrial agriculture. They are buying what they can afford. For these people, convenience is not a trap to be resisted. It is infrastructure they cannot do without. The movement that promises to liberate humanity from the tyranny of ease is, by its structure, a movement that can only be joined by the people least tyrannised by it.

Friction, in other words, has become the thing convenience was supposed to be: a privilege distributed by income.

This will not resolve itself. It will deepen. Friction-maxxing will grow, because the forces driving it are not going away. AI tools will become more capable. Convenience will become more total. The genuine erosion of cognitive independence, the real hunger for presence, the legitimate unease about what frictionless living does to human development: none of these concerns are imaginary, and all of them will intensify. More people will reach for the flip phone, the handwritten note, the home-cooked meal. Markets will form to serve them. Brands are already circling. It will not be long before friction-maxxing becomes a keyword for selling dumb phones and artisanal notebooks and premium cooking kits, at which point the rebellion will complete its transformation into the product it rebelled against.

The question that will define the next decade of “back to basics” movements is not whether friction is good for you. It probably is. The question is whether a good thing that is structurally inaccessible to the people who need it most is a movement or a market. The answer, if you look at two centuries of evidence, is that it has always been both. The commune was real and the commune was a luxury. The organic strawberry is better and the organic strawberry is a class marker. The flip phone may save your attention and the flip phone may be the newest Veblen good: an object whose value increases precisely because it signals what you can afford to give up.

The person reading this probably friction-maxes in at least one area of their life. You are not being criticised. The instinct is sound. The only honest thing to do with that instinct is to notice who else can afford to follow it, and who cannot. Because the most important friction in any society is never the kind you choose. It is the kind you cannot escape.

Sources:

Kathryn Jezer-Morton, “In 2026, We Are Friction-Maxxing,” The Cut, January 2026

Isabel Berwick, “Workers are ‘friction-maxxing’ to resist AI,” Financial Times, January 2026

Thorstein Veblen, The Theory of the Leisure Class, 1899

E.P. Thompson, The Making of the English Working Class, 1963

Jon Kabat-Zinn, Mindfulness-Based Stress Reduction, University of Massachusetts, 1979

Chade-Meng Tan, Search Inside Yourself, Google, 2007

Kate Fletcher, “Slow Fashion,” The Ecologist, 2007

Back Market, iPod sales data (reported by BGR/Yahoo Tech, 2026)

Microsoft and Carnegie Mellon University study on AI tools and critical thinking (reported by Raconteur, 2026)

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

You checked your phone before your feet hit the floor this morning. You did not check the weather. You did not check the news. You checked a notification, a message, a number. Something that told you where you stand relative to someone else. A reply that came or did not come. Likes on something you posted yesterday and have been silently counting since. An email from a name that produced a quick, involuntary spike of anxiety or satisfaction before you had finished reading the subject line. You did not decide to do this. Eight out of ten smartphone users reach for their phone within ten minutes of waking. The average American does it 144 times a day. But the frequency is not the point. The content is. Strip away the apps, the notifications, the interfaces, and what remains is a single query, fired over and over before lunch: where do I stand?

This is the invisible scoreboard. You have been keeping score on it your whole life, in a game you never agreed to play, on a board you have never consciously seen. It runs beneath every conversation you have, every room you enter, every career decision you have ever made, every restaurant you picked last Friday, every outfit you chose this morning, every scroll through your feed that left you feeling vaguely worse without knowing why. Status is not a sociological abstraction reserved for people who care too much about appearances. It is the operating system of your social brain, and it has been running without interruption since the day you were born.

Neuroscientists at the National Institute of Mental Health have mapped the architecture. In 2008, Caroline Zink and her colleagues scanned participants’ brains while they navigated a simulated social hierarchy and found that the ventral striatum, the brain’s reward centre, activated for outcomes that shifted a participant’s position in the ranking. The dorsolateral prefrontal cortex worked continuously to calculate where the participant sat relative to others. The system was automatic. It was continuous. It operated whether the participant was paying attention or not. A separate study, led by Joan Chiao in 2009, went further: the bilateral intraparietal sulci, the brain regions you use to decide which of two numbers is larger, are the same regions you use to decide which of two people ranks higher. Your brain treats “where do I rank in this room” as the same category of problem as “is seven bigger than four.” Status is not a feeling. It is a computation. And you run it every time you walk through a door.

You know this because you have felt it. You walk into a party, a meeting, a conference, a restaurant, and within seconds your brain has produced a map. You know who you would approach freely and who you would hesitate to approach. You know who you want to sit next to and who you are relieved is sitting somewhere else. You did not decide to produce this map. You could not stop producing it if you tried. It determines where you sit, who you speak to first, how loudly you laugh at whose joke, and whether you leave early or stay for one more drink. You navigate these calculations with the fluency of a native speaker in a language no one ever taught you.

The calculations have a cost. In 2003, Naomi Eisenberger and her colleagues at UCLA scanned the brains of people being excluded from a simple ball-tossing game and found that social exclusion activated the anterior cingulate cortex, the same region involved in the distress component of physical pain. Losing status does not merely feel unpleasant. It registers in the brain’s pain architecture. Gaining it lights up the same reward circuitry that fires when you eat after hunger. The scoreboard is not a metaphor your therapist uses. It is wired into the same hardware as pleasure and pain, and you were never given the option of ignoring it.

Consider, now, the career you are in. How much of it was chosen for what the work is, and how much was chosen for what the work signals? The writer Will Storr identifies three status games people play: dominance games, where status comes through force and intimidation; success games, where it comes from achievement and skill; and virtue games, where it comes from moral standing. Most careers are entries in one of these games, whether you have ever framed them that way or not. The lawyer who chose law partly because of how it sounds when someone at a dinner party asks what she does. The nonprofit worker who chose advocacy partly because of how it sounds when he explains it to his mother. The founder who chose founding partly because of the word itself: founder. There is always an audience, even when you have never consciously imagined one. You are not optimising only for income, or fulfilment, or impact. You are optimising, simultaneously and without noticing, for a version of yourself that performs well on a scoreboard inside your head.

Richard Easterlin proved the arithmetic of this in 1974, and half a century of data has not overturned it. Across nations and decades, rising income does not make people happier. What matters is not how much you earn but how much you earn relative to the people you compare yourself to. Your salary is not a number. It is a coordinate on the invisible scoreboard, and the coordinate only has meaning relative to your neighbours, your graduating class, the friends whose updates appear on your feed. Double everyone’s income tomorrow and nobody feels richer. The scoreboard recalibrates. It always recalibrates.

This is why four minutes on Instagram can leave you feeling worse than four minutes staring at a blank wall. You already know this. You have felt the small, formless deflation after a scroll that you could not quite name. A 2023 meta-analysis in Media Psychology, drawing on dozens of studies, confirmed it: upward social comparison on social media produces significant negative effects on body image, self-esteem, mental health, and subjective well-being. A 2025 study in Frontiers in Psychology pinpointed the mechanism: upward comparisons on Instagram mediate the link between using the platform and feeling worse about yourself. The app is not showing you upsetting content. It is running two hundred status comparisons in four minutes, and most of them are telling you that someone out there is thinner, richer, happier, more successful, more loved, or more at ease than you are. The feeling you could not name after the scroll was the feeling of losing a game you did not know you were playing.

The algorithm knows you are playing. It was designed to know. Variable ratio reinforcement, the same reward schedule that makes slot machines the most profitable machines in a casino, is the architecture underneath every notification system on your phone. Sometimes the comparison is favourable and you feel a small flush of warmth. Sometimes it is not and you scroll faster, looking for the next one. The inconsistency is not a flaw. It is the mechanism. Slot machines do not work if they pay out every time. They work because they pay out unpredictably, and your brain cannot stop pulling the lever because the next pull might be the one that lands. You check your phone 144 times a day for the same reason a gambler puts another coin in the slot: the variable reward has hijacked the circuitry that tracks your status, and the circuitry does not know the difference between a favourable comparison and a jackpot.

The scoreboard follows you to dinner. You chose the restaurant partly for the food and partly for what eating there says about you. At the table, the conversation is two conversations running at once: the one you can hear and the one you cannot. The one you cannot hear is the status negotiation. Who tells the story and who laughs at it. Who defers and who holds the floor. Who mentions the trip, who does not mention theirs. Who picks up the bill and what the gesture signals to everyone watching. You manage these transactions with the ease of someone who has been doing it their entire adult life, because you have, and you have never needed a manual. The manual was written a hundred thousand years ago for a primate in a group of 150 where being excluded meant death, and it is still running, unrevised, in a world of eight billion.

Here is the turn. In 2023, a team led by Akihiro Nishi and Nicholas Christakis at Yale ran a series of experiments published in Nature Mental Health. They placed 1,289 participants into networked groups, assigned them different levels of wealth, and had them interact across rounds of economic games. In some groups, participants could see how wealthy their neighbours were. In others, wealth was hidden. When wealth was visible, a steep happiness gradient appeared: the rich reported significantly higher well-being, the poor significantly lower. When wealth was invisible, the gradient collapsed. Poorer participants’ well-being rose sharply. The inequality had not changed. The resources had not changed. What changed was the visibility of the scoreboard. The mere knowledge of where you stand relative to others was doing most of the damage. Not the disadvantage. Not the deprivation. The comparison.

The implication is not that status games are bad. You could not quit if you tried. Storr puts it plainly: to our brains, status is a resource as real as oxygen or water. The neural machinery that tracks it kept your ancestors alive in a world where exclusion from the group was death. You are not going to uninstall it. But the Nishi experiment reveals something that changes the game without ending it. The scoreboard’s power depends on its invisibility. When the researchers let participants see exactly whose wealth was being compared, those participants could observe the mechanism at work. When they hid it, the mechanism ran unchecked and unhappiness followed. The same principle holds outside the lab. The moment you notice that you chose the restaurant for what it signals, that you checked your phone for where you stand, that your career is partly a costume designed for an audience that lives inside your head, you have done the one thing the invisible scoreboard cannot survive. You have made it visible. And a visible scoreboard is a scoreboard you can finally question.

This is where the personal question becomes a civilisational one. The anthropologist Robin Dunbar estimated that human social groups historically capped at roughly 150 people. The status-monitoring circuitry in your skull evolved for a village. It was never designed for a platform that runs two hundred curated status comparisons in a four-minute scroll. It was never built for a professional network that surfaces the accomplishments of ten thousand peers with a single click. It was never engineered for a world where the wealthiest people have begun signalling status not by what they own but by what they can afford to opt out of: screen-free childhoods, phoneless dinners, the conspicuous luxury of being unreachable. The scoreboard has grown so vast and so inescapable that the newest status symbol is the ability to walk away from it. And even that, of course, is just another move in the game.

You are still reading this. Your phone is near your hand. In a moment you will pick it up, and now, perhaps for the first time, you will notice the tiny calculation your brain runs before it decides what to do next: is this what someone like me would do? That question, which you have asked yourself ten thousand times without ever hearing it, is the invisible scoreboard. You cannot turn it off. But you can, starting now, ask a different question. Not “what is my score” but “whose scoreboard am I using.” Because the cruelest feature of the invisible scoreboard is not that it makes you keep score. It is that most people never write their own. They inherit a board from their parents, their class, their profession, their feed, and they spend decades climbing a ladder leaned against someone else’s wall, keeping score in someone else’s game, and wondering why winning never feels the way they thought it would. The game only owns you for as long as you never notice you are playing. And now you know.

Sources:

Zink, Tong, Chen, Bassett, Stein & Meyer-Lindenberg, “Know Your Place: Neural Processing of Social Hierarchy in Humans,” Neuron, 2008

Chiao, Harada, Oby, Li, Parrish & Bridge, “Neural Representations of Social Status Hierarchy in Human Inferior Parietal Cortex,” Neuropsychologia, 2009

Eisenberger, Lieberman & Williams, “Does Rejection Hurt? An fMRI Study of Social Exclusion,” Science, 2003

Will Storr, The Status Game, William Collins, 2021

Richard Easterlin, “Does Economic Growth Improve the Human Lot?” in Nations and Households in Economic Growth, Academic Press, 1974

McComb, Vanman & Tobin, “Upward Social Comparison on Social Media,” Media Psychology, 2023

Le Blanc-Brillon, Fortin, Lafrance & Hetu, “Instagram Use and Self-Esteem: The Mediating Role of Upward Social Comparison,” Frontiers in Psychology, 2025

Nishi, German, Iwamoto & Christakis, “Status Invisibility Alleviates the Economic Gradient in Happiness in Social Network Experiments,” Nature Mental Health, 2023

Reviews.org, “How Many Times Do You Check Your Phone Per Day?” 2023

Dunbar, “Neocortex Size as a Constraint on Group Size in Primates,” Journal of Human Evolution, 1992

“The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com”

“Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.”

THE SCENARICA SUNDAY

A surgeon named Bruce Moseley stood in an operating theatre at the Houston Veterans Affairs Medical Center and performed knee surgery on a man who could barely walk. Moseley made three small incisions in the skin. He let the patient hear the clink of instruments, the murmur of the surgical team, the splash of saline against tissue. He moved the knee through its range of motion. He sutured the incisions closed.

He had not entered the joint. The arthroscope had not been inserted. The cartilage had not been touched. Moseley had cut the skin, performed a theatre of medicine, and sewn the man back up. The knee was exactly as it had been when the patient was wheeled in.

Two years later, the patient was walking without pain. He mowed his lawn. He climbed stairs. His knee, he told the researchers, felt just like the other one. He was not an outlier. In the trial that Moseley published in the New England Journal of Medicine, one hundred and eighty patients with debilitating osteoarthritis were randomly assigned to real arthroscopic surgery or sham surgery. The sham group, i.e. the patients whose knees were never actually repaired, reported outcomes indistinguishable from those who received the genuine procedure. At some follow-up points, the patients who had received the fake operation reported better results than those who had received the real one.

A man’s belief that his knee had been fixed was sufficient to make his body behave as though it had been.

This is not the strange part. The strange part came thirteen years later.

In 2015, a team at McGill University led by a doctoral student named Alexander Tuttle did something that nobody had thought to do before. They gathered eighty-four clinical trials of neuropathic pain drugs conducted around the world between 1990 and 2013 and measured not just whether the drugs worked but how the placebo groups had changed over time. The placebo response, the degree to which patients taking sugar pills reported genuine, measurable pain relief, had been climbing. By 2013, patients in placebo groups were reporting an average thirty per cent decrease in pain levels. The sugar pills were getting stronger. Drug responses, meanwhile, had remained flat. The gap between the real drug and the empty capsule had been narrowing for twenty-three years, and by the time Tuttle’s team compiled the data, it had collapsed so far that pain drug trial after pain drug trial was failing to show any benefit over placebo, not because the drugs had gotten worse, but because the placebos had gotten better.

But Tuttle’s most disorienting finding was not the trend. It was the geography.

The increase had occurred only in the United States. Trials conducted in Europe showed no change in placebo response. Trials in Asia showed no change. American bodies were responding to empty pills more powerfully than they had a generation earlier. European and Asian bodies were not.

There are operational explanations. American trials had grown longer, from an average of four weeks in 1990 to twelve weeks by 2013. They had grown larger, from fifty patients to more than seven hundred. Bigger trials with more clinical attention can inflate the placebo response because the care itself becomes therapeutic. But those factors alone do not explain a geographic lock this clean.

The United States is one of only two countries on earth, the other is New Zealand, that permits pharmaceutical companies to advertise prescription drugs directly to consumers. American patients are exposed to roughly ten billion dollars a year in drug advertising. They see, thousands of times over, a specific narrative: a person suffers, a person takes a pill, a person is healed. The arc is always the same. The lighting is always warm. The music always resolves. The hypothesis that researchers including Tuttle’s senior author Jeffrey Mogil have raised, not yet proven, but increasingly discussed in the literature, is that this saturation has trained American brains to respond more powerfully to the act of taking any pill, because the cultural narrative that pills work has become so deeply embedded that the expectation itself now generates a physiological response that did not exist at the same magnitude in 1990.

That word, physiological, is the hinge of the entire story. The placebo effect is not a trick of perception. It is not the patient imagining they feel better and reporting inaccurately. When a patient expects pain relief, PET imaging studies show the brain releasing endorphins and dopamine along specific neural pathways. The prefrontal cortex activates. The pain matrix quiets. The endogenous opioid system fires as though an actual drug had entered the bloodstream. The relief is not imagined. It is manufactured by the brain, from raw expectation, using the same neurochemical machinery that real analgesics use. What changed in the United States was not the machinery. It was the strength of the input. Thirty years and hundreds of billions of dollars of advertising taught a nation’s nervous systems to expect that pills work. And then the nervous systems obliged.

This is where the story stops being about medicine.

Because if narrative can trigger a neurochemical cascade powerful enough to make a man walk on an unrepaired knee, then narrative is not a psychological curiosity. It is a physiological force. And once you accept that, you begin to see it operating in every system where belief shapes outcome, which is to say, in almost every system that matters.

In 2008, a team led by Dan Ariely at MIT gave eighty-two volunteers identical placebo pills and told half of them the pill cost two dollars and fifty cents per dose. They told the other half it had been discounted to ten cents. Then they administered controlled electric shocks to the volunteers’ wrists and measured their pain. Among the group told the pill was expensive, eighty-five per cent reported significant pain reduction. Among the group told the pill was cheap, sixty-one per cent did. Same pill. Same shock. Same sugar. The only variable was the price, and the price changed the body’s response. The expensive lie worked better than the cheap one.

Now consider what a stock market is. A stock market is a system in which collective belief about future value determines present price. When investors believe an asset will appreciate, they buy. Their buying drives the price up. The rising price confirms the belief that drove the purchase. The cycle is self-reinforcing, and it runs on exactly the same mechanism as the placebo: a narrative about what will happen generates the neurochemical and behavioural changes that make it happen. The dot-com boom was not a market failure. It was a market placebo. The AI valuation wave of the mid-2020s operated by the same pharmacology: a narrative powerful enough to generate the investment that generated the earnings that appeared to justify the narrative. And just as the expensive sugar pill outperformed the cheap one, the most expensively packaged market story, the one with the most prestigious analysts, the highest-profile conference appearances, the most authoritative tone, produced the strongest response. The packaging was never incidental. In a system where belief drives outcome, the packaging is the drug.

The pattern goes deeper. Political authority is a placebo in the most literal sense the word allows. A government governs because its citizens believe it governs. The institution is real, it has buildings, soldiers, bureaucrats, filing cabinets. But its authority is not contained in any of those objects. Its authority is contained in the shared belief that those objects represent legitimate power. The moment the belief withdraws, the authority vanishes as though it had never been there. Because it was always the belief doing the work, not the institution.

East Germany in November 1989 is the purest case study. The Berlin Wall was concrete and steel. The border guards had loaded weapons. The state security apparatus, the Stasi, maintained one officer or informer for roughly every sixty-three citizens and kept files on a third of the population. The infrastructure of control was massive, physical, and real. And it dissolved in a single evening, because the citizens stopped believing in it. Nobody stormed the Wall with superior weapons. Nobody overpowered the guards. The guards, vastly outnumbered, simply found that nobody in the chain of command would take personal responsibility for ordering them to shoot. The authority had been a shared hallucination, and when the hallucination broke, the concrete became irrelevant. The Soviet Union followed two years later. An empire that had defined half the planet’s political architecture for seventy years did not fall to an invading army. It fell because the story stopped working. Legitimacy is the political sugar pill, and when a society stops swallowing it, the state discovers that the filing cabinets were never the point.

The newest domain is the most unsettling. In 2022, a team led by Thomas Kosch published a study in ACM Transactions on Computer-Human Interaction that might be the most important paper about artificial intelligence that has nothing to do with artificial intelligence. They gave nearly five hundred participants a series of word puzzles. Half were told they would receive help from an adaptive AI system. The other half were told they would work alone. Every participant received exactly the same puzzles at exactly the same difficulty. No AI was present in either condition. The system was a sham.

The participants who believed they had AI assistance performed better. Their expectations about their own performance increased, and those expectations persisted even after the interaction ended. They solved more puzzles. They rated their experience higher. The belief that an intelligent system was helping them was sufficient to make them more capable, not because the system was doing the work, but because the expectation altered their engagement, their confidence, and their willingness to persist. The AI placebo.

Now extend the finding. When users believe a model is accurate, they check its outputs less. When they check less, errors pass undetected. When errors pass undetected, the system appears to work. The appearance of working reinforces the belief that the system is accurate. The cycle is identical to the clinical trial and the stock market: narrative generates the behaviour that generates the evidence that confirms the narrative. Automation bias is not a bug in human-AI interaction. It is the placebo effect wearing different clothes. And like every placebo, it works beautifully, right up until the moment someone looks inside the capsule and finds it empty.

Here is the turn. If belief can heal, belief can harm.

The medical literature calls it the nocebo effect, from the Latin “I shall harm,” and it is the placebo’s dark twin. In clinical trials, roughly one in five patients assigned to the placebo group reports side effects, nausea, headache, fatigue, dizziness from pills that contain nothing but inert filler. Patients told that a drug may cause pain experience pain. Patients warned about nausea feel nauseous. Brain imaging shows that nocebo responses are not imagined either: negative expectations trigger the release of cholecystokinin, a neurotransmitter that amplifies pain signals, and activate regions of the anterior cingulate cortex associated with threat processing. The arrow points both ways. Narrative does not merely enhance positive outcomes. It manufactures negative ones.

In a culture saturated with narratives, pharmaceutical, financial, political, technological, the nocebo is not theoretical. It is structural. A society that spends billions telling its citizens that pills heal also tells them, through side-effect disclosures and health anxiety media, precisely what to fear. A market that runs on confidence can also run on panic, and the panic is just as self-fulfilling as the confidence was. A political system that operates on shared belief can be destroyed not only by the withdrawal of belief but by the active propagation of disbelief and the mechanism by which a society talks itself into the outcomes it fears is neurochemically identical to the mechanism by which it talks itself into the outcomes it hopes for.

This is why the placebo is not a curiosity. It is the central variable. Every system that depends on human participation depends on narrative, and every narrative carries both the placebo and the nocebo in the same capsule. The question is never whether belief is shaping the outcome. The question is which belief, and in which direction, and how much of what we are measuring is the substance and how much is the story.

In a clinic in Houston, a man whose knee was never repaired walked without a cane for years. His improvement was not a delusion and it was not a trick. His brain produced the endorphins. His prefrontal cortex modulated the pain. The neurochemical cascade was real, triggered not by a scalpel but by a narrative, the narrative that a surgeon had entered his joint and made him whole. The lie was beautiful because it was, in every way that mattered to the man walking down the hall on his own two legs, true.

He never did learn, during the blinded follow-up period, that his surgery was a performance. He only knew that his knee worked and that his life had changed. Somewhere in the space between the incision that went nowhere and the leg that carried him forward, belief became biology, and the empty capsule became the most powerful drug in the building. The only question left, the one the man in Houston never had to answer but the rest of us do, every day, in every system we inhabit, is what happens when we finally look inside, and discover that the pill was empty all along. And it is working.

Sources: J. Bruce Moseley et al., “A Controlled Trial of Arthroscopic Surgery for Osteoarthritis of the Knee,” New England Journal of Medicine, 2002. Alexander H. Tuttle et al., “Increasing Placebo Responses over Time in U.S. Clinical Trials of Neuropathic Pain,” Pain, 2015. Rebecca L. Waber et al., “Commercial Features of Placebo and Therapeutic Efficacy,” JAMA, 2008. Thomas Kosch et al., “The Placebo Effect of Artificial Intelligence in Human-Computer Interaction,” ACM Transactions on Computer-Human Interaction, 2022. Irving Kirsch et al., “Initial Severity and Antidepressant Benefits: A Meta-Analysis of Data Submitted to the Food and Drug Administration,” PLoS Medicine, 2008. Ted J. Kaptchuk et al., “Placebos without Deception,” PLoS One, 2010. CSRxP, “Big Pharma’s Direct-to-Consumer Advertising,” 2025. McGill University Newsroom, “American Placebo,” 2015. Jo Marchant, “Strong Placebo Response Thwarts Painkiller Trials,” Nature, 2015.

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

On the evening of February 27, 1968, Walter Cronkite looked into a CBS camera and told the United States that the Vietnam War could only end in stalemate. It was the most famous editorial judgment in the history of broadcast journalism. Lyndon Johnson reportedly turned to an aide and said: “If I’ve lost Cronkite, I’ve lost Middle America.”

The sentence has been repeated so often that its most important word has been drained of meaning. Middle. Johnson did not say he had lost the left or the right. He said he had lost the middle, the vast, persuadable centre of American opinion that had no fixed position on Vietnam and no ideological commitment to either hawks or doves. Johnson feared Cronkite’s verdict because those people were watching. Because in February 1968, the American information ecosystem consisted of three television networks, a handful of national newspapers, and whatever your neighbour said over the fence. If you wanted to know what was happening in the world, you sat in the same room as everyone else. You heard the same facts, the same interpretation, the same thirty minutes of news. You might disagree violently with what Cronkite said. But you heard him say it.

The room was the mechanism.

James Madison understood this two centuries before anyone owned a television. In Federalist No. 10, published in November 1787, Madison laid out the problem that would define every democracy that followed: faction. Groups of citizens united by a shared passion or interest that ran against the rights of others. Madison’s concern was that a single faction would grow powerful enough to dominate. His solution was counterintuitive: make the republic larger. A bigger republic would contain more factions. More factions would mean more collisions between competing interests. The collisions would prevent any single faction from achieving majority control. Extend the sphere, Madison wrote, and you take in a greater variety of parties and interests. You make it less probable that a majority of the whole will have a common motive to invade the rights of other citizens. Madison’s bet was on encounter, not consensus. His republic worked because citizens could not avoid each other. The factions collided because the shared space forced collision, because the room was shared, because the architecture did not depend on citizens being brave.

In January 2026, the Edelman Trust Barometer surveyed 33,000 people across twenty-eight countries and measured how far the shared space has eroded. Only thirty-nine per cent of respondents said they encounter information weekly from sources with a different political leaning, down six points in a single year. Seven in ten are unwilling or hesitant to trust someone who differs from them in values, information sources, or worldview. In Japan, the figure reaches ninety per cent. In Germany, eighty-one. In the United Kingdom, seventy-six. Insularity cuts across income, gender, and age. It is not a fringe phenomenon. It is the new centre, and it has no geography and no party affiliation.

The trust gap between high-income and low-income respondents, which stood at six points in 2012, has more than doubled to fifteen points globally. In the United States, it has reached twenty-nine points. Two populations experiencing the same economy, interpreting it through entirely different information architectures, and arriving at conclusions so divergent that neither can understand the other’s anger. Only thirty-two per cent of people globally believe the next generation will be better off. The shared room is emptying. The people leaving it are not being pushed. They are being pulled, by products and preferences and platforms that learned, correctly, that comfort converts better than challenge.

The shared room emptied in stages. The first stage was selection. Social media replaced broadcast with the algorithmic feed. The crucial shift was from content that somebody else chose for you to content that a system chose based on what you had already consumed. Facebook’s News Feed, launched in 2006, did not tell you what was happening in the world. It told you what was engaging. And engagement, the algorithms discovered through billions of interactions, correlates overwhelmingly with confirmation. Content that confirms your existing beliefs gets clicked, liked, shared. Content that challenges them gets scrolled past. The algorithm optimised for attention. Polarisation was the exhaust fume of a machine built to sell it. Political scientists measured the consequences across multiple democracies: citizens who encountered opposing viewpoints moderated their positions. Citizens sealed inside ideologically homogeneous feeds became more extreme. The encountering was Madison’s theoretical mechanism, empirically verified and systematically removed.

The second stage is generation. And this is where the architecture changes in kind, not merely in degree. A large language model does not select from existing content. It creates new content, infinite, on-demand, and perfectly calibrated to the person requesting it. ChatGPT now processes over two billion queries daily, serving nine hundred million weekly active users. Consider what this means in practice. On a given morning, two people in the same city ask their AI assistant the same question: what should I know about the economy today? The economy, like all economies, is doing several things simultaneously. Wages are rising in some sectors and stagnating in others. Unemployment is low but labour force participation is uneven. Housing starts are falling while consumer spending holds. Inflation has cooled but grocery prices remain elevated. Both summaries the AI produces are accurate. Both contain real data. But the first user, whose reading history skews toward labour markets and wages, receives a summary that leads with employment strength and notes housing softness as a secondary concern. The second user, whose consumption patterns lean toward cost-of-living coverage, receives a summary that leads with persistent grocery inflation and mentions employment only in passing. Neither summary is wrong. Both are incomplete. And the two users walk into the same voting booth with coherent, data-supported, mutually incompatible understandings of whether the economy is working.

This is where the language we have inherited fails us. Misinformation is false. Disinformation is deliberately false. What the AI produces is neither. It is accurate, personalised, and fragmentary. It tells the truth, but only the truths that fit the question, in the tone the questioner prefers, at the depth they want, framed in the values they already hold. The AI does not need to lie. It needs only to select. And selection, at the scale of two billion daily queries, produces not a shared reality but nine hundred million private ones, each internally consistent, each empirically grounded, each invisible to the others.

You have already experienced this. You may not have named it, but you have felt it. Think about the last time you had a conversation with someone whose political instincts differ from yours, not a stranger online but a colleague, a family member, someone you respect, and discovered that you were not disagreeing about the interpretation of a shared set of facts. You were operating from entirely different facts. They cited something you had never seen. You cited something they had never encountered. The disagreement was not about values. It was about what had happened, and the two of you could not agree on that because your respective information architectures had been delivering you different versions of the same week for so long that the gap had become invisible to both of you. You did not enter the conversation knowing you were in a different reality. You entered it assuming you were in the same one. The shock was not the disagreement. The shock was discovering how far apart the starting points had drifted without either of you noticing.

Now think about the conversations you are no longer having. The arguments you stopped encountering, not because you decided to avoid them but because your feed, your recommendations, your search results, and now your AI assistant quietly curated them away. The perspectives that used to arrive unbidden, on the opinion page you did not choose, in the segment you did not skip, from the colleague whose desk happened to be next to yours, no longer arrive at all. The room that used to force encounter now lets you pass through an entire day, an entire week, an entire year, without hearing a single argument that makes you genuinely uncomfortable. You did not choose this. You did not have to. The architecture chose it for you, and the only evidence that something is missing is the slowly fading memory that disagreement used to feel different. It used to begin with shared premises and diverge from there. Now it begins with incompatible premises and has nowhere to go.

The marketplace of ideas, the governing metaphor of democratic theory, assumed a marketplace. A physical or institutional space where different ideas competed for the same audience. The audience had to walk past stalls they did not intend to visit. They had to hear pitches they did not seek. The marketplace worked because the architecture forced encounter. The citizen who came for grain heard the speech. The viewer who came for weather stayed for the editorial. The reader who opened to sports passed the front page. When the marketplace becomes personalised, it stops being a marketplace. It becomes a mirror. And mirrors, however high-resolution, show you only what you already are.

The Edelman data suggests we are approaching a threshold. Cross-cutting exposure at thirty-nine per cent, declining at six points per year, reaches single digits within half a decade. At that point, elections are no longer contests between competing interpretations of shared facts. They are contests between incompatible realities, each AI-reinforced, each internally coherent, each sealed against correction from the outside. And the institutions that were designed to process disagreement will find that the disagreement they are being asked to resolve is no longer the kind they were built for.

What happens to institutions that were built for the shared room once the room empties?

The answer is already visible, and it does not look like collapse. It looks like procedure without substance. In 2025, eighty-five per cent of roll-call votes in the United States Congress were party unity votes, a majority of one party facing off against a majority of the other. The figure broke the previous record by more than ten percentage points. Congress still convenes. Bills still pass. Votes are still recorded. But the deliberation that the institution was designed to produce, the collision of factions, the encounter that Madison engineered, has been replaced by two caucuses operating from two realities, each reinforced daily by the information architectures their constituents inhabit. The legislators are not failing to deliberate because they lack goodwill. They are failing to deliberate because they no longer share the factual premises that deliberation requires.

Courts assume shared facts. A trial functions because both sides, however adversarial, present evidence to the same jury drawn from the same community, and the community shares enough common ground to weigh one account against another. Elections assume shared stakes. Voters choose between competing interpretations of the same reality, and the loser accepts the outcome because they recognise the reality, even if they dispute the interpretation. Science assumes shared scrutiny. A finding is published to a common audience that can replicate, challenge, or refine it. Every one of these institutions was built for Madison’s architecture. They assume encounter. They assume the room.

When the room empties, these institutions do not break dramatically. They continue to function procedurally while losing the shared substrate that made their procedures meaningful. Congress votes, but on premises its two halves no longer share. Courts rule, but on evidence that competing information architectures frame as illegitimate before the verdict arrives. Elections produce winners, but the losers increasingly lack the shared reality that would make concession intelligible. The building is still standing. The room inside is empty.

And here is the part that concerns you directly. The emptying of the shared room is not only a problem for institutions. It is a problem for the person sitting in the private room, which is to say, for you. The shared room was not just the architecture of democracy. It was the architecture of your own intellectual growth. Every idea you hold that is worth holding was tested, at some point, against an idea that challenged it. Every opinion you have refined was refined because you encountered friction, an argument you could not dismiss, a perspective that forced you to adjust. The shared room did not just produce democratic legitimacy. It produced the conditions under which individual minds became more complex, more nuanced, more capable of holding two competing thoughts at once and weighing them honestly.

That friction is disappearing from your life, and the loss is almost impossible to detect from the inside. You feel informed. Your feed is full. Your AI answers every question. The information arrives faster, in higher resolution, with more data than any previous generation could access. What it no longer delivers is the thing that made all that information useful: the challenge. The argument you did not want to hear that turned out to be half right. The perspective you dismissed that later made you reconsider. The discomfort that was, in retrospect, the feeling of learning something. The private room is comfortable. It is also, by slow and imperceptible degrees, making your mind a smaller place to live.

Madison’s republic was a bet on architecture. He designed a system that did not need citizens to seek out disagreement voluntarily. The structure did the work. The factions collided because the room was shared. That architecture is being dismantled, quietly, commercially, and with the full participation of the people it was designed to protect. Tomorrow morning you will open a device and it will show you a world built entirely around what you already believe. It will feel like the whole world because it will contain everything you expect and nothing that disturbs it. Somewhere, in a room you will never enter, someone will experience the same sensation, complete, coherent, certain, and the two of you will share a city, a country, a ballot, and nothing else. The room is not locked. It is simply no longer shared. And the institutions that were built inside it, including the one between your ears, are learning, slowly and badly, to function in the dark.

Sources:

Walter Cronkite, CBS Evening News editorial on Vietnam, CBS, 1968

James Madison, Federalist No. 10, The Federalist Papers, 1787

2026 Edelman Trust Barometer, Edelman, 2026

Axios, “Exclusive: Global trust data finds our shared reality is collapsing,” 2026

CQ Roll Call, “Vote Studies: 2025 Sets New Mark for Partisanship on Capitol Hill,” 2026

OpenAI usage statistics reported by TechCrunch, February 2026

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI.

We don’t predict the future. We price it.

THE SCENARICA SUNDAY

Every morning at eight o’clock, a valet drew back the curtains of the most powerful man in Europe. What followed was not private. The Grand Chamberlain entered first, then the princes of the blood, then the senior officers of the household, then the physicians, then the favoured courtiers who had purchased the right to watch. In strict waves, dozens of men filed into the bedchamber of Louis XIV at Versailles to observe their king perform the act of getting dressed. One courtier handed the king his right sleeve. Another handed the left. A third held the mirror. The ceremony was called the lever du roi, and it happened every single day for over fifty years. The Sun King did not endure this because he lacked the power to stop it. He orchestrated it. Visibility was the architecture of his authority. You could not rule France from behind a closed door. You ruled it by letting France watch.

Three centuries later, in 2026, a warehouse worker in Kentucky begins her shift by scanning a badge that logs her arrival to the second. For the next ten hours, a handheld device tracks every item she picks, every step she takes between shelves, and every interval in which her hands are not touching a product. The system is called Time Off Task. If she accumulates four minutes of inactivity outside her scheduled break, she receives an automated warning. Her bathroom visits are not prohibited, but they are measured, and they count against her. She is, by any historical standard, one of the most closely watched human beings who has ever lived. She has no courtiers. She has no throne. She has an algorithm and a twelve-dollar hourly wage.

This is a story about a reversal so complete that we have not yet found the language to describe it.

For most of recorded history, visibility was the price of power. The arrangement was so consistent across civilisations, across millennia, across every form of government from theocracy to constitutional monarchy, that it barely registered as an arrangement at all. It was simply how human societies worked. The powerful were watched. The powerless were not. And the reasons had nothing to do with protecting the powerless. Nobody watched the serf because the serf did not matter enough to watch.

Consider the structure. In ancient Egypt, the pharaoh was theoretically the high priest of every temple, but because he could not be everywhere at once, he delegated his religious duties to a priesthood that performed sacrifices in his name. Every offering carried the formula “an offering in the king’s name.” The priests did not watch the pharaoh out of suspicion. They watched him because his visibility was his legitimacy. The entire theological apparatus depended on the populace knowing that the pharaoh was performing his divine role. An invisible pharaoh was a theological crisis.

In medieval Europe, the pattern intensified. The royal bedchamber was not a private room. It was a political stage. Kings held court in their bedrooms, received ambassadors beside their beds, and conducted the business of government in a space that modern sensibility would consider the most intimate in any home. The royal bed was a second throne. Thick hangings provided an illusion of seclusion, but the room was never truly private, because privacy for a medieval monarch would have been indistinguishable from irrelevance. A king nobody could see was a king nobody had to obey.

And then there were the courtiers. The lever du roi was not a French invention. It was the apotheosis of a European tradition in which proximity to the monarch’s body was the currency of political power. The right to hand Louis XIV his shirt was not a chore. It was a privilege men spent fortunes to acquire. The entire system ran on a brutal reciprocity: the king submitted to being watched, and in return, the watchers submitted to being controlled. The gaze went upward, and power flowed downward.

Now consider the other side of this arrangement. What was life like for the people nobody watched?

A Burgundian peasant in 1685, the year Louis XIV revoked the Edict of Nantes and tightened his grip on every institution in France, lived in a world without records. No government agency tracked his purchases, because he made them with coin at a local market and no receipt survived the transaction. No employer monitored his productivity, because he worked his own strip of land or his lord’s fields at a pace dictated by the season and his own body. No system logged his movements, because nobody with power had any reason to know where he was. He was illiterate, probably poor, almost certainly cold in winter, and by every modern metric, unfree. But he possessed something that no minimum-wage worker in 2026 possesses: the absolute certainty that nobody powerful was paying attention to him. Not because he was protected, but because he was ignored. His privacy was not a right. It was the condition of not mattering.

This is the fact that makes the modern arrangement so disorienting. We did not simply erode privacy. We inverted the entire structure. We took the condition that once applied to the powerless, invisibility, and sold it as a luxury to the powerful. And we took the condition that once applied to the powerful, constant observation, and imposed it on everyone who cannot afford to refuse.

The numbers are staggering, and they arrived fast. In 2026, seventy-eight per cent of employers monitor their workers digitally. The tools go far beyond checking whether someone is at their desk. More than five hundred and fifty employee monitoring products now offer keystroke logging, screen recording, webcam capture, and GPS tracking. AI productivity systems aggregate data from mouse movements, application usage, email response times, and meeting attendance into a single score for each employee. The most surveilled workers are not executives. They are hourly employees, gig workers, call-centre operators, and warehouse staff, the people with the least power to object.

The pattern extends beyond the workplace. In the United States, recipients of public assistance submit to a surveillance regime that would have been unthinkable for the aristocrats who once administered poor relief. Every SNAP transaction is electronically recorded. The government knows the date, time, dollar amount, and retailer for every purchase made with an EBT card. The system dictates what recipients may buy and flags deviations. Virginia Eubanks, a political scientist who spent years studying these systems, calls the result a “digital poorhouse,” an apparatus designed not primarily to help the poor but to profile, police, and punish them. In colonial Massachusetts, a public official called the Overseer of the Poor had the authority to warn out, that is, to expel, anyone who could not demonstrate their solvency to the satisfaction of the local Selectmen. The poor were tracked not to protect them but to manage them. Three hundred years later, the tracking has become automated, algorithmic, and vastly more comprehensive, but the function is identical. The direction of the gaze reveals who holds the power.

Meanwhile, at the other end of the income spectrum, the ultra-wealthy are spending unprecedented sums to achieve what the Burgundian peasant had for free. Privacy-focused travel now costs upwards of five hundred thousand dollars per trip, and the industry is projected to grow twenty-five per cent annually through 2030. The wealthy travel under shell companies, fly private, rent entire islands, and employ counter-surveillance teams to ensure that no photograph, no flight record, no hotel check-in connects their name to their location. They are purchasing invisibility. They are buying the peasant’s condition at a billionaire’s price.

Michael Fertik, the founder of Reputation.com, put it simply: the rich see a different internet than the poor. When you can afford to pay for services, you are the customer and your preferences are served. When you cannot afford to pay, you are the product and your data is extracted. The free version of the internet, the one used by people who cannot afford the paid version, harvests more personal information per hour than the entire apparatus of Versailles extracted in a decade.

Jeremy Bentham, the philosopher who designed the Panopticon in 1791, understood something about surveillance that took two more centuries to become obvious. His prison was a circular building with a watchtower at the centre. The guards could see every cell, but the prisoners could never see the guards. The genius of the design was not the watching. It was the uncertainty. The prisoners did not know when they were being observed, so they had to behave as though they were being observed at all times. They internalised the gaze. They became their own jailers. Michel Foucault seized on Bentham’s blueprint and turned it into a theory of modern power. The Panopticon was not just a building. It was a model for how institutions discipline entire populations: through visibility, through the threat of being seen, through the slow conversion of external control into self-regulation.

What neither Bentham nor Foucault anticipated was the direction the gaze would travel. They assumed it would point from the powerful toward the powerless. They were right about the mechanism. They were wrong about who would end up in the tower and who would end up in the cells. In the twenty-first century, the watchtower is empty at the top. The wealthiest, the most powerful, the most consequential decision-makers in the global economy have achieved a degree of opacity that Louis XIV would have found unimaginable. They make decisions in private. They communicate through encrypted channels. They structure their affairs through layers of holding companies and trusts designed to make ownership invisible. The Sun King could not eat breakfast without an audience. A billionaire in 2026 can move markets without anyone knowing his name was involved.

And in the cells below, the workers and the benefit recipients and the users of free platforms live under a surveillance regime so comprehensive that it would have struck Bentham as excessive. Their keystrokes are logged. Their purchases are recorded. Their movements are tracked. Their productivity is scored by algorithms they cannot see and cannot appeal.

Until now, the inversion at least had a theoretical escape hatch. You could delete your social media accounts. You could pay with cash. You could choose, at some cost to convenience, to be less watched. The infrastructure being assembled in 2026 is closing that exit.

A company called Clearview AI has scraped more than fifty billion images from the public internet, social media posts, news articles, and public records, and built a facial recognition database that law enforcement agencies across the United States now use to identify suspects in seconds. The US Army signed a contract in 2026 giving its Special Forces access to the system. Immigration and Customs Enforcement holds a separate contract worth over nine million dollars. You did not consent to being in this database. You were not asked. You cannot delete your face. The Burgundian peasant’s anonymity was guaranteed by the technological limitations of his era. Yours was guaranteed by nothing, and it is already gone.

Cash, the last anonymous transaction mechanism available to ordinary people, is disappearing. In the United States, cash’s share of transactions fell to fourteen per cent in 2024 and is projected to reach seven per cent by 2030. In Sweden, fewer than two per cent of transactions involve physical money. The trend has a direction and it does not reverse. Every government that has studied the question has concluded that digital payments are cheaper to administer, easier to tax, and simpler to monitor. The Burgundian peasant could buy bread without generating a data point. Within a decade, that sentence may describe no living person on earth.

And then there is the infrastructure that replaces cash entirely. As of early 2026, one hundred and thirty-four countries representing ninety-eight per cent of global GDP are actively researching, developing, or piloting central bank digital currencies. China’s digital yuan has issued 1.8 billion wallets across twenty-six cities. The European Central Bank plans a digital euro pilot for 2027. Roughly forty per cent of the central bank digital currencies under development are designed to be programmable, meaning the issuing authority can attach conditions to the money itself: where it can be spent, what it can purchase, when it expires. A government that wanted to issue stimulus payments redeemable only at approved retailers, expiring after six months, spent only on approved categories of goods, could do so. The technology is not theoretical. It is being tested. The coin that the Burgundian peasant used had no memory. The money that may replace your banknotes will remember everything: where you spent it, when, on what, and whether the purchase was permitted.

The same convergence is happening with identity. Under the European Union’s eIDAS 2.0 regulation, every member state must offer a digital identity wallet to its citizens by December 2026. The wallet links a person’s government-issued identity to their driving licence, their diplomas, their bank accounts, their health records. The EU’s stated goal is eighty per cent adoption by 2030. The wallet is not mandatory in name. But when banks, employers, hospitals, and government services begin requiring it for authentication, the distinction between optional and mandatory will be academic. Participation in modern life will require a digital identity that is, by design, verifiable, traceable, and centrally legible. Eighty-five million cameras already monitor American streets, stores, and highways. The facial recognition market is projected to reach nine billion dollars this year. Taken together, the architecture describes a world in which your face is indexed, your money is programmable, your identity is digital, and cash, the last unmonitored medium of exchange, is approaching irrelevance. The Burgundian peasant could not be watched because the technology did not exist. You will not be able to avoid being watched because the technology will not permit it.

The privacy debate, as it is usually conducted, misses all of this. We argue about cookies and consent forms and data breaches as though the problem were technical, as though better regulation or smarter encryption could restore some prior balance. But there was no prior balance. There was a prior arrangement, and it was the opposite of what we have now. For ten thousand years, the powerful paid for their power with their privacy. The visibility of the ruler was the implicit contract that made rule legitimate. You could see your king. You could watch him dress. You could, at least in theory, hold him accountable by watching what he did.

We broke that contract, and we broke it in barely twenty years. The powerful became invisible. The powerless became transparent. And we called it convenience. The infrastructure now under construction does not merely continue the inversion. It makes it permanent. It encodes it into the money supply, into the identity system, into the geometry of public space. When cash is gone and your face is in a database and your identity is a digital credential issued by the state, the exit that the Burgundian peasant never needed and that you still theoretically possess will have been sealed, not by a decree but by an upgrade.

Three centuries and one complete inversion after the lever du roi, the Sun King’s bargain looks not like an archaic ritual but like a principle we discarded before we understood what it was for. Louis XIV let France watch him dress because the alternative, a king who operated in the dark, was a king who could not be trusted. The warehouse worker in Kentucky cannot watch anyone. She is watched by a system that has no face, no chamber to enter, and no curtain to draw back. The gaze still travels in one direction. It just no longer travels up. And the door it travels through is being locked from the other side.

Sources: Shannon Selin, “Watching French Kings Rise: The Grand Lever,” 2017. WebProNews, “Privacy Becomes Ultimate Luxury for Ultrawealthy Travelers in 2026.” Virginia Eubanks, Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor, St. Martin’s Press, 2018. Michael Fertik, “The Rich See a Different Internet Than the Poor,” Scientific American, 2013. Oxfam America, “Amazon and Walmart’s Excessive Warehouse Surveillance,” 2024. State of Surveillance, “Bossware 2026: 78% of Employers Spy on Workers.” ExpressVPN, “Workplace Surveillance Trends in the U.S. 2026.” Colonial Society of Massachusetts, “The Overseers and Their Functions.” Jeremy Bentham, Panopticon; or, The Inspection-House, 1791. Michel Foucault, Discipline and Punish, 1975. Proton, “What’s Your Data Really Worth?” 2025. Atlantic Council, Central Bank Digital Currency Tracker, 2026. Biometric Update, “Clearview Facial Recognition Searches Double, Database Reaches 50B Images,” 2024. Biometric Update, “US Army Renews Clearview AI Facial Recognition Contract for Special Operations,” 2026. EU Digital Identity Framework, eIDAS Regulation 2024/1183. Capital One Shopping, “U.S. Cashless Statistics,” 2026.

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI.

We don’t predict the future. We price it.

THE SCENARICA SUNDAY

On the morning of May 8th, inside the arena at the University of Central Florida, Gloria Caulfield stepped to the podium and made a mistake. Not a factual error. Not a political gaffe. She told several thousand humanities, journalism, and communications graduates the truth. “AI,” she said, “is the next industrial revolution.” The room did not politely disagree. The room erupted. Someone near the back yelled two words that do not need repeating and a wave of booing rolled forward through the seats until it hit the stage. Caulfield paused, smiled, and offered five words that will follow her for a long time: “Oh, I love it. Passion.” She then continued her speech for another eleven minutes, roughly a third of which she devoted to the topic her audience had already made clear they did not want to hear praised. The clip went viral within hours. Social media overwhelmingly sided with the graduates. It felt, to everyone watching, like a moment of collective power.

It was not.

What happened at UCF was the latest performance of an ancient ritual. The boo. The collective, synchronised, emotionally contagious expression of rejection that feels like resistance and functions as release. It is the sound a room makes when it has no other move. And across every recorded instance in modern history, from the Luddite frame-breakers of 1811 to the London taxi drivers who blockaded Trafalgar Square over Uber in 2014, the boo has never once stopped the thing being booed. Not delayed it. Not slowed it. Not altered its trajectory by a single degree. The boo is the most satisfying and least effective response in the human repertoire. Its futility is not a bug. It is the mechanism. Catharsis is the enemy of action, because it delivers the feeling of having responded without the substance of having responded at all.

The Class of 2026 has been booing with extraordinary consistency. The day after UCF, at Middle Tennessee State University, Big Machine Records founder Scott Borchetta told graduates that AI is rewriting music production as they sat there. The audience booed. Borchetta, who discovered Taylor Swift and has never been accused of excessive diplomacy, fired back with two words: “Deal with it.” A week later, former Google CEO Eric Schmidt took the stage at the University of Arizona and was booed so persistently that at one point he paused and asked the crowd to let him finish a sentence. A petition bearing more than 1,260 signatures had sought cancellation of his appearance. Student groups had coordinated a plan to turn their backs to the stage. The boos were loud, sustained, and sincere. Schmidt finished his speech. The crowd filed out. Nothing changed.

At Glendale Community College in Phoenix on May 15th, the irony arrived gift-wrapped. President Tiffany Hernandez announced that the ceremony would use an AI system to read graduates’ names aloud as they crossed the stage. The system mispronounced names, skipped students entirely, and froze mid-ceremony, forcing officials to pause the proceedings and bring in a human announcer. The students who had spent years being told not to use AI in their coursework watched an AI fail at the one task it had been given at their graduation. They booed. The boos were justified. The boos changed nothing about the world those students were about to enter.

Between March 1811 and February 1812, bands of English textile workers calling themselves Luddites destroyed approximately one thousand machines across the Midlands and the North. They smashed stocking frames in Nottinghamshire, shearing frames in Yorkshire, and steam-powered looms in Lancashire. The movement grew so threatening that the British government deployed twelve thousand troops to suppress it, a force larger than the army the Duke of Wellington had led into Portugal during the Peninsular War. At the York Assizes in January 1813, more than sixty men were tried. Twelve were hanged. Others were transported to penal colonies on the other side of the world. The violence was real. The courage was real. The sacrifice was enormous. And the machines were universally adopted. The Luddite riots did not delay mechanisation by a single day. The economic logic driving the technology was so overwhelming that no amount of collective fury could alter its path. The riot was the boo with hammers.

Two centuries later, London’s licensed taxi drivers produced their own version. In June 2014, thousands of black cabs blockaded Trafalgar Square, Whitehall, and the roads around Parliament in protest against Uber. The images were dramatic. Traffic stopped. The story made front pages across Europe. And in the weeks that followed, Uber’s London sign-ups surged. The app reportedly saw an 850 percent increase in downloads in the days after the protest. The taxi drivers had given Uber the single most effective marketing campaign in its history. The boo, amplified to the scale of a city, had accomplished the opposite of what it intended.

The reason the boo persists despite its perfect record of failure is that it satisfies a genuine psychological need. Collective booing is driven by emotional contagion, social pressure, and shared focus. When you boo in unison with a thousand other people, three things happen simultaneously. You feel that you are not alone. You feel that the group is powerful. And you feel that something meaningful has occurred. All three feelings are real. All three are also wrong in one critical respect. They register that the group has acted. They do not register that the group has accomplished anything. The sensation of collective power and the reality of collective power are, in this particular arena, inversely related. The louder the boo, the more satisfying the release, the less likely anyone in the room is to do the harder thing afterward.

The fear underneath the booing is real, and it is growing at a pace that should concern anyone paying attention. Monster’s 2026 State of the Graduate Report found that 89 percent of this year’s graduates believe AI could replace their jobs before they even get one, a twenty-five-point spike from the previous year. For the first time in the survey’s history, job security has overtaken career growth as the top factor driving career decisions. A Gallup panel survey of 1,572 young Americans conducted in February and March found that 31 percent of Gen Z now report outright anger toward AI, up nine points from a year earlier. A study published in Scientific Reports found that AI anxiety does not merely correlate with poor career decisions among college students but causes them, and the pathway through which anxiety undermines decision-making accounts for more than 63 percent of the total effect. The graduates are not hysterical. The data supports their fear. What the data does not support is the boo as a response to it. Erica Chenoweth’s celebrated research found that every protest movement mobilising at least 3.5 percent of a population succeeded, but her data concerns the overthrow of governments, not the halting of technologies. There is no equivalent dataset showing collective protest successfully stopping a technology that was already economically viable.

So the boo has a perfect record of futility, an established psychological explanation, and documented data confirming that the fear behind it is justified. The analysis could stop here. The commencement audiences of 2026 are performing an ancient ritual that has never worked. Case closed.

Except that something happened at Arizona State University on May 11th that complicates the story in a way worth paying attention to.

Harrison Ford stood before a different audience and said something different. He told graduates that their generation has far more power than they may realise. He said nothing about AI. He talked about vulnerability, about the climate, about the mess his generation was leaving behind, and about what the next generation might build in its place. He received a standing ovation. The speech went viral, not because Ford is a movie star, though that did not hurt, but because it stood in sharp relief against every other commencement address that month.

The difference was not charisma. The difference was grammar. Specifically, the difference was who occupied the subject position in each sentence.

Caulfield’s formulation was: AI is the next industrial revolution. The subject is the technology. The audience is the object, acted upon. Borchetta’s was: AI is rewriting music production. Again, the technology acts. The graduates receive. Schmidt’s entire address was structured around what AI will do, what AI will change, what AI will make possible. In every case, the speaker placed the technology in the driver’s seat and the graduates in the passenger seat. The message, stripped to its grammatical bones, was: something is happening to you.

Ford’s formulation was: you have far more power than you realise. The subject is the audience. The technology does not appear. The message, stripped to the same bones, was: you are the thing that happens.

This is not a stylistic quibble. It is the entire explanation. The boo was never a response to artificial intelligence. It was a response to a particular sentence structure, one in which the future arrives as a force and the listener exists only as a surface it strikes. Every commencement speaker who was booed this month delivered the same underlying message: the future is here, and it does not need you. Every speaker who was applauded delivered the opposite: the future is unmade, and it cannot be built without you. The audiences were not rejecting technology. They were rejecting a casting decision. They had been assigned the role of bystander in a story about their own lives, and they used the only tool available to refuse the part.

This distinction matters because it reveals something important about the relationship between inevitability and agency that the simple “booing doesn’t work” analysis misses. The speakers who were booed were, in narrow factual terms, correct. AI is reshaping industries. It will continue to do so regardless of audience sentiment. The graduates cannot stop it by booing. All of this is true. But being correct about the trajectory of a technology and being useful to the people affected by it are entirely separate achievements, and the commencement speakers of 2026 accomplished only the first.

What Ford understood, whether by instinct or by craft, is that a room full of people who feel powerless will reject even accurate information if it is delivered in a frame that confirms their powerlessness. The information itself is not the problem. The frame is. And the frame the AI speakers chose, inevitability as the headline with compliance as the implied instruction, is precisely the frame most likely to produce a boo. Not because the audience is naive. Because the audience is paying attention. They heard the sentence structure. They understood what role they had been assigned. And they refused it in the only way the format of a commencement ceremony permits.

The most dangerous property of the boo is still that it feels like success while accomplishing nothing. That has not changed. But the more interesting finding is that the boo is not the audience’s fault. It is the speaker’s. Every commencement address that was booed this spring was booed because the speaker described a future in which the audience had no agency, no role, and no recourse except acceptance. The speakers were not wrong about the technology. They were wrong about the audience. They mistook a room full of frightened, intelligent, capable people for a room full of passengers, and they wrote their speeches accordingly. The audience corrected them. The correction was loud. The correction was cathartic. The correction was, as always, useless. But the impulse behind it was entirely sound.

The Class of 2026 will enter a workforce where AI is already embedded in the tools they use, the companies they apply to, and the industries they hope to join. The question confronting them is not whether they accept this. Acceptance is irrelevant. The technology does not require their consent. The question is what they do with the energy that produced those boos. Every generation that has booed the future has spent emotional resources on an act that left the future unchanged and left the generation slightly more exhausted, slightly more resigned, and significantly less prepared for what came next. The graduates who booed at UCF, at MTSU, at Arizona, at Glendale were performing an ancient and deeply human ritual. They were flinching. And flinching feels protective. But it has never, not once in the historical record, protected anyone from the thing that made them flinch. Somewhere in those auditoriums, there were graduates who felt the same fear and did not boo. Not because they lacked passion. Not because they agreed with the speaker. Because they were already thinking about what to build.

Sources: 404 Media, “Students Boo Commencement Speaker After She Calls AI the ‘Next Industrial Revolution,’” 2026. LiveNOW from FOX, “CEO defends AI after graduates boo commencement speech,” 2026. NBC News, “Former Google CEO Eric Schmidt booed during graduation speech about AI,” 2026. NBC News, “Arizona college skips several graduates after an AI malfunction at commencement ceremony,” 2026. ASU News, “Harrison Ford tells ASU’s Class of 2026: ‘This is your time,’” 2026. Monster, “2026 State of the Graduate Report,” 2026. Gallup and Walton Family Foundation, “Gen Z’s AI Adoption Steady, but Skepticism Climbs,” 2026. Duan, Li, and Chen, “The impact of AI anxiety on career decisions of college students,” Scientific Reports, 2026. Chenoweth and Stephan, “Why Civil Resistance Works,” Columbia University Press, 2011. Hobsbawm, “The Machine Breakers,” Past and Present, 1952. Smithsonian Magazine, “What the Luddites Really Fought Against,” 2011.

“The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com”

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI.

“We don’t predict the future. We price it.”

THE SCENARICA SUNDAY - SPECIAL EDITION

A Sunday bonus alongside today’s regular Scenarica piece, published because of what Xi Jinping said in Beijing on Thursday at one of the most geopolitically turbulent moments of our lifetimes. Some sentences are too important to absorb at the news cycle’s pace.

Thursday morning. The Great Hall of the People. Twenty-one guns fire on Tiananmen Square. The two most powerful men on earth walk a red carpet, review the People’s Liberation Army honour guard, and step inside. Cai Qi, Wang Yi, He Lifeng line the Chinese side of the table. The American CEOs Trump has brought as a kind of merchant procession line the other.

And then, before the trade folders open, before Taiwan is named, before the photographs are sorted for the evening papers, Xi Jinping pauses. He looks at Donald Trump and he reaches past him.

He reaches across the table, past the cameras, past 2026, and he picks up a phrase Athenian generals used 2,400 years ago to describe a mechanism that destroyed them. “Can China and the United States,” he asks, “overcome the Thucydides Trap and create a new paradigm of major-country relations?”

It sounds like a question. It is not a question. It is a mirror.

This piece is not about China. This piece is about the mechanism Xi held up to Trump on Thursday morning, and the way that same mechanism is, almost certainly, sitting in your life right now.

Some history, fast.

In the fifth century before Christ, two Greek city-states discovered they could no longer share a peninsula. Athens, a brilliant maritime democracy, was rising. Sparta, a brutal land power, was the established hegemon. They went to war for 27 years. The war ended the Greek classical age. Plato grew up inside it.

A general named Thucydides was exiled mid-war, and used his exile to write the war’s history. He was not a court chronicler. He was the first writer in the Western tradition to ask, in a serious adult voice, why this kind of thing keeps happening. He thought the answer mattered enough that he wrote it down in a sentence that survives him by two and a half millennia:

“It was the rise of Athens and the fear that this inspired in Sparta that made war inevitable.”

Read the sentence once and you think it is about Athens. Read it twice and you notice something strange. The verb does not sit with the rising power. The verb sits with the fear of the established one. Thucydides puts the cause of the war not in the upstart’s growth, but in the incumbent’s response to that growth.

This is what the modern world now calls the Thucydides Trap, a phrase coined by Harvard’s Graham Allison about fifteen years ago. Allison ran the numbers. In sixteen historical cases of a rising power challenging an established one, twelve ended in war. The trap is not destiny. But it is gravity.

And here is what most commentators on Thursday’s summit will miss.

The Trap is not really about the rising power. It is about the established one. It is about Sparta. It is about what happens inside the older, slower, more comfortable polity when it notices that the metabolism of its rival has surpassed its own. The trap is not the rise. The trap is the fear. And the fear is not really fear. It is wounded self-image discovering that it can no longer afford itself.

Read Thucydides closely and you find that the Spartans do not decide to go to war because they have done the maths. They decide to go to war because their assembly is full of men who can no longer tolerate the sentence “Athens is doing things we cannot do.” The grown-up Spartans, the ones who counsel patience, are voted down. The wounded ones win the vote. The wounded ones almost always win the vote.

What Thucydides identified, in other words, is not a geopolitical mechanism. It is a psychological one. It is a mechanism that lives inside any person, any institution, any civilisation that used to be the one others measured themselves against. The moment the measuring stops going one way, something in the older party convulses. The convulsion can be productive, Sparta, in theory, could have reformed itself. The convulsion can also be ruinous, Sparta, in fact, lit the world on fire to avoid looking in a mirror.

Xi knows all of this. Of course he knows it. The Thucydides Trap has been a fixture of Chinese policy speeches for fifteen years.

What was different on Thursday was the timing. He did not raise it after Taiwan. He raised it before. He raised it as the opening frame of the entire summit. He raised it in front of the American CEOs Trump had paraded in. And he raised it in front of his own translators, his own state media, his own Politburo Standing Committee, who would carry it home.

Xi was not asking a question. He was placing the mirror on the table and letting the room watch Trump decide whether to look into it.

That is what this piece is about. The mirror is not just on the table in Beijing.

I want you to think of a person.

Not a country. A person. Someone in your life who, in the last twelve months, did something you could not have done. Maybe they are younger than you. Maybe they report to you. Maybe they are your child. Maybe they used to be your protégé and quietly stopped being your protégé. You will know who I mean within five seconds, because the body remembers these things before the mind admits them.

Sit with their face for a moment. The mechanism Thucydides described is in the room with you right now.

Here is what it looks like in life.

You hire a 28-year-old. She is faster than you were at 28. She is faster than you are at 41. Her work product, even the rough drafts, has a quality you reach only on your good weeks. The first time you notice this, you feel something that does not have a clean name. It is not exactly jealousy. It is not exactly admiration. It is closer to a small private grief that you immediately reinterpret as a concern about her judgement.

You start to keep a quiet ledger of her mistakes. They are real mistakes. That is not the point. The point is that you have never kept a ledger like this before. You did not keep one for your peers. You do not keep one for your boss. You keep one for her.

In meetings, you find yourself adding caveats to her ideas before others can respond. You do not call it diminishing her. You call it stewarding the room. You tell yourself she needs the polish. She does not need the polish. She needs you to step out of the doorway. The Spartans called that strategic patience too.

One evening you draft an email to your manager. The subject line is something like “Concerns re: K’s growth trajectory.” You sleep on it. In the morning you read it back and notice that every single one of the concerns is, in fact, a place where her instinct was better than yours. You do not send the email. But you also do not delete the draft.

That is Sparta in your inbox.

Or it is a parent. You are 62. Your daughter is 31 and is about to take a job in a city you would never have moved to, for a salary you would never have asked for, with a boss you would never have chosen, doing a kind of work that did not exist when you were her age. You tell yourself you are worried about her. You make a list of the risks. The list is real and the risks are real.

But underneath the list there is something the list is doing for you. The list is letting you remain the person in the room who knows things. The list is keeping the metabolism of the relationship pointed in the direction you understand. The moment she takes the job, the metabolism reverses. She will be the one who knows things. She will be the one calling to explain a world you only half-recognise.

Some parents, faced with this, become advisors. Some parents, faced with this, become Spartans. The difference between the two is almost entirely whether they can let the mirror in the daughter’s life reflect something back without taking it as an attack.

Or it is a marriage. There is a quiet thing that happens in long marriages around year fourteen, year fifteen, year sixteen, different for everyone, but it happens. One partner, often the one who carried the household earlier, looks up and notices that the other partner has grown in a direction they did not authorise. New friends. New competence. New centre of gravity. The first partner has two options. Become curious. Or start cataloguing the other partner’s mistakes.

If you have ever been on either side of that ledger, you know exactly what Thucydides was writing about. You did not need the Greek.

Or it is a midlife reckoning with no other person in it at all. You walk into a room, a conference, a dinner, a client meeting, and you realise, with a small clean ping of horror, that you are no longer the smartest person there. Not by far. There are people fifteen years your junior who are reading papers you have not yet heard of, building things you do not understand the architecture of, holding the future in their heads in a way you used to hold it in yours.

You have, in that instant, a choice that the room cannot see you make. You can be Athens, alert, learning, slightly embarrassed, asking the next question. Or you can be Sparta, wounded, watchful, cataloguing the room’s mistakes on your phone under the table, waiting for the moment to remind everyone of the war you won fifteen years ago.

Almost everyone, the first few times, picks Sparta. The body picks it before the mind has a vote.

This is the deeper claim of Thucydides, the one most commentators on the summit will miss this weekend. The Trap is not a story about who wins. The Trap is a story about who can stand to be measured.

You have lived a version of that sentence this week. You may have lived more than one.

Now look up.

What Xi was doing on Thursday morning, with the cameras in the Great Hall and the CEOs in their rows, was not asking America for a relationship. He was asking America whether it could stand to be measured. He was asking it in front of a room full of American capital, which made the asking itself a status move. He was placing the mirror on the table and saying, calmly and in formal diplomatic register, your move.

The painful thing for an American reader is that the measurements have already happened.

BYD now launches new electric vehicles on a cadence American automakers cannot match. The gap is not a tariff gap. It is a metabolic gap, the time from concept to production line is measured in months in Shenzhen and in years in Detroit, and the gap is widening every quarter. No tariff fixes a metabolism.

DeepSeek’s V4 reasoning model, released earlier this year, hit GPT-class benchmarks on a fraction of the inference compute the American labs use. The chip export controls, the most expensive industrial policy the United States has run in a generation, were specifically designed to make this impossible. It happened anyway. It happened because the constraint forced an efficiency the American labs, swimming in capital and chips, had no reason to find. Scarcity is a strange and impolite teacher and Sparta hates it.

China is now commissioning grid-scale battery storage at a pace the United States has not approached this decade. The numbers are not contested anymore. They are in the IEA’s tracker, and American utility executives have been quietly forwarding them to each other for months.

Pick any of these and the American reader has the same two options the parent has, the manager has, the spouse has, the wounded Spartan in the assembly has. You can be curious. Or you can catalogue the mistakes.

The American conversation, in May 2026, has chosen to catalogue the mistakes. We talk about subsidies. We talk about forced technology transfer. We talk about Uyghurs and Hong Kong, which are real and which also do not explain BYD’s launch cadence. We talk about how the data is faked, which is sometimes true and which also does not explain DeepSeek V4 sitting in our own labs’ hands. We talk, in the most revealing tic of all, about how the system cannot last, which is the same sentence the Spartans told themselves about Athenian democracy for thirty years before Athens beat them at sea.

This is what Xi was doing on Thursday. He was not threatening Trump. He was not bluffing Trump. He was, in the most precise possible diplomatic language, holding up the mirror and saying: we both know who is afraid in this room, and it is not the rising power.

Trump’s response, posted to Truth Social before dawn on Friday, was to read Xi’s reference as a comment on American decline and assign it to the Biden years. The response is not embarrassing because it is partisan. The response is embarrassing because it is the exact response Thucydides predicted Sparta would have. It is the response of an established power that cannot yet bear to look at its own metabolism, so it externalises the slowdown onto a previous administration, a foreign system, a bad decade, anything that lets the older self stay the protagonist.

This piece is not arguing that America is finished. America is not finished. Athens was not even the right side of that war; Sparta won. The argument is narrower and harder. The argument is that you cannot reform what you will not look at. Sparta could not look. Sparta lost a hundred and fifty years of cultural primacy and then itself. The American question, the actual one underneath the summit, is whether the United States in 2026 can do something Sparta in 431 BC could not.

It is a question about character. Not policy. Character.

And here, if you have been honest with yourself for the last ten minutes, you already know how the question is answered. It is answered the same way it gets answered in the email about your 28-year-old colleague. It is answered in the way you talk to your adult daughter on the phone. It is answered in whether you walked into that conference and asked the next question, or whether you went home and told your spouse the room was full of fools.

The civilisation answers the way the people in it answer.

Back to the Great Hall.

Xi asked Trump three questions on Thursday morning, not one. The famous one was about the Thucydides Trap. The second was whether the two countries can meet global challenges together. The third was whether they can build a bright future for the well-being of their two peoples and the future of humanity.

Those three questions, read out loud in the order he asked them, form a kind of ladder. The first is whether America can stand to be measured. The second is whether America can act after being measured. The third is whether America can imagine a future in which it is not the only measurer. Each rung is harder than the one below. Each rung requires a kind of self-possession the wounded self does not have.

The deepest question, the one Xi did not ask out loud because he did not need to, is the one the rest of us also face. It is the question the older colleague faces in front of the 28-year-old’s better draft. The question the 62-year-old parent faces on the phone with their daughter. The question every long-married partner faces when the other has grown sideways. The question every midlife professional faces in the conference room where they are no longer the smartest.

It is the question of which polity, and which person, can hold up the mirror without breaking it.

That is the only question Thursday actually asked. It is the only one this Sunday actually answers.

Source: PRC Ministry of Foreign Affairs official readout, “President Xi Jinping Holds Talks with U.S. President Donald J. Trump,” 14 May 2026 — https://www.fmprc.gov.cn/eng/xw/zyxw/202605/t20260514_11910330.html

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

In the mid-2000s, a young researcher named Celeste Kidd was volunteering at a homeless shelter for families in Santa Ana, California. The shelter was a single large room, shared by everyone. Possessions were not safe. When one child received a toy or a treat, there was a real chance a bigger, faster child would take it. Kidd watched the children week after week and began to notice something that would eventually upend one of the most celebrated findings in the history of psychology. These children did not wait for good things. They grabbed what was in front of them, immediately, every time. And they were not being impulsive. They were being perfectly rational.

To understand why this matters, you need to know the experiment Kidd was quietly dismantling.

In the late 1960s, at the Bing Nursery School on Stanford’s campus, the psychologist Walter Mischel placed more than five hundred children, one at a time, in a small room with a single marshmallow. The offer was simple: eat it now, or wait until the researcher returns, and get two. Some children ate it within seconds. Others developed elaborate strategies, covering their eyes, singing songs, pushing the marshmallow to the far edge of the table. A few sat perfectly still for the full fifteen minutes, stone-faced and unblinking, as though engaged in a private war whose stakes only they understood.

The experiments were elegant. They were also, at that stage, unremarkable. Then, almost two decades later, the children came back. In 1990, Mischel and his colleagues tracked down 185 of the original preschoolers, now adolescents. The children who had waited longer scored significantly higher on the SAT. Their parents rated them as more competent, more capable of planning ahead. The children who had grabbed the marshmallow early were more likely to be stubborn, indecisive, and easily upset.

The narrative that emerged was irresistible, and the world consumed it whole. Self-control was destiny. The ability to delay gratification at age four was a character trait that separated the high achievers from the rest. Mischel published a bestseller in 2014. Angela Duckworth’s work on grit, which drew explicitly from the marshmallow tradition, became a TED talk viewed tens of millions of times. Schools designed curricula around delayed gratification. The test entered the language as shorthand: a marshmallow person was impulsive and unreliable, a waiter was disciplined and bound for success.

Nobody, during this long cultural moment, spent much time on a detail that would later prove critical: where the children came from.

The Bing Nursery School is not a typical preschool. It sits on the Stanford campus and has historically served the children of faculty, graduate students, and staff. The children Mischel tested were, overwhelmingly, the offspring of one of the most privileged academic communities on earth. They arrived in that small room having already spent four years in stable homes with educated parents, predictable routines, and adults who reliably kept their promises. When the researcher told them she would return with a second marshmallow, they had every reason to believe her.

The child in Kidd’s shelter had learned the opposite. Good things do not last. Promises are unreliable. Delay means loss.

Kidd, now at the University of Rochester, designed an experiment to test this. Before giving children the standard marshmallow task, she exposed them to one of two conditions. In the reliable condition, a researcher made a small promise and kept it. In the unreliable condition, the researcher made the same promise and broke it. The results, published in the journal Cognition in 2013, were stark. Children in the reliable condition waited an average of twelve minutes. Children in the unreliable condition waited three. One brief experience of broken trust was enough to collapse a child’s willingness to wait by a factor of four.

The marshmallow test was not measuring self-control. It was measuring trust. And trust is not a trait that sits inside the child. It is a product of the environment the child has already lived through.

The replications confirmed what Kidd’s shelter children had been demonstrating all along. In 2018, Tyler Watts, Greg Duncan, and Haonan Quan tested over nine hundred children from diverse socioeconomic backgrounds, not just Stanford faculty offspring. The correlation between waiting time and later achievement was roughly half the size Mischel had reported. When they controlled for family background, early cognitive ability, and home environment, the correlation shrank by two thirds. For behavioural outcomes, it vanished entirely. What looked like character was circumstance, measured at age four and mistaken for destiny.

Then came a finding that made the picture stranger still. In 2022, Yuko Munakata and colleagues ran the marshmallow test with children in both the United States and Japan, but offered half of them marshmallows and the other half wrapped gifts. Japanese children waited three times longer for food than for gifts. American children waited nearly four times longer for gifts than for food. In Japan, children are taught from toddlerhood to wait until everyone is served before eating. American children grow up in a culture where gifts are saved for specific occasions, where the ritual of unwrapping is governed by social rules about timing. Each group waited longest for the reward their culture had trained them to delay. Self-control had not varied. The rules of waiting had.

By 2024, Jessica Sperber, Watts, and Duncan had tracked 702 marshmallow test participants all the way to age 26. Marshmallow test performance was not strongly predictive of adult achievement, health, or behaviour. The grand narrative, built on fewer than two hundred children at one of the most unusual preschools in America, had not survived contact with a broader, longer, more honest look at the evidence.

Two children sit in two rooms. One has spent four years in a world where adults keep their promises. The other has spent four years learning that they do not. Both are given a marshmallow and told to wait. One waits. One grabs. We call the first one disciplined and the second one a failure, and we build an entire theory of success on the difference. But both children are running the same calculation. The only variable is the input data.

This is not a finding about children. It is a finding about systems.

In economics, the phenomenon has a name: time preference. People with fewer resources consistently discount the future more steeply. A study by Emily Lawrance in the Journal of Political Economy found that households in the bottom fifth percentile of income had subjective discount rates of 19%, compared to 12% for college-educated households in the top 5%. The poor were not less disciplined. They were responding to a different set of constraints. When your income is uncertain, your housing is unstable, and your savings buffer is thin, the rational response is to take the sure thing now. Waiting is a strategy that only pays off if the future is likely to be there when you arrive.

Sendhil Mullainathan and Eldar Shafir formalised this in their book Scarcity. Having too little, whether of money, time, or security, creates a cognitive state that narrows the mind’s horizon and compromises its bandwidth. Poverty does not merely reduce resources. It changes how people think. The mental functions we lazily call “self-control” and “long-term thinking” are not fixed capacities. They are cognitive modes that scarcity switches off. As the economist Esther Duflo put it: the richer you are, the less responsibility you need to take for your own life, because everything is taken care of for you. And the poorer you are, the more you have to be responsible for everything about your life.

Patience is not a character trait. It is a luxury good.

Once you see the pattern, it scales. Every business school teaches that great companies think long-term. Jeff Bezos built Amazon to lose money for years, investing relentlessly in infrastructure while deferring gratification on a timescale that made Wall Street physically uncomfortable. The narrative is always the same: visionary founder resists short-term pressure, proves the doubters wrong. But Bezos could afford to wait. His parents had invested $245,573 of family money in the founding. He held a Princeton degree and a Wall Street career to fall back on. Amazon operated without profit because it had access to patient capital that believed in the story. This is not a trivial detail. It is the entire mechanism.

A survey by Graham, Harvey, and Rajgopal found that almost 80% of CFOs would cut research and development spending, and more than 55% would delay starting a new project, to meet quarterly earnings targets. For a company whose credit facility depends on quarterly covenants, whose share price drop would trigger margin calls, and whose CEO’s compensation is tied to annual earnings per share, meeting the number is survival. The CFO cutting R&D to make the quarter is not lacking vision. She is eating the marshmallow because she has learned that the second one is not coming.

The pattern holds between nations. In 2022, Benjamin Enke and colleagues used time preference data from 76 countries. More patient countries were richer, had higher capital accumulation, and achieved greater productivity. The finding fed a familiar narrative: some cultures value the long term, and those cultures prosper. But countries with stable institutions, enforceable property rights, and reliable legal systems give their citizens reason to believe that investments made today will still be there tomorrow. A farmer in Denmark who plants a tree can be reasonably confident that his children will harvest the fruit. A farmer in a conflict zone where land tenure is contested and governments change by force has learned something different: take what you can, while you can. Both farmers are being rational. The difference is not culture. It is the room they go home to.

For thirty years, the marshmallow test told a comfortable story: that success is a function of personal character, that the people who fail simply lacked the discipline to wait. This story was comfortable because it located the explanation for inequality inside the individual. The test was a scientific alibi for a world that preferred to believe that outcomes reflect character rather than circumstance.

Every system that rewards patience without examining who can afford to be patient is reproducing the Mischel error. Every organisation that promotes the long-term thinker without asking what made long-term thinking possible is measuring privilege and calling it merit. The child who waits is not stronger. She is safer. The company that invests for the long term is not wiser. It is better capitalised. The country that builds institutions over decades is not more virtuous. It is more stable.

Before she ever designed an experiment, Celeste Kidd stood in a shelter in Santa Ana and watched a child receive a treat and eat it before anyone could take it away. That child had already learned everything she needed to know about waiting. She had learned that good things do not last, that promises are words people use before they leave the room, and that the only marshmallow worth anything is the one already in your hand.

The question that should keep us awake is simple: for how many people, companies, and countries have we prescribed patience while withholding the conditions that make patience possible?

We told them to wait. We just forgot to make sure it was worth it.

Sources: Walter Mischel, Ebbe Ebbesen, and Antonette Raskoff Zeiss, “Cognitive and Attentional Mechanisms in Delay of Gratification,” Journal of Personality and Social Psychology, 1972. Yuichi Shoda, Walter Mischel, and Philip K. Peake, “Predicting Adolescent Cognitive and Self-Regulatory Competencies from Preschool Delay of Gratification,” Developmental Psychology, 1990. Celeste Kidd, Holly Palmeri, and Richard Aslin, “Rational Snacking: Young Children’s Decision-Making on the Marshmallow Task Is Moderated by Beliefs About Environmental Reliability,” Cognition, 2013. Tyler Watts, Greg Duncan, and Haonan Quan, “Revisiting the Marshmallow Test: A Conceptual Replication Investigating Links Between Early Delay of Gratification and Later Outcomes,” Psychological Science, 2018. Jessica Sperber, Deborah Lowe Vandell, Greg Duncan, and Tyler Watts, “Delay of Gratification and Adult Outcomes: The Marshmallow Test Does Not Reliably Predict Adult Functioning,” Child Development, 2024. Kaichi Yanaoka, Satoru Saito, and Yuko Munakata, “Cultures Crossing: The Power of Habit in Delaying Gratification,” Psychological Science, 2022. Sendhil Mullainathan and Eldar Shafir, Scarcity: Why Having Too Little Means So Much, 2013. Emily Lawrance, “Poverty and the Rate of Time Preference: Evidence from Panel Data,” Journal of Political Economy, 1991. Benjamin Enke, Thomas Graeber, et al., “Patience and Comparative Development,” Review of Economic Studies, 2022. John Graham, Campbell Harvey, and Shiva Rajgopal, “The Economic Implications of Corporate Financial Reporting,” Journal of Accounting and Economics, 2005.

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

Two people read the same article on the same morning in March.

The first is a twenty-eight-year-old in San Francisco. He has a computer science degree from Stanford, eleven months of savings in an index fund, a network of friends who work at startups, and a mass of half-formed ideas about things he might build now that his previous company has been acqui-hired. He reads the article at a coffee shop. It tells him that most high-income skills will be irrelevant within a decade and that the future belongs to sovereign individuals who develop agency, taste, judgement, and deep generalism. He nods. He screenshots a paragraph and texts it to a friend. He opens a new Notion page and starts sketching.

The second is a forty-one-year-old woman in Dayton, Ohio. She has worked in the same call centre for six years. Last Tuesday, her manager gathered the floor and announced that the company was exploring AI-driven customer solutions and that some roles would be restructured in the third quarter. She does not have a computer science degree. She has a GED, a car payment, a twelve-year-old son whose school fees are two months behind, and four hundred and eighteen dollars in a savings account she tries not to touch. She reads the same article on her phone during her fifteen-minute break. It tells her to develop agency, build a personal brand, and start a one-person business. She closes the tab. She goes back to the phones.

They both understood the article. Only one of them can use it.

This is not an essay about whether AI will displace jobs. It will. Goldman Sachs estimates that three hundred million full-time positions worldwide will be affected by generative AI. McKinsey’s research suggests that current technology could automate roughly fifty-seven per cent of all work hours in the American economy. Anthropic’s own CEO has said that fifty per cent of entry-level white-collar jobs could disappear within five years. The displacement is not a debate. It is a schedule.

And across that landscape, a new genre has emerged. Call it the AI survival guide. The posts go viral because they speak directly to the fear. They validate the anxiety, then hand the reader a to-do list. Learn to sell. Build a brand. Develop taste. Cultivate agency. Become sovereign. The pieces are well-written, well-intentioned, and in many cases substantially correct about which skills will matter in an automated economy. One such article, published in March on Substack, gathered nearly two thousand likes and over a hundred comments. It resonated because it told frightened people that there is something they can do. It gave them a map.

But a map is only useful if you can get to the trailhead. And nobody is talking about the distance between the car park and the mountain.

The word that appears most often in the AI survival literature is sovereignty. It is a beautiful word. It implies autonomy, self-direction, mastery. In the framework these articles describe, the future splits neatly into two groups. Group one: sovereign individuals who set their own goals, solve their own problems, and ride AI to extraordinary productivity. Group two: everyone else, dependent on whatever universal basic income or government programme replaces the job that used to give their life structure.

The survival guides are written for group one. They assume group one as a starting point. And the gap between that assumption and the lives of most of their readers is where the advice quietly breaks apart.

Take the first skill: agency.

Develop agency, the articles say. Learn to identify problems, build solutions, earn from them. Agency, in this framing, is a character trait. Something internal. A muscle you train by deciding to train it.

But here is what the character-trait framing leaves out. The Ewing Marion Kauffman Foundation, the largest American foundation devoted to entrepreneurship research, has spent decades studying who actually starts businesses. Their findings are consistent and uncomfortable. More than ninety per cent of entrepreneurs come from middle-class or upper-middle-class backgrounds. Ninety-five per cent have bachelor’s degrees. Nearly half have advanced degrees. Seventy-eight per cent of new businesses are funded by personal savings. The average cost to launch a small business in the United States is thirty thousand dollars.

Now hold those numbers against this one: according to the Federal Reserve’s most recent survey, published in May 2025, thirty-seven per cent of American adults cannot cover a four-hundred-dollar emergency expense from cash or its equivalent. Not four hundred thousand. Four hundred.

Agency is not a personality trait. Agency is the surplus capacity to take a risk and survive the consequences. It requires savings that can absorb six months of failed experiments. It requires a network that can make introductions when you are starting from zero. It requires the education to recognise which problems are worth solving and the social capital to be taken seriously when you propose a solution. It requires, above all, a floor beneath you solid enough that falling does not mean losing your apartment, your health insurance, or your ability to feed your children.

Telling someone with four hundred dollars in the bank to develop agency is like telling someone in a wheelchair to develop the skill of climbing stairs. The destination is not wrong. The prescription is missing every structural prerequisite that makes arrival possible.

This is the first half of the counterpoint. The survival guides describe what sovereignty looks like. They do not describe what it is made of. And what it is made of is not psychology. It is infrastructure.

Sovereignty is infrastructure, not psychology.

It is a savings account with enough months in it to let you fail. It is health insurance that does not vanish when you leave your employer. It is a degree or credential that lets you walk into rooms where decisions are made. It is a family that can lend you money without going broke themselves. It is a postcode where the school your child attends does not consume every ounce of your remaining cognitive bandwidth. It is the accumulated, compounding privilege of having had enough stability, for long enough, to develop the habits of mind that the survival guides describe as skills.

Raj Chetty, the Harvard economist whose work on intergenerational mobility has reshaped how serious people think about opportunity in America, has mapped this with devastating precision. A child born in the bottom fifth of the income distribution in Charlotte, North Carolina has a 4.4 per cent chance of reaching the top fifth. In San Jose, the odds are nearly three times higher. Not because San Jose children are more ambitious or more talented or more sovereign. Because the infrastructure beneath them is different. The schools are better funded. The networks are denser. The safety nets are stronger. The distance from the car park to the trailhead is shorter. That is not a mindset gap. It is a concrete gap, built from decades of policy, investment, geography, and compounding advantage.

Now consider the second skill the survival guides recommend: taste and perspective.

Develop taste, they say. Your unique perspective is your unfair advantage. In a world where everyone has access to the same AI models, the person who knows what to build and why will win.

This is true. It is also the most class-encoded piece of advice in the entire genre. Taste is not innate. It is cultivated through exposure. Exposure to ideas, to travel, to different industries, to the kind of wide reading and cross-disciplinary conversation that happens naturally in university common rooms and startup incubators and nowhere else unless you can afford the time and access to seek it out. The person who has spent six years answering calls in Dayton has perspective, certainly. She has deep knowledge of what it feels like to be on the receiving end of institutional indifference. But that is not the kind of perspective the survival guides mean. They mean the perspective of someone who has read widely enough, failed safely enough, and been exposed to enough different worlds to synthesise ideas across domains. That kind of taste takes years of investment that looks, from the outside, like leisure. It is reading books that have no immediate professional application. It is taking a course in something unrelated to your job. It is spending six months on a project that might not work. Every one of those activities requires time that is not being spent earning rent.

Deep generalism, the fourth skill in the standard list, runs into the same wall. Broaden your areas of interest, the advice says. Read more, learn more, experiment more, build more. The instruction is sound. But broadening your interests requires first securing your base. Maslow understood this seventy years ago. You cannot contemplate self-actualisation when you are worried about the electricity bill. The hierarchy of needs is not a metaphor. It is a description of how human cognition actually allocates resources under scarcity. A brain consumed by financial survival does not have spare capacity for cross-domain synthesis. It is not choosing not to be generalist. It is rationing.

And then there is the most common piece of advice in the entire AI survival canon: start a business.

Start a one-person business, the articles say. AI gives you infinite leverage. Sam Altman himself has predicted that the first one-person billion-dollar company is coming. The future belongs to micro-entrepreneurs who use AI agents to do the work of entire teams.

Sam Altman made that prediction in a group chat with his tech CEO friends. He has a betting pool on it. This is a man who went from Stanford to Y Combinator to the helm of OpenAI, who has never once in his adult life wondered whether he could make rent on the first of the month. The prediction may well be correct. It is also made from a position so structurally privileged that it cannot see the floor it is standing on.

Because starting a business, even a one-person business with AI leverage, requires the ability to survive the gap between launch and revenue. The median time for a new business to become profitable is twelve to eighteen months. The Kauffman Foundation data is clear: the vast majority of startup capital comes from personal savings and family networks. If you do not have savings and your family does not have capital, you do not have a business opportunity. You have an idea and no runway.

Eighty-two per cent of bootstrapped startups fail. The number one reason, across every study, is running out of cash. Not running out of ideas. Not running out of agency. Running out of the money that keeps the lights on while the agency develops.

There is a phrase the survival guides love: skin in the game. It is borrowed from Nassim Taleb, and in his formulation it means accepting downside risk as a condition of participation. It is a fine principle for people whose downside is a dent in their savings. It is a catastrophic principle for people whose downside is homelessness.

This is not an argument against the advice. Let me be precise about that.

The four skills are real. Agency matters. Taste matters. Judgement matters. Generalism matters. The survival guides are not wrong about what the future rewards. They are wrong about who can hear the advice and act on it. They describe the summit. They do not describe the cliff between the reader and the first foothold.

The articles are written by people who became sovereign and then looked backwards and described what they saw. This is natural. It is what humans do. We narrate our own success as a sequence of decisions and skills, because the alternative is to admit that the infrastructure we stood on was doing most of the work. The Stanford grad in San Francisco does not experience his savings, his network, and his degree as structural advantages. He experiences them as the background of his life. They are invisible to him in the way that water is invisible to a fish. When he writes an article telling other people to develop agency, he means it sincerely. He just does not realise he is describing the view from a platform that most of his readers cannot reach.

And this brings us to the deepest problem with the sovereignty frame. It is not just that the advice only works for the privileged. It is that the advice, by framing sovereignty as a psychological achievement rather than a structural one, implicitly blames the people who cannot follow it. If agency is a skill, then lacking it is a failure. If sovereignty is a mindset, then dependence is a character flaw. The framework that is meant to empower people ends up doing something more insidious: it tells them that their inability to thrive in an AI economy is their own fault. You had the same article. You had the same advice. You just did not execute.

Develop high agency is not advice. It is a description of the outcome, presented as a prescription.

There is an older version of this trick, and it is worth naming. In the early 2010s, the standard response to any anxiety about economic displacement was learn to code. It had the same structure. A real skill, genuinely valuable, offered as universal advice without acknowledging that access to coding bootcamps required tuition, time, a laptop, broadband, and the cognitive bandwidth that comes from not being in financial crisis. Learn to code was correct about the skill and wrong about the audience. Become sovereign is its direct descendant. The skill has changed. The structural blindness has not.

This matters now more than it has ever mattered, because the displacement that AI is producing has a specific and vicious geometry. Brookings Institute research shows that AI disproportionately benefits high-skill, high-income workers while reducing opportunities for lower-skill workers. The productivity gains flow upward. The job losses concentrate downward. The people best positioned to ride the AI wave are the people who were already riding every previous wave, because they have the savings, the education, and the network to adapt. The people worst positioned are the ones the survival guides are ostensibly written for: the anxious, the uncertain, the person scrolling Substack at midnight wondering what to do.

You can automate production completely. You cannot automate demand.

This is the sentence the survival guides never write. AI can build the product. It can write the code. It can generate the content and design the interface and optimise the funnel. What it cannot do is make anyone care. Demand is a human phenomenon. It is built on trust, on relationships, on the feeling that the person selling you something understands your life. And the infrastructure that creates demand, the audience, the brand, the network of trust, is precisely the infrastructure that takes time, money, and stability to build.

The one-person billion-dollar company, if it arrives, will not be built by someone who decided to become sovereign. It will be built by someone who already was. Someone with the savings to survive three years of building. Someone with the network to get the first hundred users. Someone with the education to understand the problem deeply enough to solve it in a way that AI alone cannot. The company will look, from the outside, like a triumph of individual agency. From the inside, it will be a triumph of accumulated structural advantage, amplified by technology.

So what is the honest advice?

It is not that the four skills do not matter. It is that they cannot be developed in a vacuum. They require a floor. And for millions of people, the floor is missing. The honest conversation about AI displacement is not a conversation about skills. It is a conversation about infrastructure. About savings. About healthcare that is not tied to employment. About education that does not leave you with six figures of debt before you have earned a dollar. About childcare that does not consume half your income. About the basic, unglamorous, policy-level scaffolding that makes it possible for a human being to take a risk.

Finland ran a universal basic income experiment from 2017 to 2018. The results were not dramatic. Employment rates barely changed. But something else did. The participants reported lower stress, better health, and more confidence in their future. They were not transformed into entrepreneurs overnight. They were given the breathing room to think clearly. That breathing room, the cognitive space that opens when you are not one missed paycheck from catastrophe, is the actual prerequisite for every skill the survival guides describe.

The Stockton SEED programme in California gave a hundred and twenty-five residents five hundred dollars a month. The majority used it to buy groceries and pay bills. It did not turn them into sovereign individuals. It turned them into people who could think about something other than whether they could make it to the end of the month. That, not a podcast about agency, is where sovereignty begins.

Here is what is coming.

Within three years, the entry-level white-collar jobs that currently employ millions of people will thin dramatically. Customer service, data entry, paralegal research, basic financial analysis, content production at scale. The people in those jobs will be told to adapt. They will be shown articles about the four skills. They will be told that the future belongs to the sovereign, the agentic, the tasteful, the generalist. And most of them will be unable to act on any of it, because they will be trying to pay rent.

The survival guides will get more popular as the displacement accelerates. They will become a genre unto themselves, a vast and growing literature of advice that is technically correct and structurally useless for the majority of its audience. And the gap between the people who can follow the advice and the people who cannot will widen, because AI’s productivity gains compound for those who can invest in them and pass over those who cannot.

This is not a story about AI. It is a story about floors.

The person in San Francisco has a floor. It is made of savings, education, networks, family money, health insurance, and the kind of deep, inherited stability that lets you take a risk and know that failure means starting again, not falling through. He will read the survival guides and build something. He will attribute his success to agency, taste, and judgement. He will be right about the skills and wrong about why he had them.

The woman in Dayton has no floor. She has a ceiling that is getting lower. She will read the same survival guides and feel a mixture of inspiration and quiet despair, because she can see the summit and she can see the cliff between her and the first foothold, and no amount of motivation can bridge a gap that is made of money she does not have.

The honest version of the AI survival guide would start with a question that none of them ask: what is beneath your feet right now? Because the skills they describe are real, and the future they predict is plausible, and the advice they offer is sound for people who can take it. But the distance between hearing advice and acting on it is not measured in motivation. It is measured in months of savings, in the quality of schools within driving distance, in whether losing your job means a career pivot or a housing crisis. It is measured in infrastructure.

And until someone builds that into the conversation, the survival guides will keep going viral, and the people who need them most will keep closing the tab and going back to the phones.

Sources:

Goldman Sachs, “The Potentially Large Effects of Artificial Intelligence on Economic Growth,” March 2023

McKinsey Global Institute, “A New Future of Work: The Race to Deploy AI and Raise Skills in Europe and Beyond,” 2024

Federal Reserve Board, “Report on the Economic Well-Being of U.S. Households in 2024,” May 2025

Ewing Marion Kauffman Foundation, “The Anatomy of an Entrepreneur” and “Who Is the Entrepreneur?” research series

Raj Chetty, Nathaniel Hendren, Patrick Kline, and Emmanuel Saez, “Where Is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States,” Quarterly Journal of Economics, 2014

Brookings Institution, “AI’s Impact on Income Inequality in the US,” 2024

Kela (Social Insurance Institution of Finland), “Results of Finland’s Basic Income Experiment 2017-2018,” 2020

Stockton Economic Empowerment Demonstration (SEED), “SEED’s First Year: Results from a Guaranteed Income Pilot,” 2021

Abraham Maslow, “A Theory of Human Motivation,” Psychological Review, 1943

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

In July of 1844, Charles Darwin finished 230 pages that would change human civilisation. The essay laid out the theory of natural selection in meticulous detail: the mechanism of variation, the pressure of competition, the slow accumulation of advantage over generations. It contained every essential argument that would appear, fifteen years later, in On the Origin of Species. He had been developing the idea since his return from the Beagle voyage eight years earlier, filling notebook after notebook with observations, counterarguments, and evidence that pointed, with increasing force, toward one conclusion. The essay was complete. The argument was sound. And Darwin, rather than publish it, wrote a letter to his wife Emma asking her to arrange its publication in the event of his death (to spend four hundred pounds on an editor, to ask Hooker or Lyell to supervise), sealed both documents together, and placed them in a drawer.

He did not open that drawer for fifteen years.

He classified barnacles instead. He classified them so exhaustively, for so many years, that his children grew up assuming this was simply what fathers did. One of his sons, visiting a friend’s house, reportedly asked where the friend’s father did his barnacles. Darwin bred pigeons. He corresponded with botanists about seed dispersal, with pigeon fanciers about plumage variation, with entomologists about beetles. He accumulated evidence for a theory that was already, by his own private admission, ready. And then, in June of 1858, a letter arrived from a young naturalist named Alfred Russel Wallace, written from the island of Ternate in the Malay Archipelago. It contained a short essay describing a theory of evolution by natural selection that was, in its essentials, identical to the one sitting in Darwin’s drawer. Darwin wrote to his friend Charles Lyell in anguish: Wallace could have written a better summary of Darwin’s own theory than Darwin could himself. The joint presentation to the Linnean Society was arranged within days.

Fifteen years. Fifteen years of barnacles and pigeons and letters about beetles. And in the end, the theory that Darwin published in 1859 was the theory he had finished in 1844. The additional deliberation had added volume to his evidence and nothing to his argument. What ended the delay was not a better dataset or a deeper conviction. It was a letter from a stranger that did what Darwin could not bring himself to do: it made the decision for him. He had prepared for his own death before he could face the act of deciding to publish while alive.

Darwin’s drawer is the most prestigious version of a universal experience. You recognise it. The specifics vary. It is a job you have been meaning to leave, a relationship you have been meaning to end or begin, a conversation you have been meaning to have, a city you have been meaning to move to or away from. But the architecture is identical. You have the information. You have had it for weeks, probably months. Both options are defensible. Both have been evaluated, re-evaluated, discussed with friends, weighed against your values, tested against your fears. Both are roughly equal. And so you do what Darwin did. You classify barnacles. You reorganise the kitchen. You read one more article, listen to one more podcast, have one more conversation with the same friend who said the same thing last time. You mistake the pain of choosing for a signal that more analysis is needed. It is not. The pain is not caused by the decision. It is caused by the deciding.

There is a woman sitting in a parked car in her own driveway at eleven o’clock on a Tuesday night, engine off, hands still on the steering wheel, not going inside. She has been weighing whether to leave her job for seven months. She has a spreadsheet on her laptop with two columns (Reasons to Stay, Reasons to Go) that she last updated eleven days ago, and neither column has changed in six weeks. She has had the same conversation with her sister four times. The information is not insufficient. The information was sufficient in March. Every new input, the podcast episode, the article, the friend’s opinion, is another barnacle classified in a study that was already finished. She will go inside. She will go to bed. She will not decide tonight. She has not decided any night for seven months. And every morning she wakes up a little more tired, a little more irritable, a little less able to enjoy the job she might keep or imagine the job she might want, because the act of not deciding is not neutral. It is consuming her.

But on a freezing stretch of sand on the Outer Banks of North Carolina, on the morning of December 14, 1903, two brothers solved exactly this kind of problem in three seconds.

Wilbur and Orville Wright had spent four years building the Flyer, three of them at a desolate camp at Kill Devil Hills, sleeping in a wooden shed with sand drifting through the floorboards, battling mosquitoes so thick in summer that they wore netting over their faces, and frostbite so severe in December that they could barely grip a wrench. The machine was ready. One of them would attempt the first sustained, controlled, powered flight in human history. The question of who goes first had no correct answer. They had built the Flyer together, argued over its wing camber together, crashed its predecessors together. Neither could claim precedence the other would accept. So they flipped a coin.

Wilbur won. He climbed onto the lower wing, settled his hips into the wooden cradle that controlled the wing-warping mechanism, and opened the throttle. The Flyer lurched forward along the launching rail, lifted briefly, pitched upward too steeply, stalled, and buried its front elevator in the sand. Three days of repairs in freezing wind.

On December 17, it was Orville’s turn. At 10:35 in the morning, with Wilbur running alongside to steady the right wingtip and a man named John T. Daniels, who had never operated a camera before in his life, standing behind a tripod, the Flyer lifted off the rail and flew one hundred and twenty feet in twelve seconds. The photograph Daniels took at that instant became one of the most famous images of the twentieth century. It remains the only photograph he ever made.

The coin had given the first attempt to the brother who would fail. It did not matter. The coin’s job was never to choose correctly. Its job was to end a conversation that had no correct answer and begin an attempt that did.

In 2016, Steven Levitt, the University of Chicago economist behind Freakonomics, tested this idea at a scale no researcher had attempted. He recruited over twenty thousand people who were genuinely stuck on a major life decision, not mildly uncertain but paralysed, sitting-in-the-driveway stuck, and gave each of them a coin. The decisions ranged from the enormous to the moderate. Should I quit my job? Should I end my relationship? Should I go back to school? Heads meant make the change. Tails meant maintain the status quo. Levitt followed up at two months and again at six months, surveying both the participants and independent third-party verifiers they had nominated to keep them honest.

The results, published in the Review of Economic Studies in 2021, were as simple as they were disorienting. For important decisions, the people who got heads, those who made the change, were substantially happier at the six-month mark. Not slightly. Markedly, and confirmed independently by the third-party observers. And it held regardless of what the specific decision was. Quitting a job, ending a relationship, starting a business: the content of the decision did not predict the happiness of the decider. The act of deciding did.

Jeff Bezos understood this before Levitt proved it. In 1994, at thirty, he was a vice president at D.E. Shaw, the hedge fund, earning the kind of salary that makes career changes feel reckless. He had conceived the idea for an online bookstore. He told his boss. David Shaw took him on a long walk through Central Park and told him it was a good idea, but probably a better idea for someone who did not already have a good job. Bezos went home and built himself a coin. He called it the regret minimisation framework: imagine yourself at eighty, and ask which decision you will regret more. The answer came in seconds. He would not regret trying and failing. He knew with certainty that he would regret never having tried. The framework did not analyse the decision. It ended the analysis. It was a coin made of a single question, and it landed the moment he asked it.

Why does this work? Why does the act of committing produce happiness that the act of deliberating cannot?

Because the pain of a hard decision is not what most people think it is. It is not confusion. It is grief. Every hard decision between roughly equal options is a bereavement conducted in advance: the mourning of the life you will not live, the job you will not keep, the city you will not stay in, the version of yourself that belongs to the road you do not take. The deliberation is not analysis. It is the postponement of loss. And loss postponed does not diminish with time. It compounds. Sylvia Plath saw this before the psychologists did. In The Bell Jar, she described a woman sitting in a fig tree, watching every fig, each one a different possible life, darken and shrivel and drop to the ground while she sat paralysed in the branches, unable to pick one. The figs do not wait. They rot.

But here is what Plath did not write, and what the psychologist Leon Festinger proved in 1957: the grief has an expiry date, but only if you choose. Once a decision is truly made, with the bridge burned and the other option released, the human mind performs a remarkable and largely unconscious recalibration. It increases its valuation of the chosen path and decreases its valuation of the abandoned one. The person who quits does not merely accept the new job. Over the following weeks, their brain rewrites the old one: its frustrations amplified, its rewards diminished, its daily texture remembered as greyer than it was. The new path brightens in the same proportion. Not because the facts changed, but because the commitment changed the lens through which the facts are seen. The choice creates the preference. The preference does not create the choice. This is the psychological cascade that months of deliberation were actively preventing. The machinery of healing cannot start until the loss is real. And the loss cannot become real until you decide.

Festinger called this cognitive dissonance reduction, and he published it in 1957. In the same year, on the other side of the Princeton campus, a twenty-six-year-old graduate student named Hugh Everett III was completing a doctoral thesis that asked the same question from the opposite end of reality.

Festinger had asked: what happens inside the mind when you choose? His answer: the mind rewrites reality to make the chosen path feel right. Everett asked: what happens inside the universe when a quantum event has two possible outcomes? The standard answer, the one taught in every physics classroom for thirty years, was that observation forces the event to choose one outcome and the other ceases to exist. Everett proposed something stranger. Nothing ceases to exist. The universe does not choose. It branches. Both outcomes happen. In one branch, you see the particle go left. In another branch, another version of you sees it go right. Both branches are real. Neither knows about the other.

The idea was so radical that it was largely ignored for a decade. Everett left academia entirely, never published in physics again. But his interpretation, later named the many-worlds interpretation by the physicist Bryce DeWitt, did not go away. It grew. Today it is taken seriously by a substantial and growing number of physicists, among them Sean Carroll at Johns Hopkins, David Deutsch at Oxford, and Max Tegmark at MIT. The double-slit experiment, in which a single particle fired at two openings produces an interference pattern as though it passed through both simultaneously, is the oldest and most vivid demonstration of the phenomenon. In the standard interpretation, observing which slit the particle takes collapses the superposition. In Everett’s interpretation, the observation does not collapse anything. It branches. The observer who sees the particle go left and the observer who sees it go right both exist. They simply cannot see each other.

Whether the many-worlds interpretation scales from particles to people, whether a coin flipped in a driveway genuinely divides the universe the way an electron divides at a slit, is a question physics has not yet settled. But the thought experiment it offers is, for anyone sitting in Darwin’s drawer, the most radical consolation imaginable. If Festinger is right, the act of choosing rewrites your perception of the path not taken, and the grief fades. If Everett is also right, there is nothing to grieve at all. The path not taken is being taken. The version of you that stays at the job and the version that leaves both walk away from the coin. The flip does not destroy one version and preserve the other. It sends each of you through a different door. Every fig on Plath’s tree is picked. Every branch is lived. The mercy of the coin, in this reading, is not just that it ends your agony. It is that in the most expansive understanding of what happens when a coin lands, nothing is lost.

The thought is vertiginous. It may not be true. But consider the alternative.

Bronnie Ware spent years as a palliative care nurse in Australia, sitting with people in their final weeks. She asked them what they regretted. The answers were not about bad decisions. They were not about risks that went wrong or leaps that did not land. The most common regret, the one Ware heard more than any other, the one that came with a specific kind of quiet devastation, was this: I wish I had had the courage to live a life true to myself, not the life others expected of me.

Not I wish I had chosen differently. I wish I had chosen at all.

These were people who had sat in Darwin’s drawer for decades. Who had stayed in the driveway until the engine rusted. Who had watched the figs fall one by one until there were none left to pick. In Everett’s framework, these are people who never branched, who spent a lifetime in superposition, refusing to let the coin land, and arrived at the end holding not two possible lives but none. The coin is not reckless. The coin is not flippant. The refusal to flip it, the belief that more time, more data, more deliberation will eventually reveal the correct answer to a question that has no correct answer, is the real recklessness. It is the slow, respectable, well-reasoned path to a deathbed regret.

December 17, 1903. Three days after the coin chose wrong. Orville Wright lies flat on the lower wing of the Flyer, his hips in the cradle, his hands on the elevator lever. Wilbur has just released the restraining wire. He is running alongside in the sand, his hand steadying the right wingtip, his coat flapping in the December wind. Five men from the lifesaving station stand a few yards back, squinting into the grey morning. Nobody is thinking about the coin. Nobody is regretting that Wilbur went first and stalled. And if Everett was right, somewhere in the branching architecture of everything that exists, another Wilbur is running alongside another Flyer that never stalled, and another Orville is watching from the ground, and neither of them is thinking about the coin either.

The mercy of the coin is not that it knows something you do not. It has no knowledge. It has no preferences. It cannot lie awake at three in the morning running the same comparison for the seven hundredth time. It cannot grieve the option it eliminates. Or, if a twenty-six-year-old graduate student in 1957 was right, it does not eliminate anything at all. It simply lands. And landing, on either side, it does not matter which, is the one thing the agonising mind cannot do for itself.

Sources:

Janet Browne, Charles Darwin: Voyaging, Jonathan Cape, 1995

Charles Darwin, letter to Charles Lyell, 18 June 1858, Darwin Correspondence Project

David McCullough, The Wright Brothers, Simon and Schuster, 2015

First Flight Foundation, “Coin Toss,” firstflightfoundation.org

Steven D. Levitt, “Heads or Tails: The Impact of a Coin Toss on Major Life Decisions and Subsequent Happiness,” Review of Economic Studies, Vol. 88, No. 1, 2021

Jeff Bezos, interview and public remarks on regret minimisation framework, various, 1997-2018

Sylvia Plath, The Bell Jar, Heinemann, 1963

Leon Festinger, A Theory of Cognitive Dissonance, Stanford University Press, 1957

Hugh Everett III, “Relative State Formulation of Quantum Mechanics,” Reviews of Modern Physics, Vol. 29, No. 3, July 1957

Sean Carroll, Something Deeply Hidden: Quantum Worlds and the Emergence of Spacetime, Dutton, 2019

Bronnie Ware, The Top Five Regrets of the Dying, Hay House, 2012

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

On 15 December 1971, the British Home Secretary Reginald Maudling held a press conference in Belfast. The Troubles were at their bloodiest. Soldiers patrolled the streets. Bombs were a weekly occurrence. Maudling told the assembled journalists that the British Army could not eliminate the IRA. It could, however, reduce the violence to, in his words, “something which is acceptable.” Ian Paisley condemned him. Opposition parties boycotted him. Editorialists were appalled. Maudling had simply said aloud what every institution that confronts a persistent problem eventually does in silence: negotiate the problem down to a level it can tolerate, call the result a policy, and move on.

The deeper mechanism was identified, in an entirely different context, by a fisheries scientist named Daniel Pauly. In 1995, Pauly published a one-page essay in the journal Trends in Ecology and Evolution that introduced a concept he called the shifting baseline syndrome. His subject was fish. Specifically, the collapse of fish stocks across the world’s oceans, a collapse so slow and so generational that the people managing the fisheries could not see it happening.

Pauly’s observation was precise. Each generation of fisheries scientists, he wrote, accepts as a baseline the stock levels that exist at the beginning of their careers, and uses those levels to evaluate change. When the next generation begins, the stocks have declined further, but the new scientists take the lower level as their starting point. The result is a ratchet. The baseline shifts downward with each generation, and nobody experiences the full decline, because nobody’s career spans it.

The numbers in the North Atlantic tell the story. In 1968, the northern cod catch off Newfoundland peaked at eight hundred and ten thousand tonnes. That catch was roughly three times the maximum yearly haul before the era of factory trawlers. Between 1647 and 1750, fishermen took approximately eight million tonnes of cod from those waters across a hundred and three years. The factory trawlers matched that figure in fifteen. By the time Canada imposed a moratorium on cod fishing in 1992, the stock had declined by ninety-nine percent from its historical levels. Almost nobody experienced this as a ninety-nine percent decline. Each generation of fishermen and scientists had a different baseline. A fisherman who started working in the 1980s thought the depleted stocks of that era were normal, because they were the only stocks he had ever seen. His grandfather, who remembered the teeming waters of the 1950s, was dead or retired, and his anecdotes were treated as the exaggerations of old men rather than as data.

Pauly called this a syndrome, and the word was chosen carefully. A syndrome is not a single symptom. It is a pattern of symptoms that reinforce one another. The shifting baseline causes people to recalibrate their ambitions downward along with their measurements. If today’s depleted stock is your baseline, then a modest recovery looks like success. You are measuring progress against a standard that has already been degraded, and the degradation is invisible because it happened before you arrived.

The same mechanism operates inside organisations. The sociologist Diane Vaughan studied the 1986 Challenger disaster and found that engineers had been documenting O-ring erosion for years before the shuttle broke apart. Launch after launch, the seals showed damage. Launch after launch, nothing catastrophic happened. Each flight that survived became evidence that the damage was tolerable, and the tolerable became the expected, and the expected became the standard. Vaughan called this the normalisation of deviance: a slow institutional drift in which the acceptable level of risk crept upward, one safe landing at a time, until cold weather in January exposed what the standard had become.

This is the violation floor. The level of harm, risk, dysfunction, or degradation that a system has negotiated down to and then stopped seeing. The problem’s new address, the place it lives after everyone stops trying to evict it. The violation floor is set by a thousand small accommodations, each of which is rational in isolation. Each time the cod stock drops a little and the fishery stays open, the floor sinks. Each time the O-ring erodes and the shuttle flies safely, the floor rises. Each time a city records another year of a stable murder rate and calls it progress, the floor hardens.

The psychology has a name. Adaptation-level theory, formalised by Harry Helson in the 1940s and extended by Brickman and Campbell, demonstrates that humans recalibrate their internal reference points to match persistent stimuli. The well-known finding is that lottery winners return to baseline happiness within months. The less discussed finding runs in the other direction: people adapt to persistent negatives with the same efficiency. Chronic pain patients report lower pain scores over time, not because the pain diminishes, but because their reference point shifts. Citizens of cities with entrenched pollution describe their air quality as normal. Perception filters out constants so the organism can attend to changes. If the background noise holds steady, you stop hearing it. If the floor holds steady, you stop seeing it. The floor becomes the ground.

The mechanism operates with particular efficiency inside the home. A relationship that escalates does so in increments small enough that no single incident registers as the one that changed everything. The first time a partner raises his voice during an argument, the other partner registers it as unusual. The second time, she notes it. The fifth time, she has stopped noting it, because raised voices have become part of how arguments happen in this house. The calibration has already moved. What follows moves it further. A slammed door. A thrown object that misses. A thrown object that does not miss. Each escalation arrives against a baseline that the previous escalation has already adjusted, and the adjustment is invisible because it happens inside the same nervous system that is trying to evaluate it. The survivor often cannot say when the floor began to sink. She can describe the early months of the relationship and she can describe last Tuesday, and the distance between them is vast, and she crossed it one accommodation at a time, each one feeling like a reasonable response to the current situation because the current situation was the only situation she could feel. By the time someone from outside the household describes what they see, the survivor’s most common response is confusion. The description does not match her experience. Her experience is that this is her life. A shove was a Monday.

Ask an emergency department nurse what a bad shift looks like. If she has been working for five years, she will describe something that would have been unrecognisable to her on her first day. Her first month in the department, a patient grabbed her arm hard enough to leave a bruise. She filed an incident report and went home shaken. Her colleagues were sympathetic but unsurprised. By her second year, she stopped filing reports for grabs. A grab was not an incident. By her third year, a shift without a verbal threat was a good shift, and a shift with only shouting was unremarkable. By her fifth year, she noticed she had stopped flinching. She could remember flinching. She could remember the version of herself that flinched. That version felt naive, like someone who had not yet learned what the job actually was. Nobody had told her to stop reporting. Nobody had drawn a line and said: this much violence is acceptable. The floor had drifted there on its own, one unreported shift at a time, and the reason she could not feel it sinking was that she had been sinking with it. A grab was a Tuesday.

What makes the violation floor dangerous is not that it exists. Every complex system involves tradeoffs, and some level of residual risk is a mathematical certainty. The danger is in the direction: the floor almost always sinks. It sinks because the forces that push it down, budget pressure, fatigue, generational turnover, the simple human desire to stop worrying about something that has not yet killed anyone, are constant and structural. The forces that push it up, scandal, disaster, a charismatic reformer who remembers what the original standard was, are sporadic and temporary. The asymmetry is built into the system. Over enough time, the floor will always find a lower level, and the people standing on it will always believe it has been there all along.

Pauly understood this asymmetry. His prescription was institutional: keep the data, publish the historical baselines, make the original standard visible so that no generation can mistake its degraded inheritance for the natural state of things. The defence against the shifting baseline is better measurement, measurement against a fixed reference point that exists outside the system and cannot be renegotiated by the people inside it. Publication alone has not proved to be remembrance. The cod data has been public for thirty years. The stocks have not returned. Knowing where the line used to be and acting on that knowledge are separated by the full width of human inertia.

There is one institution on earth that has solved this problem, and it solved it by rejecting Pauly’s premise entirely. Ise Jingu, the most sacred shrine in Shinto, stands in Mie Prefecture on the coast of central Japan. It has been demolished and rebuilt from scratch every twenty years for thirteen hundred years. The current cycle, the sixty-third, began in 2025. The trees used in the reconstruction are planted a century before they are needed. The twenty-year cycle is deliberate. It is one human generation. The master carpenters who lead each rebuild train the apprentices who will lead the next one. The cost is approximately three hundred and ninety million dollars per cycle. The Shinto term for the principle is tokowaka: eternal youth through renewal. The institutional translation is simpler. By changing, to remain unchanged.

The logic inverts every assumption the essay has made so far. Pauly’s prescription is to keep records, to publish the original baseline so that future generations can measure their inheritance against it. Ise’s prescription holds that records do not preserve standards. People do, and only when they are forced to enact the standard rather than read about it. There is no old shrine to compare the new shrine to. Every shrine is the first shrine. The baseline cannot shift because there is no baseline. The drift cannot start because the system resets before the first accommodation can take hold. The knowledge of how to build the shrine does not live in a filing cabinet or a published paper. It lives in the hands of carpenters who will teach it to carpenters who will teach it to carpenters. The chain is unbroken because the chain is the point.

This is a structural argument about what it would take to defeat the violation floor anywhere. The standard cannot survive as data. Data can be published and ignored. The standard cannot survive as policy. Policy can be revised and diluted. The standard survives only when it is performed, repeatedly, by people who learned it from people who performed it, in a cycle short enough that no single generation can forget it before the next performance begins. Pauly was correct that the original standard must be visible. He was incomplete. Visibility is necessary. It is not sufficient. The sufficient condition is ritual: the deliberate, costly, recurring refusal to let the original recede into history where it can be doubted or discounted or simply allowed to fade.

Most institutions cannot rebuild themselves every twenty years. Most do not. For the rest, the trajectory is one directional. Each generation’s tolerable becomes the next generation’s normal and the next generation’s invisible. The drift compounds across decades and then across centuries, each accommodation absorbed into the operating assumptions of the people who inherit it. The Roman Empire did not fall in a year. It fell across four centuries of accommodations, each one rational at the time, each one negotiated by administrators who were solving the problem in front of them with the resources they had. The frontier contracted. The currency debased. The civil service thinned. Each contraction became the new baseline against which the next contraction was measured, and by the time the last emperor was deposed the distance between the Rome of Augustus and the Rome of Romulus Augustulus was so vast that no single lifetime could have perceived the full arc of the decline. The collapse was not a moment. It was the eventual visibility of a floor that had been sinking for four hundred years.

Maudling said on camera what every institution prefers to do in silence. Every policy, every ceasefire, every regulation, every managed decline draws the same line. Somewhere in the filing cabinet of every organisation is the original target, the one that was set before the accommodations began, before the floor started sinking, before the people who remembered the old standard retired and took the memory with them. In Mie Prefecture, they do not keep a filing cabinet. They keep carpenters. The rest of us keep the drawer. Most do not open it. The floor is the ground. The ground is all there is.

Sources:

Pauly, Anecdotes and the Shifting Baseline Syndrome of Fisheries, Trends in Ecology and Evolution, 1995
Vaughan, The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA, University of Chicago Press, 1996
Helson, Adaptation-Level Theory, Harper and Row, 1964
Brickman and Campbell, Hedonic Relativism and Planning the Good Society, Academic Press, 1971
Hamilton and Haedrich, The Fall and Future of Newfoundland’s Cod Fishery, World Development, 1999
Kahn, The Human Relationship with Nature, MIT Press, 1999
Emergency Nurses Association, Workplace Violence, 2024
American College of Emergency Physicians, Rising Violence in the Emergency Department, AAMC, 2024
Adams, Rebuilding Every 20 Years Keeps This New Old Japanese Shrine Eternally Young, National Geographic, 2025

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

On the morning of 6 April 2026, Sam Altman published a thirteen-page paper. It was titled “Industrial Policy for the Intelligence Age,” and it proposed, among other things, a public wealth fund that would give every American citizen a dividend stake in the AI economy, a tax on automated labour, and a four-day workweek. The document compared the present moment to the Progressive Era and the New Deal. It cited historical precedent. It used measured language. Four days later, at approximately 3:40 a.m. on a Friday, a twenty-year-old man named Daniel Moreno-Gama threw a Molotov cocktail at the metal gate of Altman’s home in San Francisco’s Russian Hill neighbourhood. He had travelled from Houston with multiple incendiary devices, a jug of kerosene, a firearm, and a manifesto predicting that artificial intelligence would cause human extinction. The manifesto contained a hit list of AI executives. It contained a letter addressed directly to Altman. It contained the sentence: “If I am going to advocate for others to kill and commit crimes, then I must lead by example.”

The paper proposed a future. The fire proposed something older.

There is a pattern in the history of transformative technology that is so consistent it might be called a law. The pattern is this: when a new technology generates enough anxiety about the future, the society that cannot control the technology turns on the person who made it. Not the technology itself. The person. The technology is abstract. It is diffuse. It is already loose in the world and cannot be unmade. But the inventor has a name, an address, and a face. The inventor can be sued, humiliated, stripped of credentials, or firebombed. The inventor is the lightning rod onto which an entire civilisation’s ambivalence about its own future is discharged.

Consider Eli Whitney. In 1794, Whitney patented the cotton gin, a device so simple that a single worker could clean fifty pounds of cotton in a day, a task that had previously required a full day’s labour from ten people. The machine made the Southern planter class fabulously wealthy almost overnight. It also made Whitney the most hated man in Georgia. Within months, unauthorised copies of the gin spread across the South. Planters built their own versions and claimed they were new inventions. When Whitney and his business partner Phineas Miller tried to enforce their patent, local juries – composed almost entirely of the plantation owners who were getting rich from the stolen design – ruled against them. Whitney filed more than sixty lawsuits. He won almost none. A loophole in the 1793 patent act meant he could not even collect damages until the law was changed in 1800, by which point the infringement was total. When legal resistance proved insufficient, someone burned down Miller and Whitney’s warehouse in Georgia. The man who had made their fortunes had become the obstacle to those fortunes, and they removed him with the tools available: litigation, piracy, and arson.

Whitney spent the rest of his productive years not on the invention that changed American agriculture but on patent battles that consumed every dollar the cotton gin had generated. His company went bankrupt in 1797. The states of South Carolina, North Carolina, Tennessee, and finally – reluctantly – Georgia eventually paid settlements totalling roughly ninety thousand dollars. After legal costs, Whitney and Miller netted practically nothing. The machine that built the antebellum South destroyed the man who designed it, not because the machine failed but because it succeeded beyond anyone’s capacity to absorb the change.

The Luddites understood this instinct with a precision that their popular reputation obscures. The standard telling is that they were anti-technology: simple-minded weavers smashing looms because they feared the future. The historical record says something different. The Luddites did not attack all looms. They did not burn all factories. They attacked the specific workshops of specific owners who had dismissed experienced artisans or cut wages to starvation levels. They sent threatening letters to named individuals. One letter, sent to a factory owner in Huddersfield, read: “Information has been given that you are the holder of those detestable shearing frames. I shall send one of my lieutenants with at least 300 men to destroy them and burn your building to ashes.” The target was not the machine in the abstract. The target was the man whose name was attached to the machine, the person onto whom the anxiety could be projected.

This distinction matters because it reveals the mechanism. On 28 April 1812, a mill owner named William Horsfall was riding home from the Huddersfield market when he was shot by a group of Luddites led by George Mellor. Horsfall had been vocal in his contempt for the croppers whose livelihoods the shearing frames were destroying. He had boasted publicly that he would ride up to his saddle in Luddite blood. Mellor and two accomplices were hanged at York Castle the following January. But notice what happened: the technology kept running. The shearing frames were not dismantled. The factories were not closed. Horsfall’s murder did not slow the Industrial Revolution by a single hour. The violence was not functional. It was psychological. Killing Horsfall did not solve the problem the shearing frames created. It solved a different problem: the unbearable feeling of powerlessness in the face of change that could not be stopped. The named individual absorbs the fury that the nameless process cannot.

Robert Oppenheimer is the purest case. He was appointed director of the Los Alamos Laboratory in 1943, tasked by the United States government with building the atomic bomb. He succeeded. The Trinity test lit up the New Mexico desert in July 1945, and within weeks the bombs fell on Hiroshima and Nagasaki. Oppenheimer later recalled a line from the Bhagavad Gita – a text he had studied in the original Sanskrit at Berkeley – and gave it the English rendering that would define him: “Now I am become Death, the destroyer of worlds.” After the war, Oppenheimer did what Altman did eighty-one years later. He tried to write the rules. He confronted Secretary of War Henry Stimson and demanded that nuclear weapons be banned. He told President Truman he felt he had blood on his hands. He opposed the development of the hydrogen bomb, arguing that it would accelerate a dangerous arms race without making anyone safer.

The government that had commissioned the bomb destroyed the man who built it. Lewis Strauss, chairman of the Atomic Energy Commission, orchestrated a security hearing in April 1954 that was designed from the start to end Oppenheimer’s influence. Twenty-three of the twenty-four charges concerned alleged Communist associations from decades earlier. The last charge was the real one: that Oppenheimer had opposed the hydrogen bomb. The hearing panel acknowledged that Oppenheimer was a loyal citizen. It revoked his security clearance anyway, by a vote of two to one. He spent his remaining years at the Institute for Advanced Study in Princeton, an exile in plain sight, cut off from the world he had helped create and then tried to govern.

The pattern is precise. The inventor builds the thing. The thing changes the world. The world becomes anxious about the change. The anxiety needs a target. The technology is too large, too diffuse, too embedded to be a target. The inventor is right there.

But Altman’s case adds a dimension that Whitney, the Luddites’ targets, and Oppenheimer illuminate only partially. Altman was not merely the builder of the technology people feared. He was the builder who then published the paper. He was the inventor who proposed the rules. And proposing the rules makes the inventor more of a target, not less. This is the part of the pattern that is hardest to see from the inside.

When Altman published “Industrial Policy for the Intelligence Age,” he did something that looked, on the surface, like responsibility. He acknowledged that the technology his company was building would disrupt labour markets, concentrate wealth, and require new social contracts. He proposed concrete mechanisms: wealth funds, robot taxes, shorter workweeks. Critics called it a “policymercial” and accused OpenAI of regulatory nihilism dressed up as civic duty. But the more dangerous consequence was subtler. By proposing solutions to the anxiety that AI was generating, Altman implicitly confirmed the premise of the anxiety. The person who says “I know how to fix this” is also the person who admits that “this” is broken. The paper did not calm the fear. It validated it. And it attached a return address.

Oppenheimer learned the same lesson. His opposition to the hydrogen bomb was not received as the counsel of the man who understood nuclear weapons better than anyone alive. It was received as an admission that the weapons were as dangerous as the most frightened voices claimed. If the father of the bomb says the next bomb is too dangerous to build, then how dangerous is the bomb we already have? The government did not need Oppenheimer to be wrong. It needed him to be silent. When he refused silence, he became more threatening than the Soviet Union.

Whitney’s case is structurally identical. His patent enforcement was not merely a commercial claim. It was a reminder that the cotton gin belonged to someone, that the revolution the planters were enjoying had an author, and that the author expected to be compensated. The planters did not want an author. They wanted an orphaned technology – powerful, profitable, and free of any person who might claim credit, demand payment, or impose conditions. Whitney’s lawsuits were not just inconvenient. They were existential reminders that the transformation had a source, and the source had rights.

This is the mechanism that every transformative technology activates. The technology itself disperses into the economy, the culture, the daily habits of millions of people. It becomes ambient. It stops feeling like someone’s invention and starts feeling like the weather. But the inventor remains singular, visible, and claimable. And the moment the inventor speaks – to propose rules, to demand credit, to express doubt – the inventor shatters the illusion that the technology is a natural force. The inventor reminds the world that a person did this. And a person can be held responsible in a way that a technology cannot.

Moreno-Gama’s manifesto predicted the extinction of humanity. His Molotov cocktail landed on a metal gate and was extinguished by a security guard. Two days later, two more people shot at Altman’s house from a car. The technology kept running. OpenAI did not pause its models. The shearing frames did not stop. The cotton gin was not unbuilt. The bomb was not disassembled. The violence was not aimed at the thing. It was aimed at the name on the thing. And it solved the same problem it has always solved: not the problem of the technology, but the problem of having no one to blame for the future.

Every civilisation that has lived through a technological transformation has produced this moment. The question is never whether it comes. The question is whether the civilisation recognises it for what it is – a confession of helplessness disguised as an act of agency – before it mistakes the destruction of the inventor for the governance of the invention.

We are in ours.

Sources:
OpenAI, “Industrial Policy for the Intelligence Age,” April 2026. Axios, “Behind the Curtain: Sam’s Superintelligence New Deal,” April 2026. SF Standard, “OpenAI Published a New Deal for AI. Days Later, Someone Firebombed Sam Altman’s House,” April 2026. Fortune, “After Attacks on Sam Altman’s Home, Experts See Parallels to the Industrial Revolution,” April 2026. CNBC, “Suspect in Attack Aimed to Kill OpenAI CEO, Warned of Humanity’s Extinction from AI,” April 2026. Eli Whitney, patent records, US National Archives. Britannica, “Eli Whitney.” Luddite Bicentenary Project, “The Assassination of William Horsfall,” April 1812. History.com, “How Lewis Strauss Orchestrated Robert Oppenheimer’s Downfall.” Britannica, “Oppenheimer Security Hearing, 1954.”

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

On the twentieth of July, 1969, Buzz Aldrin became the second human being to set foot on the surface of the Moon. He stood in the Sea of Tranquility, looked out across the grey emptiness stretching to the curved horizon, and said two words into his radio that would follow him for the rest of his life: “Magnificent desolation.” It was a perfect description of the lunar surface. Within two years, it would become a perfect description of Buzz Aldrin.

By 1971, the man who had walked on the Moon was being hospitalised for clinical depression. His twenty-one-year marriage collapsed. He drank heavily and could not stop. He wandered through speaking engagements and ceremonial dinners where strangers asked the same question — “What was it like up there?” — and he discovered he had no answer that satisfied them or himself. Aldrin later wrote that as he toured the world in the months after Apollo 11, with no idea how he was going to top his moonwalking adventure, it dawned on him that “magnificent desolation” described not just what he had seen on the Moon but what he felt inside. He had done the most extraordinary thing a human being had ever done. He was thirty-nine years old. And the question that consumed him was the one nobody had prepared him for: now what?

This is a story about the year after you get what you wanted. Not the year of striving, sacrifice, and sleepless nights. Not the triumphant moment of arrival. The year after. The one that nobody warns you about, because from the outside it looks like everything you ever dreamed of, and from the inside it feels like standing in a landscape that is magnificent and desolate in equal measure.

Harvard psychologist Tal Ben-Shahar gave this experience a clinical name: the arrival fallacy. It is the belief that reaching a particular destination will deliver lasting happiness. The word “fallacy” does the heavy lifting. We are not merely wrong about how happy a given achievement will make us. We are wrong in a specific and predictable way. We overestimate the intensity and duration of the happiness it will bring, and we are blindsided by the emptiness that follows when the thing we built our identity around is suddenly behind us rather than ahead.

The science beneath this runs deep. In 1978, psychologist Philip Brickman and his colleagues at Northwestern University published a study that should have changed how every ambitious person thinks about success. They tracked twenty-two major lottery winners — people who had received between fifty thousand and one million dollars — and compared their happiness levels to a control group and to twenty-nine people who had been paralysed in accidents. The finding was jarring. Lottery winners were no happier than the control group. They rated their present happiness at 4.0 on a five-point scale. The controls came in at 3.82. Statistically, the difference was noise. But here was the truly unsettling part: the lottery winners reported taking significantly less pleasure from ordinary activities — reading a magazine, hearing a joke, eating breakfast — than people who had won nothing at all. The peak experience of winning had not added to their happiness. It had subtracted from everything else.

Brickman called the mechanism “hedonic adaptation.” The brain recalibrates. Whatever becomes your new normal stops producing the dopamine hit that made it feel extraordinary. The promotion, the book deal, the championship, the exit — each delivers a spike of euphoria and then the treadmill resumes, except now you are running on a higher setting with the same legs. Brickman himself, in a grim footnote that his own research might have predicted, took his own life in 1982, four years after publishing the paper that explained why getting what you want does not make you happy.

The pattern shows up with brutal consistency in elite athletes. Michael Phelps, the most decorated Olympian in history with twenty-three gold medals, fell into what he later called “post-Olympic depression” after every single Games he competed in. After Athens in 2004, it was confusion and listlessness. After London in 2012, it nearly killed him. He sat alone in his bedroom for three to five days at a stretch, not eating, barely sleeping. “I didn’t want to be in the sport anymore,” he said. “I didn’t want to be alive anymore.” This was a man who had just won four more gold medals. From the outside, he was the greatest swimmer who had ever lived. From the inside, he was a person whose entire identity had been structured around a goal that was now in the past tense.

British swimmer Adam Peaty, who has won Olympic gold and broken world records, compressed the experience into a single devastating image. “A gold medal,” he said, “is the coldest thing you will ever wear. It feels so warm because you have achieved your goal, but at what cost?” Peaty went on to describe years of depression and alcohol problems that followed his greatest victories. He had been on what he called “an endless search” for the next gold medal or world record, and when the searching paused, he found he did not recognise the person who remained.

The pattern extends far beyond sport. Researchers at Columbia Business School interviewed twenty-two entrepreneurs who had sold their companies — every one of them in deals they considered successful. The finding was unanimous. Every single founder experienced what the researchers described as a deep and prolonged sadness after the sale. Not some of them. Not most of them. All of them. The emptiness arrived precisely when everyone around them was telling them to celebrate. Neuroscience offers a partial explanation: studies have shown that entrepreneurs display brain activity when viewing their company’s brand similar to what parents show when viewing images of their children. The company is not something you do. It becomes something you are. Selling it does not just change your calendar. It severs a piece of your neurological identity.

A startling number of exited founders divorce within two years of their sale. They develop health problems. They describe a disorientation so profound that several used the same phrase independently: “I didn’t know who I was anymore.” One entrepreneur who later interviewed more than two hundred people who had sold their businesses reported that similar struggles were not the exception but the norm. The period after selling his own company, he said, was one of the hardest of his life. He had more money, more time, and more freedom than he had ever possessed, and he was miserable.

What makes the arrival fallacy so dangerous is not merely that it disappoints. Disappointment is survivable. The danger is that it strikes at identity. When you spend years organising your life around a goal — making partner, finishing the novel, building the company, winning the medal — the goal does not just sit in your future as a target. It colonises your present. It tells you who you are each morning. It gives you a reason to tolerate the difficult parts of each day. It provides a narrative: I am the person who is working toward this thing. When the thing is achieved, the narrative collapses. You are not a person working toward something. You are a person who did something. And the past tense is a terrible place to build an identity.

This is why the most common advice given to people who have just achieved something enormous — “enjoy it,” “take some time off,” “you’ve earned it” — is precisely the wrong advice at precisely the wrong moment. Enjoyment requires a self that knows what it enjoys. Time off requires a self that exists independently of the work. Earning a rest implies a temporary pause before resumption, but resumption of what? The goal is gone. The structure is gone. The daily answer to “why am I doing this?” is gone. What remains is a person surrounded by congratulations, holding a trophy or a cheque or a title, experiencing what Aldrin experienced on his world tour: everyone else is certain you should be the happiest person in the room, and you cannot explain why you are not.

The psychologists Timothy Wilson and Daniel Gilbert identified the cognitive mechanism behind this. They called it the impact bias: the systematic tendency to overestimate both the intensity and the duration of our emotional reactions to future events. We are confident that the promotion will transform our daily experience. We are certain that the house will change how we feel each morning. We are sure that the achievement will settle something restless within us. We are wrong every time, and we are wrong in the same direction every time, and we do not learn from being wrong, because each new goal feels different from the last one. This time, we tell ourselves, it will be the one that sticks.

Andre Agassi won eight Grand Slam tennis titles and tens of millions of dollars. He later revealed in his autobiography that he had hated tennis from childhood and that the pain of losing always outweighed any joy of winning. Victory, for Agassi, was not happiness. It was the temporary absence of the specific misery of defeat. He drank. He withdrew from friends and family. He described a depression that shattered his confidence and a hollowness that winning could not fill. The man ranked number one in the world was, by his own account, one of the unhappiest people in any room he entered. He did not make peace with tennis — and, by extension, with himself — until he was twenty-seven and had reframed the game as something other than a destination to reach.

That reframing is the key. The people who navigate the year after best are not the ones who lower their ambitions. They are the ones who rebuild their relationship with ambition itself. Ben-Shahar, who coined the term “arrival fallacy,” does not argue against pursuing goals. He argues against the belief that goals are finish lines. The distinction matters. A finish line is a place you cross and then stop. A direction is something you travel in and keep travelling. The people who collapse after achievement are almost always the ones who treated the goal as a finish line. The ones who recover quickly are the ones who already had the next direction in mind — not because they are restless or unsatisfied, but because they understood that the goal was never the point. The movement was the point.

Phelps eventually found his way out not by winning more medals but by becoming an advocate for mental health in athletes, redirecting the same intensity that had powered his swimming into a different channel. Aldrin, after years of struggle, became a tireless campaigner for Mars exploration — not looking back at the Moon he had already walked on, but forward to a planet nobody had reached. The exit was not retirement from ambition. It was the transfer of ambition from a destination to a direction.

There is a line that gets attributed to Alexander the Great, usually by way of Plutarch: that when Alexander saw the breadth of his domain, he wept, for there were no more worlds to conquer. The quote is almost certainly apocryphal. What Plutarch actually wrote was quite different. He wrote that Alexander wept upon hearing the philosopher Anaxarchus discourse about the infinite number of worlds in the universe, and that when asked what ailed him, Alexander replied: “Is it not worthy of tears, that, when the number of worlds is infinite, we have not yet become lords of a single one?” The real Alexander was not weeping because he had run out of things to conquer. He was weeping because he understood how much remained. That is a very different kind of tears. And it is, perhaps, the healthier kind.

The arrival fallacy is not a warning against ambition. It is a warning against a particular relationship with ambition — one in which the future achievement is the load-bearing wall of your entire psychological architecture. Pull it out and the building collapses. The people who thrive after getting what they wanted are the ones who built more walls. They had relationships that existed independently of the goal. They had curiosities that had nothing to do with the medal or the deal. They had a sense of self that could survive the past tense.

If you are in the year after, or heading toward it, the question worth sitting with is not “what do I do next?” That question will answer itself. The question is: “Who am I when I am not chasing anything?” If you do not have an answer, the arrival will find you unprepared. Not because the achievement was not worth it. But because you were never going to live there. Nobody does. The arrival is not a place. It is a moment. And moments, by definition, pass.

Buzz Aldrin spent twelve minutes on the surface of the Moon. He spent the rest of his life on Earth. The magnificent desolation was real in both places. But only on Earth did he have to learn to build something in it.

Sources: Tal Ben-Shahar, Happier: Can You Learn to Be Happy?, McGraw-Hill, 2007. Philip Brickman, Dan Coates, and Ronnie Janoff-Bulman, “Lottery Winners and Accident Victims: Is Happiness Relative?”, Journal of Personality and Social Psychology, 1978. Timothy Wilson and Daniel Gilbert, “Affective Forecasting,” Advances in Experimental Social Psychology, 2003. Andre Agassi, Open: An Autobiography, Knopf, 2009. Buzz Aldrin and Ken Abraham, Magnificent Desolation: The Long Journey Home from the Moon, Harmony Books, 2009. Scientific American, “For Olympic Athletes, First Come the Games, Then Come the Post-Olympics Blues,” 2024. Academy of Management Perspectives, “Depression and Entrepreneurial Exit,” 2018.

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.

THE SCENARICA SUNDAY

At two minutes past midnight on June 1, 2009, somewhere over the Atlantic Ocean between Rio de Janeiro and Paris, the autopilot on Air France Flight 447 disconnected itself. Ice crystals had clogged the aircraft’s pitot tubes, the slender probes that measure airspeed. Without reliable speed data, the Airbus A330’s flight computers did what they were designed to do in ambiguity: they handed control back to the humans.

What happened next took three minutes and thirty seconds. In that window, three trained pilots – professionals with thousands of flight hours between them – could not diagnose what was wrong. The co-pilot at the controls pulled the nose up when he should have pushed it down. The captain, who had been resting, rushed into the cockpit and asked: “What’s happening?” The answer, delivered by his co-pilot in a voice stripped of composure, has become one of the most haunting lines in aviation history: “We completely lost control of the airplane and we don’t understand anything. We tried everything.”

They had not, in fact, tried everything. They had not tried the one thing that might have saved them: pushing the nose down to break the aerodynamic stall. It is the most fundamental recovery technique in flying, taught to student pilots in their first weeks of training. But the pilots of Flight 447 had spent their careers managing automation, not hand-flying aircraft. When the magenta line on their screens disappeared, they did not become pilots. They became passengers.

Two hundred and twenty-eight people died that night. The plane fell for three and a half minutes, from 38,000 feet into the Atlantic, and no one on the flight deck understood why.

This is a story about what we lose when we stop practising the things our tools do for us. And it is the most important story of our time, because we are now offloading more thinking to machines than at any moment in human history.

In April 1997, two years before Flight 447’s doomed aircraft even entered service, a veteran American Airlines pilot named Warren Vanderburgh stood before a room of his colleagues at the airline’s training academy in Dallas, Texas. He had been given a problem: 68 percent of the incidents and violations the airline was investigating could be traced back to automation mismanagement. Pilots were not crashing because they lacked skill. They were crashing because they had a different skill – the skill of following a line on a screen – and it was the wrong skill for the moment when the screen went blank.

Vanderburgh had a phrase for these pilots. He called them “children of the magenta line.”

In a modern cockpit, the navigation computer paints a bright magenta path across the flight display. Follow the magenta line and the plane goes where it should. The automation handles altitude, speed, heading, descent profiles, even the approach to the runway. The pilot’s job, increasingly, is to supervise. To watch. To monitor the machines that fly the machine.

Vanderburgh’s argument was not that automation was bad. It was that automation had produced a generation of professionals who could manage systems but could not fly airplanes. They could program the computer in calm conditions but could not hand-fly through turbulence. They had become, in his words, so dependent on the magenta line that when it vanished, they had nothing to fall back on.

The phrase stuck. It entered aviation culture and never left. Nearly three decades later, flight instructors still use it. And the problem it describes has only grown.

Fourteen years before Vanderburgh’s presentation, a British psychologist named Lisanne Bainbridge published a short paper in the journal Automatica. She called it “Ironies of Automation,” and if you want to understand the single most important tension in human-machine interaction, those six pages are where to start.

Bainbridge identified a paradox so clean it almost reads as satire. The more advanced the automation, the more crucial the human operator becomes – and the less practised that operator is. Here is why: designers automate precisely because human performance is variable and error-prone. But automation fails too, in rare and unpredictable ways. When it does, the human must intervene. The problem is that the human has been sitting in a monitoring role, possibly for hours, possibly for years, doing nothing that resembles the actual task. Their skills have atrophied. Their situational awareness has dimmed. They are being asked to perform at expert level in the exact moment they are least prepared to do so.

Bainbridge saw something else, too. She noted that training could not fully solve the problem because the scenarios requiring human intervention are, by definition, the rare and unusual ones. You cannot practise for every surprise. And the more reliable the automation becomes, the rarer the surprises, and the less experience the human accumulates for handling them. It is a vicious circle: better machines produce less capable humans, who are then most needed precisely when the better machines unexpectedly fail.

The paper was published in 1983. It has since accumulated thousands of citations and its influence is only accelerating. Every year, the ironies Bainbridge identified become less theoretical and more lethal.

The magenta line is not confined to cockpits.

In 2000, Eleanor Maguire and her colleagues at University College London published a study that would reshape our understanding of the brain’s relationship to navigation. They scanned the brains of London taxi drivers – men and women who had spent years memorising the city’s 25,000 streets, 20,000 landmarks, and 320 routes through a gruelling process known as “the Knowledge.” What the scans revealed was startling: the posterior hippocampus, the brain region responsible for spatial memory and navigation, was physically larger in taxi drivers than in the general population. More remarkably, the size correlated with experience. The longer someone had been driving a cab, the larger that region had grown.

The brain had literally reshaped itself around the demands of navigation. Use it, and it grows.

The implication of the reverse is obvious, and subsequent research has confirmed it. People who rely on GPS navigation show reduced activity in the hippocampus. They are not building cognitive maps of their environment. They are following a magenta line, and the part of their brain that would otherwise be constructing a rich, flexible model of the world around them is quietly disengaging. A study published in December 2024 in The BMJ found that taxi drivers die at lower rates from Alzheimer’s disease than people in other professions, likely because decades of active navigation exercised exactly the brain structures that the disease attacks.

This is not a metaphor. The brain is an organ that strengthens with use and weakens with disuse. When we outsource navigation to a device, we are not just saving time. We are choosing not to exercise a capacity. Over years, the capacity diminishes.

Now multiply this across every domain of cognition.

We are living through the largest cognitive offloading event in human history. In the span of three years, tools powered by artificial intelligence have begun handling tasks that previously required active human thought: writing, summarising, researching, analysing, coding, reasoning through complex problems. For adults who built these skills before the tools arrived, the effect may be temporary. The skills exist. They can be recalled with effort, like a pilot who hasn’t hand-flown in months but once knew how.

For children, the picture is different and more troubling. A March 2026 article in Psychology Today framed the distinction bluntly: adults lose skills to AI; children never build them. When a student uses a large language model to write their essay, they are not offloading a skill they possess. They are skipping the cognitive struggle that would have built the skill in the first place. The essay is not the point. The thinking required to produce the essay is the point. The product was always a by-product. The process was the education.

Research published in early 2025 by Michael Gerlich at SBS Swiss Business School found a significant negative correlation between frequent AI tool usage and critical thinking abilities, with cognitive offloading as the primary driver. Younger participants showed higher dependence on AI tools and lower critical thinking scores than older participants. The study also found a buffer: individuals with advanced educational backgrounds maintained more robust critical thinking skills despite regular AI usage, suggesting that once the cognitive architecture is built, it is more resilient. But the architecture has to be built first.

Bainbridge would recognise this instantly. The irony is the same irony, just applied to cognition itself rather than to operating a machine. We are automating thinking. The automation will fail – it already does, regularly, in ways that range from subtle hallucinations to confident fabrications. When it fails, the human must step in. But the human who has been outsourcing the thinking may not have the capacity to evaluate whether the machine’s output is right or wrong, because they never developed the judgement that comes from doing the thinking themselves.

We are building a generation of children of the magenta line. Not for cockpits. For minds.

None of this is an argument against technology. The autopilot has made flying extraordinarily safe. GPS has made navigation effortless. AI tools are already making knowledge workers more productive. The question is not whether to use these tools. The question is what to do with the cognitive space they free up.

Vanderburgh’s solution at American Airlines was not to disable the autopilot. It was to train pilots to recognise when to step down from automation, to hand-fly regularly, to maintain the foundational skills that automation had made easy to neglect. The best airlines adopted this philosophy. They mandated manual flying during portions of flights. They built simulator scenarios around automation failures. They treated skill maintenance not as a luxury but as a safety protocol.

The principle translates directly to the AI moment. The answer is not to refuse the tools. It is to be deliberate about which skills you continue to exercise and which you allow to atrophy. It is to understand that every task you hand to a machine is a rep you are not doing at the cognitive gym. Some of those reps do not matter. Some of them are the only thing standing between you and a moment when you need to think clearly and the tool is not there to think for you.

The Scenarica lens on this is simple. Uncertainty is not a bug. It is the operating environment. The people who thrive in uncertainty are not the ones with the best tools. They are the ones who have built the deepest capacity to think without tools when the moment demands it. Every automation is a bet that the tool will be available when you need it. That bet has a failure rate. And the failure rate is not zero.

On the night of June 1, 2009, the pitot tubes on an Airbus A330 froze for less than a minute. The ice cleared, the sensors recovered, and the airspeed data came back. By the time the instruments were reading correctly again, the pilots no longer trusted them. They had lost the thread. They could not distinguish good information from bad because they had spent too long in a world where the magenta line did the distinguishing for them.

The autopilot was ready to be re-engaged. The airplane was ready to fly itself to Paris. But the pilots, in their confusion, never gave it the chance. They held the nose up all the way to the ocean.

The tools were not the problem. The tools were fine. It was the space between the human and the tool – the gap where skill and judgement are supposed to live – that had gone hollow. And in three minutes and thirty seconds, that hollow space became a grave.

Every time you let a machine think for you, you are making the same implicit bet those pilots made: that the machine will always be there when you need it, that the magenta line will never disappear. It is, on most days, a very good bet. But Scenarica does not exist for most days. It exists for the days when the line vanishes and all you have left is what you built when no one was watching.

Build something.

Sources: Lisanne Bainbridge, “Ironies of Automation,” Automatica (1983). Eleanor Maguire et al., “Navigation-related structural change in the hippocampi of taxi drivers,” PNAS (2000). Bureau d’Enquetes et d’Analyses, Final Report on the accident on 1st June 2009 to the Airbus A330-203, Flight AF 447 (2012). Michael Gerlich, “AI Tools in Society: Impacts on Cognitive Offloading and the Future of Critical Thinking,” Societies (2025). Warren Vanderburgh, “Children of the Magenta Line,” American Airlines Training Academy (1997). The BMJ, taxi driver Alzheimer’s mortality study (2024).

The Scenarica Sunday is a weekly essay exploring how to think, decide, and see the world more clearly. Published every Sunday morning. Subscribe free at scenarica.substack.com

Scenarica publishes daily probability intelligence on geopolitics, bitcoin, the global economy, and AI. We don’t predict the future. We price it.