Perchance to Economics

Having spent a decent amount of time in-between airports recently, the observation of the rhythms and pastimes of my fellow travelers became an enjoyable hobby. Frankly, it killed time. Between Marrakech, Madrid and Paris, the seeming liminality of each airport was fascinating to observe. Truly, if one could make the argument, time and space are all relative in these transits of travel. 

The spectacles were plentiful. From the dubious airport beer at eight in the morning, followed by the yoga-esque sleeping positions of the victims of overnight layovers, to the influencer videotaping and observable confusion of certain septuagenarians among all the technology and hassle, it was magical. 

The reason for my travel was primarily academic. I had been fortunate to have been accepted for presentation at an international conference in Morocco (EDI 2026), while after its conclusion I also spent two days in Paris as recuperation. So, an aloof 20-year-old found himself across the Mediterranean carrying suits, ties, an inexcusable amount of linen and internal trepidation. The scope of the trip was initially horrific: a solo travel mainly for work with impromptu and relative logistics. For some it can be a dream come true. For me, it was a challenge. 

In this context, Baudrillard’s analyses came to mind. The passive observance of the settings around me proved to be not just a pastime, but also a means of intellectual stimulation and an avenue for contemplation under the scorching Saharan heat. Airports can be inferred as liminal spaces, where participants are left to their own devices in a crude attempt to navigate a maze of directions until the final departure… or arrival. 

As such, one can approach such a liminal space under the guise of a simulacrum of stability. Airports have a strict and ordered layout, with prescient staff ready to help and acquainted with most questions before they’ve even left our heads. They resemble an orderly stability and constitute an idol of coherence. All these make sense for the untrained or uninitiated eye. Upon closer inspection however, everything is relative. The staff have gotten used to the questions so they either repeat the answers or synthesize them on the spot if something truly novel appears. Directions and layouts make sense until concomitant delays and impertinent maintenance appear, whereupon even provisionally the conversance shifts into extraordinary and irregular solutions. However, all such extremities still occur under the illusion, the simulacrum of stability and safety; there is still the reference to order, yet the referent has all but disappeared. 

On a flight to Paris, a delay was announced, so, to kill time, I went to the extortionate airport café and had some espresso and a pastry. There I chatted with a fellow traveler, heading to Texas. Our conversation read as such;

 

“Do you know if your flight is delayed?”
“Yeah”, he told me, “…about four hours”
I couldn’t believe it. “So, what are you going to do?”
His answer was beyond laconic; “I’ll have a couple beers and call my wife”.

 

This brief exchange reinforced the perception of an illusory sense of safety I was feeling. I damn well knew four hours at a gate is boredom exemplified, however both he and I were unphased by the event. Have a beer, get the buzz, call the missus and jump in. Why bother? Everything will be sorted out eventually. 

This simulacrum of security only exists because of the scope of the enterprise taking place within and without the realm of the airport. Besides the millions of passengers, airports operationalize security because the consequences of failure are catastrophic. So, the solution is an induced sedation, an inundation of time and space to the extent where anything seems relative and at hand. The justification is “the delay”, or “the fatigue” or anything else conjured up in a haste of preparation. 

As the saying goes, all bets are off. Why? Because the point of being at an airport is to get where you want to get, either away from home or close to it. This intermittence is where the liminality comes in, because in such a transitional state it is permissible to stop caring for minutiae, given every other thing one has to worry about (passports, baggage, luggage, hair etc.). What difference does a beer at 09:00 make, or an espresso martini at 11:00 constitute? These are simulacral happenings, as most of the people who engage in such activities can hardly be assumed to be alcoholics. What is the reference here? Leisure. What is the absent referent? Frankly, that this is not a bar but a public space and the sun is still up. 

So, people like my fellow traveler have absolved themselves of the pretense of societal preconceptions as per their actions, all while persisting in an illusion of comfort and tranquility. Similar sightings can be observed anywhere across an airport. Art thou in the mood for some pizza or burgers early in the morning or some dubious cocktails with hideous prices? Go ahead, your plane is leaving in a while, not now. Enjoy! 

This can be amalgamated into a critique of consumerism and a society of spectacles (as per Guy Debord) or even as a tirade against whatever phantom has appeared and is popular this week among the computational literati. Others could argue that such, grotesque at worst and surreal at best, scenes are the natural result of a sanitized area: long gone are the days of carefree travel and no security checks prior to boarding. As such, perhaps, the compensation for the hassle of unending aisles of individuals, obscure checks, mysterious rays and bizarre machinery can very well be this over-availability of indulgence and decadence, with the prices as the only counterweight and necessary reminder of reality.

The Dissolution of Identity

There is another curious phenomenon that unfolds within the airport. Outside its perimeter, individuals carry a plethora of identities: professions, responsibilities, relationships, and routines. They are students, parents, executives, physicians, or laborers. Once inside the terminal, however, these distinctions become largely irrelevant. They all become one: passengers.

The airport reduces identity to information. A traveler becomes a passport, a boarding pass, a seat assignment, a baggage allowance, and eventually a gate number. The individual is temporarily translated into a logistical entity whose primary purpose is efficient movement through a highly regulated system. They become an automaton the likes of which defy categorization.

One can arrive in Athens, Madrid, Paris, Dubai or Singapore and require only a glance at the language displayed overhead to know where one is. Everything else, from the polished floors, luxury boutiques, chain cafés, security procedures, to the departure boards and announcements, appears almost eerily interchangeable. Airports no longer imitate cities; increasingly, they imitate one another.

This transformation is largely accepted without resistance. It is not unusual to hear someone refer to themselves by their destination rather than by occupation or circumstance: “I’m flying to Madrid,” “I’m connecting through Doha,” or “I’m waiting for Gate B17.” Biographical identity momentarily recedes behind operational identity. The reference is speaking, not the referent.

In this sense, the airport represents a peculiar inversion of everyday society. Rather than expressing individuality, it demands procedural conformity. Everyone follows identical rituals, submits to identical inspections, and moves according to identical instructions. The illusion of personal autonomy remains intact, yet the overwhelming majority of decisions have already been prescribed by the architecture of the system itself. One is powerless to stop it; after all, there is a flight waiting to be boarded, and they’ve paid for it.

The Commodification of Waiting

Perhaps nowhere is contemporary consumer culture more visible than in the management of waiting. Airports are not designed merely to transport passengers; they are equally designed to occupy them. Attendance is not a structural headache but a desirable feature, via which commerce survives and expands, if not predates the tired and potentially overwhelmed traveler as its prey.

This ordinarily understood interruption of productive activity becomes a tremendous opportunity. Restaurants, cafés, luxury boutiques, duty-free stores, lounges, massage chairs, sleeping pods, premium security lanes, and countless other services exist not because passengers actively sought them, but because the architecture of travel inevitably produces idle time, it manufactures it.

The delay itself becomes economically productive.

This is where Guy Debord’s notion of the spectacle becomes particularly illuminating. Consumption no longer follows necessity but instead fills temporal emptiness. The airport offers a carefully curated sequence of spectacles intended to transform enforced inactivity into voluntary participation. One does not merely wait for departure; one shops while waiting, drinks while waiting, browses while waiting, or upgrades the experience of waiting altogether. Inactivity somehow produces activity, albeit onesided and purely transactional, thus vapid and void of any transcendent meaning.

Even hierarchy is not spared of the intransigence of commodification. Priority boarding, business lounges, and fast-track security transform identical journeys into differentiated experiences purchased through additional consumption. The transition itself acquires market value.

Temporal Dislocation

Time itself appears to function differently within the airport.

Outside the liminal gates, time possesses an unseeming social rhythm. Breakfast belongs to the morning, work occupies the afternoon, and sleep arrives at night. Airports weaken these conventions, if not outright trounce them. A departure at six in the morning may justify dinner at four, cocktails before sunrise, or sleep beneath fluorescent lighting in the middle of the afternoon. Habits become transactional and routines are rendered as temporary exceptions in the context of the trip.

The clock continues to measure hours, yet its social meaning becomes increasingly unstable. It is no longer absolute. No, the clock still measures hours, but nobody consults it for the purposes for which clocks ordinarily exist; the referent (routine & obligations) has been made impractical and thus a new meaning has been assigned to the old machine, as per the demands of the new situation.

International travel complicates matters further. Watches display one timezone, departure boards another, while one’s biological rhythm often conforms to neither. Jet lag? Perhaps a more fitting term is deracinated circadian woes. Travelers cease organizing themselves around conventional daily routines and instead orient themselves toward a sequence of departures, arrivals, boarding announcements, and connections. They escape the tangible and embark onto a realm of relativities, where everything can shift and be altered, usually by factors outside of their control.

Time becomes operational rather than experiential. It is reduced to a reference.

In this transitional environment, ordinary temporal expectations lose much of their authority, and liminality prevails. The question ceases to be “What time is it?” and instead becomes “When does boarding begin?” Chronological time gradually yields to logistical time. Can this however really be called “time” in its conventional and accepted sense?

The Suspension of Ordinary Norms

This temporal uncertainty coincides with a broader suspension of everyday social conventions. Behaviors that might appear unusual elsewhere become almost entirely unremarkable inside the terminal. Boundaries, much like time and space, are blurred; they are rendered liminal.

Alcohol is consumed before sunrise. Individuals sleep across rows of chairs or directly upon the floor. Expensive meals are purchased at unconventional hours. Conversations begin between complete strangers and end without expectation of future contact. Dress codes fluctuate dramatically between business suits, athletic clothing, pajamas, and holiday attire.

Such behaviors rarely attract attention because everyone present implicitly recognizes the exceptional nature of the environment. The destination, if not making it to the destination, matters more than the trip to get there.

Anthropologically, this reflects one of the defining characteristics of liminality. Transitional spaces often relax the norms governing ordinary life, permitting behaviors that would otherwise appear incongruous, if not outright absurd. The airport becomes a temporary suspension of routine social expectations, not because those expectations disappear entirely, but because they are subordinated to the singular objective shared by every traveler: reaching somewhere else. It is perhaps one of the last places in modern society where strangers routinely initiate conversation without suspicion.

The absurdity of paying excessive prices for mediocre coffee or eating pizza before breakfast is accepted without protest because such inconveniences are understood as incidental costs of transition rather than violations of normal life. The coffee per se is not the actual product of the transaction. Instead, it can be argued that ultimately, the coffee is the simulacrum of normality, the pretense of stability and reason in an otherwise unstable, fluctuating and hyperreal setting. Same for the questionable food and alcoholic beverages. Nobody is really purchasing them with themselves in mind; they serve as vehicles for comfort and reassurance, a much needed break from the persistent anxiety of travel and the uncertainty of a world where circumstance can change faster than in an instant.

Security as Performance

The airport also presents one of the most visible performances of institutional authority in contemporary society, if not a procedural zealotry that has all but eclipsed from our post-modern world.

Security procedures are highly ritualized. Shoes are removed. Liquids are separated into transparent bags. Laptops emerge from backpacks only to disappear again moments later. Passengers proceed through scanners, submit to (random) inspections, and await approval before continuing their journey.

These rituals undoubtedly possess practical functions. Yet they also carry a symbolic dimension.

Security must be seen as much as it must be exercised. There needs to be control, even if at times ridiculous and pointless.

The visible repetition of these procedures reassures travelers that order is being maintained, regardless of whether each individual action contributes equally to overall safety. The performance itself becomes part of the experience of security.

This observation aligns closely with Baudrillard’s discussion of simulation. The image of control becomes inseparable from control itself. Confidence is generated not merely through actual protection but through the continual visibility of protective practices. It is a hyperreal spectacle in the most performative of ways, where most involved understand the occasional vapidity of the process yet follow it anyway so as to prevent the potential disaster.

Airports as the Archetype of Postmodernity

It may therefore be mistaken to regard airports as exceptional environments. Their apparent peculiarity may simply render visible tendencies already present throughout contemporary society. The airport is not a place; it is a liminal institution between humanity as a referent and the perspectives humanity has for itself as references.

Increasingly, daily life resembles perpetual transit. Employment is impermanent. Residence is flexible. Relationships begin and end across digital platforms. Identity is authenticated through passwords, biometric verification, and algorithmic databases. Navigation relies upon screens. Consumption fills moments of inactivity. Standardized global spaces replace distinct local identities. Everything is a simulacrum of what existed in the past, yet void of the essence the original referent possesed.

Perhaps airports do not represent an interruption of ordinary society, but its logical culmination, the apogee of incessant pursuits that led us to specters of what was once idyllic and unattainable.

If Baudrillard argued that simulation gradually replaces reality, then the airport may be understood as one of the clearest manifestations of that transition: a carefully managed environment in which movement supersedes place, procedure supersedes identity, and representation increasingly substitutes for lived experience. In other words, as a transcedent institution, the airport can be approached as the ultimate liminality – an incessant confusion.

One enters intending only to depart.

Instead, one briefly inhabits a world in which geography, time, identity, and ordinary social conventions become strangely negotiable, alterable, if not outright manufacturable. Only after leaving does their apparent solidity quietly return. Or perhaps, having experienced the airport, one begins to suspect that they were never quite as solid as they first appeared. They only believed they were solid until liminality caught up with them.

There is a price for a unit of thinking, and it is falling faster than almost any price in the modern economic record. In November 2022, querying a machine that could perform at roughly the level of a competent generalist cost about twenty dollars for a million tokens. By October 2024 the same capability cost seven cents. A two-hundred-and-eighty-fold collapse in eighteen months, and the curve has not flattened since.

Prices do not usually behave this way. The cost of a loaf of bread, a kilowatt-hour, an hour of plumbing, they drift, they do not fall through the floor. When a price detaches from gravity like this, it is worth asking what, exactly, has become cheap. The answer is uncomfortable, because the thing that has become cheap is a passable imitation of the one commodity the developed world had organized its entire labor market around selling: cognitive effort. Reading, drafting, summarizing, answering, reconciling, coding the routine parts. The work of the desk.

And once a thing becomes cheap, capital goes looking for the places it is still expensive.

That is the whole story, mechanically. Somewhere between the salary line on a payroll and the metered line on a cloud-services invoice, a gap has opened. It is a wide, structural, exploitable gap between the price of human cognition and the price of its synthetic substitute. Markets do not leave gaps like that alone. They build machinery to stand inside them and collect the difference. Call it the labor arbitrage machine: not a single tool or company, but the emergent apparatus; models, agents, vendors, procurement teams, investor expectations. All that converts the price gap between a human hour and a thousand tokens into margin.

Engineers have a phrase for what happens to a well-designed system when the load exceeds what it can serve: graceful degradation. The system does not crash. It sheds the least essential functions, runs slower, drops quality at the edges, and stays standing. The labor market is now being asked to degrade gracefully, to shed the most automatable cognitive tasks, quietly, function by function, without anyone declaring a crisis. Totale, because the shedding is not confined to one trade. It reaches the call center and the back office and the junior desk and the first draft, all at once. This essay is about that machine: how it prices its inputs, where it is already running, and the three postures a firm can take toward it – refusal, the loop, and reinvention – each with a real cost and a real casualty.

The Price of a Thought
Begin with the input that makes the machine possible, because without the price collapse none of the rest is even arithmetic. Stanford’s AI Index, drawing on Epoch AI, tracks the cost of running a model at a fixed level of capability over time. The headline series is the one above: from twenty dollars per million tokens to seven cents in a year and a half. Depending on the task, inference prices have been falling somewhere between nine and nine hundred times per year, with the fastest declines arriving after January 2024. The floor keeps dropping as competition thickens. This is a price war between a handful of frontier labs and a long tail of open-weight challengers willing to undercut on raw cost.

Figure 1. The price of a thought, collapsing: the cost to query a model at GPT-3.5 capability fell roughly 280-fold in eighteen months. Source: Stanford AI Index 2025, citing Epoch AI.

What that line means in practice is that intelligence, or at least the part of it that can be reduced to pattern, retrieval, and fluent paraphrase, has been transformed from a scarce, expensive, human-embodied resource into something closer to a utility. Metered. Abundant. Billed by the unit. And like every utility before it, the moment it became cheap and reliable, the question stopped being whether to use it and became where to point it.

Here the two ledgers diverge in a way that organizes everything that follows. A human worker is a fixed cost with a floor: a salary, benefits, a desk, a manager’s attention, sick leave, the slow expensive business of being a person. A model is a variable cost with no floor; you pay per token, the unit price falls every quarter, and when the work stops the meter stops with it. One of these lines bends downward on its own. The other does not. A finance function that has internalized that asymmetry does not need to be told what to do; the spreadsheet tells it.

The Arithmetic of Substitution
Abstractions persuade no one. The machine becomes legible only when you put a single unit of work on the table and price it both ways. Take the most automatable cognitive job in the economy by sheer headcount: the customer service representative. In May 2024 the United States employed roughly 2.8 million of them, at a median wage of about $42,830 a year. The Bureau of Labor Statistics already projects the occupation to shrink five percent over the decade to 2034, a loss of some 154,000 posts, which is itself a quiet forecast about who wins this comparison.

Run the numbers per resolved conversation. A human agent handling perhaps fifty chats a day across roughly two hundred and fifty working days resolves on the order of 12,500 conversations a year. Against base wages alone, that is about $3.43 per conversation, and the true figure is higher once benefits, supervision, premises, and attrition are loaded in. Now price the same conversation as tokens. Assume a realistic agentic exchange: a system prompt, retrieved account and policy documents, a few turns of back-and-forth, call it fifteen thousand tokens in and two thousand out. At early-2026 list prices, that single resolved conversation costs roughly twelve cents on a frontier model, six cents on a capable mid-tier model, and about a fifth of a cent on a budget model.

Figure 2. The arithmetic of substitution. Illustrative cost to handle one resolved customer conversation: human base wage versus per-token model cost at three price tiers. Assumes ~15,000 input + 2,000 output tokens per resolved conversation; excludes build, integration, and human-oversight costs. Sources: BLS (wages); published API list prices, early 2026.

Three cents. Six cents. Twelve cents. Against three dollars and forty-three.

Hold the obvious objections, because they matter and we will return to every one of them: the human handles the furious customer, the ambiguous edge case, the conversation that needs a person to simply be a person; the model needs building, integrating, supervising, and correcting; the cheapest model is not the most reliable; and the comparison flatters the machine by costing only the easy tier of work. All true. But none of it closes a gap of this magnitude. When the per-unit cost of a substitute is one-thirtieth to one-fifteen-hundredth of the incumbent, the substitute does not have to be as good. It only has to be good enough on enough of the volume, while the arbitrage takes care of the rest.

There is a subtler effect, easy to miss if you think only in terms of substitution. When the marginal cost of a cognitive task falls toward zero, it does not merely make the existing task cheaper; it makes entirely new things economic that were not before. Support a firm could never have afforded to staff becomes a free feature. A product that once justified a single tier of service can suddenly justify infinite, instant, multilingual service. Prompt caching and batch processing push the per-unit cost down by a further order of magnitude on repeated work. The arbitrage, then, is not a one-time saving banked against a fixed list of tasks. It is a moving floor that keeps descending.

If that sounds theoretical, it has already been run at scale, in public, by a company willing to put a number on it. In February 2024 the fintech Klarna launched an OpenAI-powered customer-service assistant. Within its first month it had handled 2.3 million conversations, about two-thirds of the company’s total chat volume, doing what Klarna described as the work of 700 full-time agents. Resolution time fell from eleven minutes to under two; repeat inquiries dropped a quarter; the assistant operated in more than thirty-five languages, around the clock. Klarna put the profit improvement at roughly forty million dollars in the first year.

Notice what that figure actually is. Seven hundred agents at a loaded cost of fifty to fifty-five thousand dollars each is, near enough, forty million dollars. Klarna’s headline savings number is the labor arbitrage, stated in dollars, by the firm capturing it. The market understood the signal immediately: on the day of the announcement, shares of Teleperformance, one of the world’s largest call-centre operators, fell sharply as investors repriced the future of an entire industry built on human conversational labor. The machine had been demonstrated, and the demonstration was the point.

Adoption Without Transformation
One demonstration does not make a regime. So how widely is the machine actually running? McKinsey’s 2025 survey of nearly two thousand organizations found that 88 percent now use AI in at least one business function. That is an increase of 78 percent a year earlier and 55 percent the year before that. Adoption, in other words, is no longer a differentiator; it is the baseline. Stanford’s index tells the same story from a different sample: organizational use jumping to 78 percent in 2024 from 55 percent, and the use of generative AI in at least one function more than doubling to 71 percent.

And yet. Beneath the adoption curve sits an uncomfortable second number. Only around 39 percent of McKinsey’s respondents reported any measurable impact on enterprise earnings, and only about six percent qualified as genuine high performers seeing material EBIT effects. The majority sit in what the report’s readers have taken to calling pilot purgatory: experiments that never graduate to production. Nearly everyone has the machine plugged in. Almost no one has rewired the building around it.

Figure 3. Adoption without transformation. AI use is nearly universal and still climbing, but enterprise-level impact remains concentrated in a small minority of firms. Source: McKinsey, The State of AI (2023–2025).

This gap is not a footnote; it is the texture of the moment. The discourse speaks of 2025 as the year of the AI agent; systems that do not merely answer a prompt but plan, decide, and execute multi-step workflows. In the consultancy’s phrase they are known as “digital employees”. Twenty-three percent of firms report scaling such agents somewhere in the organization. But in any given function, fewer than one in ten has agents in genuine production. The machine’s reach, for now, exceeds its grip. Which means the interesting question is not whether firms will pursue the arbitrage but how. And here the field narrows to three postures, three answers to the same spreadsheet.

The workforce expectations buried in the same survey are where adoption and arbitrage finally touch. A median of thirty percent of respondents now expect headcount to fall in the functions where AI is deployed and roughly a third anticipate an enterprise-wide reduction of three percent or more. These are not yet cuts; they are intentions, the leading edge of the spreadsheet becoming policy. The reason most firms have not yet realized the savings is mundane and revealing: they are good at running AI projects and bad at rebuilding the operating model around them. Capturing the arbitrage at scale means redesigning the workflow end to end, not bolting a model onto the old one. The gap in Figure 3 is the distance between buying the machine and rewiring the building.
Three Postures Toward the Machine

A firm staring at the gap between $3.43 and three cents has, in the end, only three things it can do. It can refuse the trade and keep its people. It can split the difference and put a human in the loop. Or it can take the trade in full, shed the labor, and redeploy the freed capital. Each is defensible. Each is being chosen, right now, by serious companies. And each extracts a price that is paid by someone.

I. The Refusal
The first posture is to decline. To treat human judgement, presence, and accountability as the product rather than a cost to be optimized away, and to say so. This is not nostalgia; in several markets it is a strategy. There are domains where the human is the value proposition: private wealth, complex care, high-trust B2B relationships, the bespoke and the consequential. A bank that automates the furious client at the worst moment of their financial life has not saved money; it has taught that client to leave. The refusal banks on the premise that as synthetic interaction becomes infinitely abundant, a real human becomes a luxury good.

The history of automation offers more examples of this dynamic than either its advocates or critics usually admit. Mechanical looms did not eliminate handmade textiles; they transformed them into luxury goods. Industrial agriculture did not eliminate artisanal food production; it narrowed it into premium markets whose value derived precisely from their departure from industrial scale. Recorded music did not eliminate live performance. In each case abundance increased the relative value of scarcity. The mass-produced product won the volume market, but the handcrafted product often captured the margin.

The same possibility exists for cognition.

As synthetic intelligence becomes abundant, instantaneous, and nearly free, certain forms of human judgement may acquire value precisely because they remain expensive. A legal opinion signed by a named partner carries a form of accountability that no model can assume. A physician delivering a difficult diagnosis provides not merely information but responsibility for that information. A private banker managing generational wealth sells trust as much as analysis. In each case the customer is purchasing not only the answer but the person attached to it.

There is evidence that this distinction already matters. Studies of automation in medicine repeatedly find that patients express lower trust in fully automated decision-making, particularly in consequential domains involving diagnosis, treatment, or uncertainty. Trust rises when AI serves as an advisor rather than a replacement, preserving visible human accountability. Likewise, surveys of financial and legal services continue to show that clients place disproportionate value on access to identifiable experts even when routine analysis becomes commoditized. The technical quality of the answer matters; the ability to assign responsibility for it matters more.

This creates an unusual economic possibility. The firm that refuses full automation is not necessarily rejecting efficiency. It may instead be repositioning itself into a different market entirely. When machine-generated interaction becomes ubiquitous, human interaction becomes differentiated. The customer who cannot tell whether they are speaking to a machine is unlikely to pay a premium. The customer who knows they are speaking to a person sometimes will.

The refusal therefore treats humanity not as a cost centre but as a scarcity asset. Its wager is that while cognition can be automated, accountability cannot; while information can be reproduced infinitely, trust remains stubbornly attached to individuals. The firm choosing refusal is not betting against technology. It is betting that abundance creates its own demand for authenticity.

The supporting evidence is not merely sentimental. Recall that the arbitrage flatters the machine by costing only the easy tier of work; the moment volume tilts toward the ambiguous, the empathetic, or the legally consequential, the model’s failure rate becomes the firm’s liability. Tacit knowledge – the practical understanding learned on the job, never written down, and therefore absent from the text the models were trained on – remains stubbornly human. There is a reason the senior practitioner is so hard to automate: the thing she knows was never typed anywhere for a model to read.

The asymmetry of failure is the refusal’s strongest card. A human agent who mishandles a conversation does so once, quietly, to one customer. A model deployed across the whole volume can fail in the same way to thousands at once. In the age of the screenshot, a single grotesque failure travels further than a year of competent service. The firm that automates its most sensitive interactions is not trading a small per-conversation saving against a small per-conversation risk; it is concentrating a diffuse, survivable human error rate into a single systemic one. For a brand whose entire value rests on trust, that can prove catastrophic at the precise moment it looks most efficient.

But refusal has a cost, and it is brutal in its simplicity. A competitor who takes the trade can underprice you on the entire automatable tier of the market and use the margin to compete for the rest. Hold the line on human-only service and you may find yourself defending a premium that the market has quietly decided not to pay. As I have written before in another context, a neo-Luddite in 2026 only has his future to miss while the world accelerates around him. The refusal is honorable; it is also, for any firm exposed to price-sensitive volume, a wager that the gap will not widen. The gap is widening.

II. The Loop
The second posture refuses the binary. It keeps the human and the machine in the same workflow – the human in the loop – and tries to capture most of the efficiency without surrendering judgement, accountability, or the apprenticeship that produces the next generation of competence. This is the posture the evidence treats most kindly, and it is the one Klarna itself eventually adopted after overreaching.

The foundational study here is Brynjolfsson, Li and Raymond’s Generative AI at Work, which tracked the staggered rollout of a conversational AI assistant across 5,179 customer-support agents handling some three million chats. Access to the tool raised productivity, defined as “issues resolved per hour”, by 14 percent on average. But the average concealed the finding that matters: a 34 percent improvement for novice and low-skilled workers, and close to zero for the experienced and highly skilled. The model worked by capturing the tacit know-how of the best agents and disseminating it to the rest, helping newer workers move down the experience curve faster. It also improved customer sentiment and raised employee retention.

One detail from that study deserves to be set apart, because it is the entire argument for the loop in a single number. The agents adopted only about 38 percent of the model’s suggestions. They were not stenographers transcribing an oracle. They exercised judgement over which advice was good. The system’s value came precisely from the combination of the model’s breadth and the human’s discrimination. The loop is not a compromise between two inferior options. On the evidence, it can be better than either alone.

The machine raises the floor. The human still has to choose.

Klarna’s second chapter is the cautionary half of the same lesson. Having declared the work of 700 agents automated, the company cut its workforce sharply from around five thousand to roughly thirty-five hundred, largely through attrition. Then, in May 2025, its chief executive told Bloomberg the company had cut too deep on humans, and began reopening hiring for premium human support. The most-cited automation success story in the industry had quietly rebalanced toward the loop. The lesson it drew in public was the one the data already implied: automate the high-volume, low-empathy tier, and move the humans up the value chain rather than out of the building.

But the loop carries its own concealed cost, and it is the one I find most worth dwelling on, because it does not show up on any spreadsheet for years. If the model most helps the novice, lets the beginner perform like a veteran on day one, then the firm’s incentive to actually train that beginner quietly evaporates. Why invest in the slow, expensive manufacture of competence when a subscription delivers a passable version of it instantly? This is the skills-ceiling problem, and it connects directly to something I have argued at length elsewhere: the apprenticeship layer of cognitive work – the beginner tasks through which novices have always become professionals – is precisely the layer the machine performs most fluently. Raise the floor for the novice today and you may find you have removed the staircase by which novices once became experts. The loop preserves the human in the room. It does not, on its own, preserve the path that put her there.

There is a second cost to the loop, quieter than the first. A human-in-the-loop system is, by construction, a measured human. Every suggestion accepted or refused, every resolution time, every sentiment score becomes data. The same apparatus that augments the worker also surveils her at a granularity no clipboard ever achieved. I have written before about how digital systems reduce human interaction to quantifiable engagement; the loop imports that logic into the workplace itself, turning judgement into a metric and the worker into a node whose deviations from the model can be flagged. The augmented worker is more productive. They are also more legible, more comparable, and should the arbitrage ever tilt far enough, more easily dispensed with. Augmentation and replacement are not opposites; often they are the same project at different stages.

III. The Reinvention
The third posture takes the trade in full and is unapologetic about it. It treats AI not as an assistant to existing workers but as a reason to have fewer of them. It reframes the resulting layoffs not as loss but as reinvention: capital freed from routine cognitive labor and redeployed toward higher-value work. This is the posture that generates the headlines, and the fear.

The displacement is real and it is being named. The outplacement firm Challenger, Gray & Christmas, which has tracked AI as a stated reason for job cuts since 2023, recorded 54,836 layoffs explicitly attributed to AI in 2025 alone. That is more than triple the combined total of the two prior years, and part of a cumulative figure of nearly seventy-two thousand since tracking began. Against a year in which US employers announced roughly 1.17 million cuts in total, the highest since the pandemic, and in which the technology sector alone shed over 154,000 posts, the AI-attributed number looks almost modest. Almost.

Figure 4. The named cause is rare and rising. US layoffs explicitly attributed to AI more than tripled from the 2023–2024 combined total to 2025, and are widely thought to undercount the real effect. Source: Challenger, Gray & Christmas.

But the named figure is the wrong number to watch, and the firm that compiles it said so plainly. “Regardless of whether individual jobs are being replaced by AI,” its workplace expert observed in early 2026, “the money for those roles is.” That sentence is the reinvention posture in miniature. Firms rarely announce that a chatbot fired a department. They announce restructuring, efficiency, a pivot toward AI investment. The budget that used to fund a row of desks is rerouted to compute and to a smaller number of higher-paid roles. The arbitrage rarely appears in the layoff notice. It appears in the capital-expenditure line.

IBM is the instructive case, because it shows the reinvention working exactly as advertised and reveals what that means. The company used AI agents to automate the bulk of its routine human-resources tasks; its internal system, by its own account, now handles the overwhelming majority of such requests, and a few hundred HR roles were displaced. Yet IBM’s total headcount rose. “Our total employment has actually gone up,” its chief executive told the Wall Street Journal, “because what it does is it gives you more investment to put into other areas” such as programmers, salespeople, the “critical-thinking” roles where humans face other humans rather than process rote work. This is the optimistic ledger: automation as a capital-reallocation engine, shrinking the routine and funding the strategic.

The optimistic case has history on its side, and it deserves to be stated at full strength. Every previous wave of automation provoked confident predictions of permanent unemployment, and every previous wave was falsified by adaptation: new industries, new roles, work no one had thought to want. The reinventionist bets that this time is no different; that the surplus released by cheap cognition funds the next thing, as it always has. The wager is reasonable. But it rests on a structural assumption this particular wave quietly violates. For two centuries, automation climbed the ladder from the bottom rung of physical drudgery upward, leaving cognitive work for last. Generative AI has inverted the climb: it is strongest at precisely the entry-level cognitive tasks through which beginners once became professionals. Adaptation needs a pathway, and a pathway needs beginners permitted to exist long enough to turn into something else.

And it is genuinely optimistic for IBM, and for the programmers it hired. The harder questions are distributional, and they are the ones the reinvention posture is least eager to answer. The HR clerk whose work was automated is not, in general, the person hired into the better-paid engineering role. The surplus the arbitrage releases is real; the claim that it accrues to the displaced is mostly aspirational. Which returns us to a point I have made before about a different transfer of value: in a system under structural pressure, value does not disappear – it is captured. The arbitrage generates a surplus. The only question that matters is who captures it.

A price advantage, however dramatic, is not a law of nature. Economic history contains numerous cases in which a technically superior or cheaper alternative spread far more slowly than its arithmetic suggested. Nuclear power promised electricity too cheap to meter, yet regulation and public opposition constrained its deployment for decades. Electronic health records took years to penetrate healthcare despite obvious efficiency gains because institutions, incentives, and workflows resisted change. Even the shipping container, perhaps the most economically transformative logistics innovation of the twentieth century, spent years waiting for ports, unions, railroads, and regulators to reorganize themselves around it.

The lesson is not that economics loses. It is that economics often travels at the speed of institutions. A sufficiently large price gap exerts relentless pressure toward adoption, but pressure and transformation are not the same thing. The machine may be economically inevitable without being organizationally immediate.

The Terms of the Trade
If the arbitrage is going to run regardless, which the price gap guarantees it will, then the useful work is not to denounce it or to cheer it, but to argue over its terms. The machine is indifferent to how its surplus is distributed. People are not, and neither, in the end, are the economies they compose. A few principles follow from everything above, addressed to the parties who can actually set those terms.

For firms: the loop beats the binary on the evidence we have. The reflex to chase the cheapest model into the largest layoff ignores both the failure-rate liability and the slower erosion of the talent pipeline. The serious move is to automate the genuinely routine, route the consequential to humans, and keep manufacturing competence on purpose, because the firm that stops training beginners is borrowing its senior talent from a future that may not arrive. The World Economic Forum’s employers already report that 39 percent of workers’ core skills will be outdated between 2025 and 2030, and name the skills gap as their single greatest barrier to transformation. Reskilling is not corporate charity. It is the maintenance schedule for the only input the machine cannot yet supply.

For workers: the defensible ground is the work the model performs worst: judgement, accountability, the management of other humans, and the tacit craft that was never written down. The IMF estimates that around 40 percent of jobs globally are exposed to AI, rising to roughly 60 percent in advanced economies, where about half of the exposed roles may be harmed rather than helped. Exposure is not destiny, but it is a map. The instinct to compete with the machine on speed and volume is a losing one; the instinct to climb toward what it cannot do is not.

For policymakers: the apprenticeship problem is a public problem, because no individual firm has an incentive to solve it. For most of economic history the transition from novice to professional was engineered, not left to chance. Guilds, indentures, the post-war ladder of cheap training and internal promotion; all those provided a pathway to professional augmentation and amelioration. If the machine is dismantling the entry-level rung faster than the market is rebuilding it, then some modern equivalent of the indenture, like subsidized training posts, public apprenticeships, a deliberate manufacture of the next cohort, is not nostalgia but infrastructure. Transparency helps too: if firms were required to name AI as a cause of displacement rather than laundering it through “restructuring,” we would at least be arguing over a real number instead of an undercount.

And the distribution question is not merely one of fairness; it is one of stability, which even the most hard-nosed reinventionist has reason to weigh. A cohort that does everything it was told; accumulates the credentials, masters the tools and still finds the entry-level rung sawn off, does not stay quiescent forever. Thwarted expectation is among the most reliable raw materials of unrest. An economy that captures the arbitrage narrowly, routing its surplus to capital and a thin tier of senior labor while hollowing out the path beneath them, is not only being unjust. It is manufacturing a frustrated electorate, and billing the invoice to a later decade.

The arbitrage is a fact. Its terms are a choice.

Degradation, Gracefully
Return to the word. Engineers prize graceful degradation because the alternative – the sudden, total, catastrophic failure – is so much worse. A system that sheds load quietly and keeps standing is, by the standards of engineering, a success. That is precisely what makes the economic version so difficult to resist, and so easy to wave through. There is no crash to point at. No date on which the labor market fell over. Only a long, soft shedding: the call centre thinned, the back office automated, the junior desk never filled, the first draft now machine-made. Each decision is locally rational, priced by the same widening gap, however none of them announcing themselves as a turning point.

The machine does not hate the workers it displaces, any more than a spreadsheet hates the cell it overwrites. It simply notices a price difference and collects it. That is its nature, and the price difference is not going to close; the line in Figure 1 does not bend back up. What remains undecided is not whether the arbitrage runs but who writes its terms. That is, whether the surplus it releases is captured narrowly or shared, whether the staircase from novice to expert is dismantled or rebuilt, whether the human is moved up the value chain or simply out of the building.

Degradation totale, then, is not a prophecy of collapse. It is a description of a posture: an economy quietly optimizing away the most automatable fraction of human cognitive work, function by function, gracefully, without ever quite admitting that is what it is doing. The system stays standing. The feeds refresh. The margins improve. And somewhere in the slow shedding, a category of human effort is being repriced toward zero; not with a crash, but with a discount.

Keep cool. Watch the price of a thought. The machine is already running.

On paper, adulthood should be arriving later than ever, and arriving in better condition than ever.

Young adults today are, on average, the most schooled cohort in human history. They hold more credentials than their parents did, command more information than any generation before them, and operate tools that would have read as miracles a few decades ago. They spend longer in formal education and emerge into a world whose productive capacity dwarfs anything humanity has previously assembled.

And yet the milestones that once marked a finished adult keep sliding out of reach. Home ownership comes later, if it comes. Marriage is postponed. Birth rates have fallen beneath replacement. Independence is acquired slowly, partially, and late. The doors are visible. Reaching them has become the difficulty.

In the shallow lakes that once ringed Mexico City lives a creature that declines to grow up. The axolotl is a salamander, and like most salamanders it begins life as a gilled, aquatic larva. Unlike most salamanders, it never leaves that stage. It feeds, matures, and breeds while keeping the feathered external gills and soft body of a juvenile. Biologists call the condition neoteny: the retention of larval form into reproductive adulthood. The axolotl is, in the most literal sense, a permanent adolescent. It is perfectly capable of becoming an adult salamander. It simply never receives the signal to do so.

That signal is a surge of thyroid hormone, which in most amphibians is set off by the environment: a drying pond, a shift in chemistry, the pressure to find dry land. The axolotl evolved in deep, stable water where that pressure never arrived, so its metamorphic machinery sits intact and idle. Supply the hormone in a laboratory and the animal will, belatedly, transform into a land-dwelling salamander. Withhold it, and the creature lives out its life in the only body it has ever worn.

This essay is about a generation in much the same position. Not biologically, but economically. A cohort fully equipped for adulthood, suspended in warm and stable water, never receiving the signal to change.

Call it economic neoteny.

Not adolescence in the developmental sense, and not a failure of individual nerve. Economic neoteny is a structural condition: a prolonged retention of dependency, deferral, and provisional status well past the age at which earlier generations had completed the passage to independence. The body is adult. The economic form is not.

We tend to diagnose its symptoms one at a time. Housing becomes a housing problem. Fertility becomes a demographic problem. Loneliness becomes a mental-health problem. Youth joblessness becomes a labour-market problem. Examined together, they look less like four separate ailments and more like one organism that has quietly stopped metamorphosing.

The Pond that will not drain

The evidence is not subtle.

In the United States, the median age at first marriage has climbed almost without pause for seventy years. In 1956 it touched its modern floor, 22.5 years for men and 20.1 for women.[1] By 2025 it had reached 30.8 for men and 28.4 for women, up from 23.5 and 21.1 as recently as 1975.[2] An entire decade of life has been inserted between leaving school and the altar.

Figure 1. The receding altar: median age at first marriage in the United States, 1950 to 2024.

The household data points the same way. In 2025, some 58% of American adults aged 18 to 24, and 16% of those aged 25 to 34, were still living in a parental home.[3] During the summer of 2020 the share of all 18 to 29 year-olds living with their parents reached 52%, around 26.6 million people, the highest figure ever recorded. The previous peak, 48%, dated to the 1940 census at the close of the Great Depression; the low, 29%, came in 1960, after which the line has risen more or less steadily.[4] The pandemic did not create the trend. It merely revealed how full the pond already was.

Europe tells the same story with sharper regional accents. Across the EU in 2024, the average young person left the parental home at 26.2 years.[5] The spread, however, is enormous. In the Nordic countries the water drains early: Finland at 21.4, Denmark at 21.7, Sweden at 21.9. In the south and east it barely drains at all: Spain at 30.0, Italy at 30.1, Greece at 30.7, Slovakia at 30.9, and Croatia at 31.3. Housing is the obvious culprit. Nearly one in ten young Europeans now lives in a household that spends 40% or more of its disposable income on shelter, a heavier burden than the population at large carries.[6]

Figure 2. How long the pond holds them: average age of leaving the parental home across the EU, 2024.

And then the most consequential deferral of all. Across the OECD, the total fertility rate has fallen from 3.3 children per woman in 1960 to 1.5 in 2022, well beneath the 2.1 needed merely to hold a population steady. In Italy and Spain it sits at 1.2; in South Korea, around 0.7. The average woman who does have a child now does so at 30.9, against 28.6 at the turn of the century.[7] A generation that cannot establish itself does not, on the whole, reproduce itself.

Each figure describes a different door. Together they describe a single antechamber, and a pond that refuses to drain.

It would be a mistake to read these as purely financial statistics. A milestone deferred is not only an expense postponed; it is a form of selfhood postponed. The job that does not arrive withholds more than income. It withholds structure, routine, responsibility, a place among colleagues, and the particular dignity of being depended upon. A generation suspended between dependence and autonomy is left neither children nor fully independent adults, and the psychological residue of that suspension, the drift, the disillusionment, the sense of being perpetually almost, is not a separate crisis from the economic one. It is the same condition viewed from the inside.

The Ladder that manufactured Adults

For most of economic history, the transition we are now failing to complete was not left to chance. It was engineered.

England’s Statute of Artificers, passed in 1563, made the point with Tudor bluntness: no person could practise a trade until he had served a seven-year apprenticeship, bound by written indenture, in most cases until the age of 24.[8] The seven years were not merely instruction. They were a slow, supervised manufacture of competence. The master absorbed the cost of the novice’s blunders; the guild guaranteed the standard at the end; and society accepted that an apprentice would produce little of value for years before he produced anything worth selling. The arrangement was rigid, hierarchical, and often exploitative. It was also, in its way, a reliable machine for converting children into adults.

The twentieth century built a faster and broader version of the same machine. In the decades after the Second World War, a compressed and remarkably dependable sequence took hold across the developed world: education, then a job that paid enough to live on, then a home, then a family, frequently before thirty and sometimes well before. The 1956 marriage trough was no accident; it was the demographic signature of an economy in which a single income could secure a house and a young worker could expect his employer to train him. Cheap mortgages, expanding firms with internal promotion ladders, veterans’ education subsidies: these were the warm, shallow waters in which adulthood formed on schedule. The system was unequal and exclusionary in ways we rightly no longer accept. But it metamorphosed people on time.

Labour markets, in other words, used to perform a function we rarely name. They did not only allocate workers. They produced adults. Banks trained analysts; law firms trained associates; hospitals trained residents; newsrooms trained reporters; universities trained researchers. No one started as an expert. Competence was a by-product of being tolerated, for a while, as a beginner, of being permitted to be useless on the way to becoming useful. The ladder mattered as much as the destination, because the ladder was where the metamorphosis happened.

The first crack in that ladder appeared not with the robots but with the paperwork. As more young people earned degrees, employers began treating the degree first as a convenient filter and then as a flat requirement, for work that had never needed one. The economists Joseph Fuller and Manjari Raman documented the phenomenon in a 2017 Harvard Business School study tartly titled Dismissed by Degrees. In 2015, they found, 67% of American job postings for production supervisors demanded a college degree, while only 16% of the people actually doing that job held one. The proportion of secretaries with a bachelor’s degree had tripled, from 9% in 1990 to 33%. By their estimate, credential inflation had placed roughly six million middle-skill jobs out of reach of anyone without a diploma.[9]

The effect was to lengthen the larval stage itself. More years in school, more debt, a later start, for tasks the schooling did not actually require. The pond grew deeper before anyone quite noticed the water was rising.

Credential inflation → delayed entry to work

Delayed entry → delayed income

Delayed income → delayed independence

Delayed independence → delayed adulthood

The Vanishing of the Beginner

This is where artificial intelligence enters, and it enters from an unexpected direction.

Most of the public argument about AI and work concerns displacement in the aggregate. Will it take the jobs? The World Economic Forum’s Future of Jobs Report 2025, drawing on more than a thousand employers across 55 economies, offers the familiar, reassuring arithmetic: 92 million roles displaced by 2030, 170 million created, a net gain of 78 million, set against a churn equal to 22% of the world’s 1.2 billion formal jobs.[10] On that ledger the machines are net creators of work. The IMF, for its part, estimates that around 40% of jobs worldwide are exposed to AI, rising to roughly 60% in advanced economies, where about half of the exposed jobs may be harmed rather than helped.[11]

But the aggregate conceals the structure, and the structure is the whole point. Consider which tasks generative AI performs most fluently today. Summarising documents. Drafting first versions. Conducting preliminary research. Reviewing records. Organising information. Writing routine code. These are not, for the most part, the tasks of the expert. They are the tasks of the beginner. They are the apprenticeship layer itself.

For the whole of industrial history, automation climbed the ladder from the bottom: it took the rungs of physical drudgery first and left cognitive work for last. Generative AI has inverted the climb.

Industrial automation → the lowest physical rungs first

Generative AI → the lowest cognitive rungs first

It is strongest precisely at the entry-level cognitive work through which novices have always become professionals. It does not, at least not yet, replace the senior analyst. It replaces the work the junior analyst was hired in order to learn on.

The first hard evidence is already in. In 2025, researchers at the Stanford Digital Economy Lab analysed payroll records covering some 25 million American workers and published their findings under a pointed title, Canaries in the Coal Mine? Since late 2022, when generative tools became widely available, early-career workers aged 22 to 25 in the most AI-exposed occupations had seen their employment fall by 13% relative to the rest of the workforce, even after controlling for firm-level shocks. In the hardest-hit fields, software engineering and customer service, entry-level employment dropped by roughly a fifth. Older workers in those very same occupations saw their employment hold steady or grow.[12]

Figure 3. The ladder loses its bottom rung: employment change in the most AI-exposed occupations, United States, since late 2022.

The asymmetry is the finding. The senior workers are protected not by seniority alone but by tacit knowledge, the kind of practical understanding that is learned on the job, never written down, and therefore absent from the text on which the models were trained. Bank of America’s researchers flagged a related milestone: for the first time in recent memory, the unemployment rate for recent graduates has risen above the rate for the workforce as a whole.[13] The IMF’s managing director, Kristalina Georgieva, put the asymmetry plainly at Davos, observing that the tasks AI removes first are largely the ones entry-level jobs are made of.[14] The canary in the coal mine is not the economy. It is the beginner.

From Unemployment to Unemployability

Here the vocabulary we have inherited begins to fail us.

Mustafa Suleyman, among others, has argued that the deeper risk is not unemployment but unemployability.[15] The distinction is not rhetorical. An unemployed worker possesses skills for which demand has temporarily lapsed; the cure is a recovering market. An unemployable worker never acquired the skills in the first place, because the rungs on which they were once acquired have been sawn off. The first is a market problem, and we have centuries of practice managing market problems. The second is a developmental problem, and it is far less tractable, because it strikes at the very mechanism by which competent adults are manufactured.

Recall the axolotl. Its metamorphic machinery remains perfectly intact; it simply never receives the trigger. A generation can be in exactly that state: chronologically adult, formally credentialed, fully capable of becoming an accomplished professional, and yet never handed the sequence of beginner tasks through which accomplishment is actually built. The capacity is present. The signal is missing.

And competence, once it stops being produced, is not easily restocked. The senior professionals who currently anchor every field were themselves manufactured by an apprenticeship system that is now quietly closing behind them. Tacit knowledge passes from one cohort of practitioners to the next only if there is a next cohort in the room to receive it. Remove the juniors for a decade and the shortfall does not stay an entry-level shortfall. It ages upward, into a missing tier of mid-career expertise that no model can supply, because the humans who would have held that expertise were never permitted to begin.

The political arithmetic is worth stating too, because societies are not infinitely patient. A cohort that has done everything it was told to do, accumulating the credentials, deferring the family, postponing the home, and still finds the doors narrowing, does not stay quiescent forever. Thwarted expectations are among the most reliable raw materials of unrest. An economy that produces qualifications faster than it produces the means to use them is, over a long enough horizon, manufacturing not only frustrated individuals but a frustrated electorate. The cost of a stalled metamorphosis is paid first by the young and eventually by everyone.

The Hormone that never comes

The institutions whose task is to watch these things are, in their measured way, sounding the alarm.

The WEF reports that 39% of workers’ core skills are expected to become outdated between 2025 and 2030, and that 63% of employers already name the skills gap as their single greatest barrier to transformation. The OECD records that across its member states roughly 13% of people aged 15 to 29 are not in employment, education, or training, a figure that runs to one in five or worse across much of southern Europe; the EU rate stood at 11.0% in 2025, still short of the bloc’s 9% target for 2030.[16] Georgieva has likened the arrival of AI in the labour market to a tsunami. The metaphors are piling up, and they are all metaphors of water.

None of this guarantees catastrophe. History is littered with confident forecasts of technological unemployment that adaptation quietly falsified. New industries appeared; old professions evolved; human ingenuity outran the pessimists again and again. But adaptation has always required a pathway, and a pathway requires that beginners be allowed to exist long enough to turn into something else. The question is not whether technology advances. It will. The question is whether our institutions can refill the conditions under which a beginner can still become an expert.

What induces metamorphosis in an axolotl is a change in its surroundings: the water recedes, the chemistry shifts, the body finally receives the instruction to grow into its other form. The economic equivalents are not mysterious. Housing that an ordinary single income can actually reach. Entry-level work that builds skill rather than merely extracting cheap labour or being automated out of existence. Training arrangements, modern indentures of a kind, that once again make it somebody’s explicit business to manufacture the next cohort of adults. Family formation that does not first demand a decade of deferral. These are the triggers. Their common feature, at present, is absence.

It is worth keeping the cautionary half of the analogy in view. When biologists force metamorphosis on an axolotl by injecting the hormone directly, the animal does transform, but the abrupt change frequently shortens its life. Transitions imposed without the supporting environment tend to go badly. The remedy for economic neoteny is therefore not to shove a generation through the door by decree, with lectures about grit or a sudden withdrawal of support, but to restore the conditions under which the change occurs on its own.

The Warm Water

The most serious challenge posed by artificial intelligence may, in the end, have very little to do with intelligence.

It may have to do with becoming.

For as long as there have been crafts, societies have maintained machinery for turning dependent novices into independent adults: the guild, the indenture, the post-war ladder, the entry-level desk. The machinery was never fair and never gentle. But it ran, and people came out the other side changed. Economic neoteny is the name for what happens when that machinery is allowed to idle. Not a sudden collapse, not a crisis with a date attached, simply a generation that keeps its larval form because nothing in its environment ever tells it to do otherwise.

The danger does not announce itself. There is no alarm. The water stays warm, the credentials accumulate, the milestones recede one quiet year at a time. The doors are still there. The gills are still feathered. And metamorphosis, perpetually available, keeps not arriving.

The pond is not draining on its own. The only question is whether we intend to drain it.

---

[1]U.S. Census Bureau, Estimated Median Age at First Marriage, by Sex (Table MS-2). The modern low of 22.5 years for men and 20.1 for women was recorded in 1956.

[2]U.S. Census Bureau, Families and Living Arrangements, Current Population Survey (ASEC) 2025. Median age at first marriage of 30.8 (men) and 28.4 (women), against 23.5 and 21.1 in 1975.

[3]Same source: in 2025, 58% of adults aged 18 to 24, and 16% of those aged 25 to 34, lived in a parental home.

[4]R. Fry and J. S. Passel, A majority of young adults in the U.S. live with their parents for the first time since the Great Depression, Pew Research Center, September 2020. The July 2020 figure of 52% (26.6 million) surpassed the prior peak of 48% in the 1940 census; the series low was 29% in 1960.

[5]Eurostat, When do young people in the EU leave home? (dataset yth_demo_030), 2024 data. EU average 26.2 years; Croatia 31.3, Slovakia 30.9, Greece 30.7, Italy 30.1, Spain 30.0; Finland 21.4, Denmark 21.7, Sweden 21.9.

[6]Eurostat, Young people, housing conditions, 2024: 9.7% of those aged 15 to 29 lived in households spending 40% or more of disposable income on housing, against 8.2% of the total population.

[7]OECD, Society at a Glance 2024. Total fertility rate down from 3.3 children per woman (1960) to 1.5 (2022); 1.2 in Italy and Spain; about 0.7 in Korea (2023). Mean age at childbirth 30.9 (2022), up from 28.6 (2000).

[8]Statute of Artificers 1563 (5 Eliz. 1 c. 4); see A short history of apprenticeships in England, UK Parliament. Compulsory seven-year apprenticeship, generally to age 24; repealed in the early nineteenth century.

[9]J. B. Fuller and M. Raman, Dismissed by Degrees, Harvard Business School, Accenture and Grads of Life, October 2017. In 2015, 67% of production-supervisor postings sought a degree against 16% of incumbents; secretaries holding a bachelor’s degree rose from 9% (1990) to 33%; roughly 6 million jobs judged at risk of degree inflation.

[10]World Economic Forum, Future of Jobs Report 2025. 92 million roles displaced and 170 million created by 2030 (net +78 million), a 22% churn of 1.2 billion formal jobs; survey of more than 1,000 employers across 22 industry clusters and 55 economies. Also: 39% of core skills expected to be outdated by 2030; 63% of employers cite the skills gap as their leading barrier.

[11]International Monetary Fund, Gen-AI: Artificial Intelligence and the Future of Work, January 2024. About 40% of jobs worldwide are exposed to AI, rising to roughly 60% in advanced economies (40% in emerging markets, 26% in low-income economies); about half of exposed jobs in advanced economies are likely to be negatively affected.

[12]E. Brynjolfsson, B. Chandar and R. Chen, Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence, Stanford Digital Economy Lab, 2025. A 13% relative decline in employment for workers aged 22 to 25 in the most AI-exposed occupations since late 2022 (about 20% in entry-level software and customer service), against stable or rising employment for older workers in the same fields; based on ADP payroll records covering some 25 million workers.

[13]Bank of America Global Research, cited in coverage of the Stanford study: the unemployment rate for recent graduates has risen above the overall rate for the first time in recent memory.

[14]K. Georgieva, International Monetary Fund, remarks at the World Economic Forum, Davos (2024 and 2026), describing AI’s labour-market effect as resembling “a tsunami” and noting that the tasks eliminated first are largely those of entry-level work.

[15]M. Suleyman, The Coming Wave (2023) and subsequent public remarks, on the prospect of widespread unemployability rather than mere unemployment.

[16]OECD, Youth not in employment, education or training (NEET), and Education at a Glance: roughly 13% of those aged 15 to 29 across the OECD are NEET, rising to one in five or more across much of southern Europe. The EU rate was 11.0% in 2025, against a 2030 target of 9%.

Nowadays, one defining feature of modern society is perpetual connection. Social media galore; endless entertainment; an uninterrupted stream of noise. One could argue there is barely a moment left in which someone is truly bored, truly unreachable, or even truly alone.

Never before have people spoken so constantly, shared so much of themselves, or remained so continuously accessible to one another. Concurrently, never before have so many reported feeling emotionally detached, socially exhausted, romantically disillusioned, existentially adrift, and youthfully alienated.

Across much of the developed world, particularly among younger generations, an uncomfortable contradiction has begun to emerge: humanity has become hyperconnected technologically while simultaneously fragmenting socially.

We are amusing ourselves into desensitization, a kind of communal coma.

The lights remain on. The feeds refresh endlessly. Notifications arrive like machine-gun fire, sometimes by the minute. Yet somewhere amidst the scroll, something distinctly human appears to be receding quietly into the background.

As Gen Z increasingly enters adulthood, this contradiction becomes impossible to ignore. This is the first generation raised almost entirely alongside the internet, social media, algorithmic entertainment, and the early integration of artificial intelligence into everyday life. Unlike previous generations, technology was not merely adopted by Gen Z; it formed the environment in which they socially developed. Reality itself became partially digitized from childhood onward.

This is not to suggest that every young adult has become disconnected from the physical world, nor that technology itself is inherently destructive. A neo-Luddite in 2026 only has his future to miss while the world accelerates around him. The issue is far more nuanced, and perhaps far more concerning.

What has emerged is not the disappearance of social interaction, but its mutation.

Human connection increasingly exists in mediated, curated, and algorithmically filtered forms. Physical presence remains, yet digital interaction has overtaken it in both frequency and influence.

Food arrives through applications. Friendships are maintained through applications. Relationships begin, evolve, and often end through applications. Entertainment, political discourse, identity formation, validation, and increasingly even emotional support have become integrated into digital systems fundamentally designed around engagement and retention.

Human connection has slowly been reduced into something measurable: clicks, impressions, activity, engagement.

Metadata and Wi-Fi signals increasingly shape outcomes as much as the people involved

Alienation itself, however, is hardly a modern phenomenon. Long before smartphones and social media, thinkers such as Émile Durkheim identified forms of detachment emerging from industrial modernity. Karl Marx described the alienation of workers from both their labor and themselves within industrial capitalist systems, while Durkheim warned of “anomie” — the breakdown of social cohesion and shared purpose amidst rapid societal transformation.

What modern technology accomplished was not the creation of alienation, but its acceleration.

Industrialization weakened traditional communal structures. Urbanization disrupted extended family networks and local social bonds. Hyper-individualistic consumer culture increasingly reframed human beings as self-contained economic actors rather than members of cohesive communities.

Technology slammed the gas onto an already existing condition.

Social media transformed identity itself into performance, reducing human interaction into quantifiable engagement metrics: likes, reposts, visibility, impressions. Individualism became commercialized. Emotion became commodified. Attention became currency.

And perhaps most dangerously of all, isolation became profitable.

The modern internet increasingly resembles not a public square, but an attention economy in which outrage, anxiety, insecurity, and performance generate engagement, which generates revenue.

Drama sells.

Attention sells.

Alienation sells too.

This is particularly evident among younger generations. Increasingly, identity formation occurs not through stable communities or interpersonal interaction, but through fragmented digital ecosystems driven by algorithms optimized for retention. The result is a generation simultaneously overexposed to information and deprived of genuine grounding.

One may know everything occurring politically across the globe while remaining disconnected from neighbors living a few meters away.

One may possess hundreds of online acquaintances while struggling profoundly with intimacy.

One may be constantly perceived, yet rarely truly known.

The paradox of the digital age is that exposure has become mistaken for connection.

At the center of this transformation lies a broader shift in humanity’s relationship with convenience itself. Modern technological systems are designed to minimize friction. Food delivery eliminates waiting. Streaming eliminates boredom. Online shopping eliminates travel. Algorithms eliminate uncertainty.

Increasingly, digital life reshapes human expectations around immediacy.

Meaningful human relationships, however, have always depended upon forms of friction.

Friendship requires patience.

Trust requires consistency.

Love requires vulnerability, compromise, sacrifice, misunderstanding, and emotional risk.

Human connection derives part of its meaning precisely from the fact that it cannot be fully optimized.

Algorithms, however, are designed to optimize relentlessly.

This distinction matters far more than it initially appears.

As modern life becomes increasingly frictionless, human tolerance for emotional discomfort declines alongside it. Interpersonal relationships become vulnerable to the same logic governing digital consumption: if something becomes difficult, replace it; if something becomes emotionally inconvenient, disengage from it.

Dating applications transform intimacy into marketplaces of perceived infinite alternatives. Social media encourages individuals to construct curated identities optimized for validation rather than authenticity. Emotional expression itself risks becoming performative, a shallow transaction in which emotions are reduced to reactions, emojis, reposts, and recycled references.

The result is not merely loneliness.

It is emotional instability emerging from the collapse of stable social structures, a crisis of the internal world.

Importantly, this phenomenon cannot be understood solely through psychology or culture. Its roots are also deeply economic.

Gen Z enters adulthood amidst rising housing costs, stagnant wages, precarious labor conditions, credential inflation, and declining long-term economic security across much of the developed world. Traditional milestones associated with adulthood — home ownership, stable careers, marriage, family formation — increasingly feel delayed or unattainable for large segments of the population.

The post-war dream is dead.

And perhaps, collectively, we killed it ourselves.

Such conditions inevitably shape social behavior. A generation unable to afford independence often remains economically dependent longer. A generation uncertain about its future naturally gravitates toward escapism, digital immersion, and alternative forms of emotional fulfillment.

Social atomization is not occurring independently of economic pressures; those pressures reinforce it continuously.

Digital life no longer merely distracts from instability.

Increasingly, it compensates for it.

This becomes even more significant with the rise of artificial intelligence and synthetic forms of interaction. AI assistants, emotionally responsive chatbots, recommendation systems, and increasingly human-like digital communication introduce an unsettling possibility: emotional simulation may eventually become easier to access than emotional vulnerability itself.

This is not simply convenience.

It is emotional outsourcing.

If an algorithm can provide endless affirmation, endless responsiveness, endless patience, and endless personalization, many individuals may gradually begin preferring synthetic interaction over the unpredictability of human relationships.

This possibility should not be dismissed as science fiction. In many ways, it has already begun.

Human relationships are meaningful partly because they involve another independent consciousness: a person with agency, unpredictability, flaws, contradictions, and emotional complexity. Genuine connection requires negotiation between separate human realities.

Artificial systems, by contrast, are increasingly designed around personalized emotional satisfaction.

They simulate humanity.

They do not participate in it.

Convenience may therefore begin replacing communion.

Simulation may begin replacing intimacy.

And perhaps most concerningly, many people may not immediately recognize the difference.

None of this suggests that technology itself is inherently malicious, nor that society must retreat nostalgically into some imagined pre-digital past. Technological progress has improved countless aspects of modern life. The issue is not technology alone, but humanity’s inability to socially and psychologically adapt to the consequences accompanying it.

Civilizations ultimately depend upon trust, communal participation, and stable interpersonal bonds. A society increasingly composed of isolated individuals connected primarily through algorithmically mediated systems risks becoming politically volatile, emotionally exhausted, and socially fragmented.

A civilization cannot remain cohesive indefinitely if its members increasingly experience one another as abstractions rather than communities.

The danger facing modern society is therefore not sudden collapse beneath technological progress.

It is something quieter.

Something slower.

Something almost comfortable.

It is the gradual normalization of emotional substitution in place of emotional connection; convenience in place of community; simulation in place of intimacy.

And perhaps the most unsettling part of all is that this transformation does not arrive violently.

It arrives illuminated softly by screens, accompanied by curated feeds, personalized algorithms, and endless entertainment, while somewhere, almost unnoticed, the human experience itself slowly begins fading into the sunset.         

[12 min. read]

On an unassuming July morning a couple of decades ago, a series of Asian economies entered freefall. It was on July 2nd 1997 that the Thai baht — fuelled by hot money and propping up Thailand’s bubble economy — unexpectedly, or rather inevitably, crashed. What followed across the region was a sequence of fiscal collapses so closely choreographed that, in hindsight, they read less like discrete national tragedies than like patients in adjacent beds of the same ward, each presenting different symptoms of the same underlying pathology.

That is the lens I want to apply here. The 1997 Asian crisis is usually told as a country-by-country narrative — Thailand falls, Indonesia burns, Korea restructures, Hong Kong defends. Useful enough as history. But it underplays something more revealing: the crisis exposed an entire architecture of fiscal vulnerability that the global system was, at the time, structurally unwilling to address. Hot money flowed in because pegs invited it. Pegs held because reserves backed them. Reserves drained because corporates had borrowed in dollars they did not earn. And when the fever broke, the IMF arrived — clipboard in hand — to administer treatments whose side effects often outlasted the disease.

Call it the Fiscal Sanitarium. A place where ailing economies are admitted, diagnosed, treated, and — in some cases — billed for the privilege.

The Pathology

Before describing the patients, it is worth naming the pathology. Across the Asian Tigers, four conditions overlapped in the years preceding July 1997:

-> Currency pegs to the U.S. dollar → invited speculative capital -> Speculative capital → inflated asset bubbles -> Bubbles → encouraged dollar-denominated corporate borrowing -> Dollar-denominated borrowing → made local devaluation catastrophic

Each link in that chain was rational on its own terms. Pegs offered exchange-rate stability and lower borrowing costs. Capital flowed in chasing the yield differential. Local corporates discovered they could borrow in dollars at lower rates than in their domestic currencies, and as long as the peg held, this looked like a free lunch. The trouble is that free lunches in macroeconomics tend to come with deferred bills.

When the peg broke, every link in the chain reversed direction at once. Capital fled, asset prices collapsed, and corporates suddenly owed dollars they could no longer service in their devalued local currencies. The bill came due simultaneously across an entire region.

Patient Zero: Thailand

Thailand presented the textbook case. By the spring of 1997 it had accumulated substantial foreign debt and was relying on speculative capital flows to sustain growth. Anticipated interest and exchange-rate differentials allowed Thailand — and its neighbours, Malaysia and Indonesia among them — to capture short-term profits at the expense of longer-term stability. The bubble was not unique to Thailand; it was regional. Thailand was simply where it popped first.

In May 1997 the baht came under successive speculative attacks. On June 30th the government declared it would not devalue. Within forty-eight hours, that statement was inoperative: Thailand lacked the foreign reserves to defend the USD–Baht peg, and on July 2nd the currency was floated. The market took over from there.

The collapse was extraordinarily fast.

-> Massive layoffs in finance, real estate, and construction -> Rural-bound migration of urban workers; hundreds of thousands of foreign workers deported -> Stock market down 75% -> Baht devalued by more than half, bottoming at 56 to the U.S. dollar in January 1998 -> Finance One — the country’s largest finance company — collapsed

By August the IMF unveiled an initial $17 billion rescue, later topped up by $2.9 billion, conditional on bankruptcy law reform and stronger financial-sector regulation. By 2001 Thailand’s economy had recovered enough that it repaid the IMF in 2003 — four years ahead of schedule. A relatively clean discharge from the sanitarium, all things considered.

Hong Kong: the Defended Patient

Hong Kong is the case that breaks the pattern, and it is worth dwelling on. The crisis arrived in October — three months after Thailand, and just months after the territory’s transition from British to Chinese rule on July 1st 1997. The Hong Kong dollar came under speculative pressure because local inflation had been running well above the U.S. rate for years, leaving the peg looking stretched.

Unlike Thailand, however, Hong Kong had ammunition. Foreign reserves stood at over $80 billion, and the Monetary Authority (HKMA) was prepared to use them. More than $1 billion was spent defending the local currency directly. To squeeze out short positions, the HKMA raised overnight interest rates from 8% to 23%, briefly spiking to 280%. The rate hike worked against the currency speculators — but it also drove down equity prices, and speculators promptly shifted to shorting Hang Seng-listed shares.

So the HKMA adapted. It identified that speculators were exploiting the city’s currency-board mechanism, in which overnight rates (HIBOR) automatically rise in response to large net sales of the local currency, while simultaneously betting against the equity market. The HKMA and Donald Tsang — the then Financial Secretary — declared war on them.

The government bought approximately HK$120 billion (US$15 billion) worth of shares directly, becoming, at one point, the largest shareholder of HSBC at 10%. The hostilities ended in late August with the closing of that month’s Hang Seng Index futures contract. In 1999 the government began divesting via the Tracker Fund of Hong Kong, eventually booking a profit of around HK$30 billion (US$4 billion).

A patient who, when admitted, refused to lie down. The defence is often cited approvingly, and it deserves to be — but with a footnote: it was only available because Hong Kong, uniquely, had reserves vast enough and a regulatory architecture flexible enough to mount it. Most patients in the ward did not have that option.

Indonesia: the Fatal Case

Of all the cases, Indonesia was the most disturbing — and the most instructive about what fiscal crises do when they meet pre-existing political fragility.

By June 1997, Indonesia looked, on paper, healthy. Low inflation. A trade surplus of more than $900 million. Foreign reserves above $20 billion. A reasonable banking sector. The textbook would have predicted resilience.

What the textbook missed was the corporate balance sheet. A large number of Indonesian corporations had been borrowing in U.S. dollars, and this had worked beautifully for years, as the rupiah had strengthened against the dollar; effective debt loads had been declining. When Thailand floated the baht, Indonesia’s monetary authorities widened the rupiah’s trading band from 8% to 12%. The currency promptly came under severe attack.

The IMF arrived with $23 billion. It did not stop the slide. Fears over corporate dollar debts, massive selling, and the scramble for hard currency drove the rupiah down further. The Jakarta Stock Exchange touched a historic low in September. By December 1998, Indonesia had lost 13.5% of its GDP.

President Suharto sacked the central bank governor. It changed nothing. The country was already volatile — allegations of fraud in the 1997 legislative election, the 1998 Trisakti shootings, accumulated discontent with Suharto’s rule and his policies toward Chinese-Indonesians. The crisis was the spark in a room already full of accelerant. May 1998 saw violent riots across the country, thousands of deaths, the collapse of the Suharto regime, and the beginning of a turbulent transition.

The lesson is uncomfortable. Fiscal crises do not happen in political vacuums. The same shock that produced an orderly bailout in Bangkok produced a regime collapse in Jakarta. Pathology meets pre-existing conditions.

South Korea: the Conglomerate Disease

South Korea was geographically distant from the Southeast Asian epicentre but no less exposed. Its banking sector was burdened with non-performing loans because its largest corporations — the chaebols — had been funding aggressive global expansions on debt. The strategic ambition was explicit: build conglomerates large enough to compete on the world stage. The economic reality was that many of those expansions failed to generate returns, while the conglomerates kept absorbing capital.

The Hanbo scandal exposed both the weakness and the corruption to international markets. In July 1997, Kia Motors — the country’s third-largest automaker — requested emergency loans. From there the dominoes fell in sequence:

-> Hyundai absorbed Kia (1998) -> Samsung Motors’ $5 billion automotive venture was dissolved -> Daewoo Motors was eventually sold to General Motors

The IMF stepped in with a controversial $58.4 billion bailout. The conditions were severe. The ceiling on foreign investment in Korean companies was raised from 26% to 100%. Financial-sector reform shut down or merged 787 insolvent institutions by June 2003. The Korean won weakened from around 800 to over 1,700 per dollar before recovering. National debt-to-GDP more than doubled, from roughly 13% to 30%.

Foreign ownership of the Korean financial system increased dramatically as a result. Whether one reads that as healthy reform or as the price extracted for treatment depends largely on where one is standing.

The Doctor’s Bill

This brings up the question that the country-by-country narrative often glosses over: what, exactly, did the IMF bring to the sanitarium?

The official answer is liquidity and discipline. Bailout funds in exchange for structural reforms — banking-sector regulation, bankruptcy frameworks, opening to foreign capital, fiscal tightening. In Thailand and Korea, the patients eventually recovered, repaid, and in Korea’s case repaid early. By that measure the treatment worked.

The less official answer is that conditionality was, in practice, an instrument that reshaped each country’s economic architecture along lines that were not always chosen domestically:

-> Foreign-investment ceilings were raised -> Domestic financial institutions were restructured or absorbed -> Fiscal tightening was imposed during recessions -> Sovereign policy autonomy was, for the duration of the programme, conditional

Was this necessary medicine or extracted concession? The honest answer is some of both, in proportions that varied by country and by treatment. The closer one looks, the more the cleanly clinical framing of “the IMF rescued these economies” gives way to something more textured — a process in which crisis created leverage, and leverage produced terms.

Closing the Ward

Malaysia and the Philippines were also affected, though properly accounting for them would require more space than a single essay allows. The Philippines absorbed the shock with relative resilience; Malaysia, controversially, refused IMF assistance and imposed capital controls instead — a heterodox choice that, two decades on, scholars are still arguing about.

What the 1997 crisis ultimately revealed was less a sequence of national failures than the fragility of the architecture itself. Hot money will flow toward yield differentials. Pegs will hold until reserves cannot hold them. Dollar-denominated corporate debt will price in the absence of devaluation risk and will detonate when that absence ends. These are not lessons from a particular cohort of Asian economies in 1997. They are persistent features of any system that combines fixed exchange rates, open capital accounts, and dollar-dominated credit markets — a system whose contradictions Asia happened to expose first, but which has since produced its own variations elsewhere.

The Fiscal Sanitarium, then, is not a place. It is a posture. A way of organising the international response to fiscal sickness that accepts crisis as an event to be managed rather than as a recurring feature to be designed against. Each cycle produces its own patients, its own treatments, its own discharge papers. None of them, so far, has produced an architecture under which the patients stop arriving.

Keep cool. Focus on the facts. The next ward is already filling.

As of writing, a massive intergenerational wealth transfer is underway – if not already happening as we speak. This is not a passive transfer of wealth. It is an active competition over who captures it. Who is preying on whom? To put it simply:

-> Sellers under time pressure → become price-takers
-> Capital exploits opacity → buy quality assets below intrinsic value
-> Intermediaries → extract fees from fragmentation
-> Platforms (future AI layer) → control deal flow

As Baby Boomers and the Silent Generation retire or plan to do so, estimates of trillions of dollars are expected to change hands. The New York Times on May 14th 2023 estimated that by 2045, upwards of $84.4 trillion in assets will be bequeathed. Of that, a staggering $72.6 trillion will be heading directly to the heirs.

Now we are almost halfway done with 2026. What has happened since? McKinsey & Co.’s Institute for Economic Mobility authored a report on the matter in February, and it is quite revealing, albeit limited on the American fiscal frontier. It describes a broken, fragmented, opaque market – exactly the type of system AI historically restructures.

According to McKinsey’s IEM, the issue is not a lack of businesses or buyers. Rather, the challenge is a broken system for transferring ownership. A $5 trillion opportunity exists – not because of innovation, but because of inefficiency. It seems strange but when examined, it makes sense:

-> Inefficiency → opportunity
-> Opacity → extraction
-> Fragmentation → predation

Fragmented markets do not eliminate value – they redistribute it to those best positioned to navigate them.

This is not just a demographic shift – it is a structural moment that will determine who owns the next generation of the economy. The data is more than intriguing:

~6 million SMBs will transition ownership by 2035

~1 million viable firms (~$5T value) could be transferred instead of closed

Failure → mass closures, job losses, and local economic decline

Success → one of the biggest wealth redistribution opportunities in decades

It is important to note that per IEM, most businesses do not get sold – they die. Around 92% of exits are closures, not transfers. Only ~5% sold and ~3% are transferred. Thus, that is the central inefficiency: viable businesses disappear because the system fails, not because they are bad.

So, who will buy them? Most businesses are too small for private equity, yet also too complex for informal buyers. As a result, they fall into a dead zone of capital + advisory support. This is where the majority of closures will happen. Even worse, the buyer side is fragmented. The three main buyer types are:

I. Institutional (PE, corporates) → efficient but only for large deals
II. Independent buyers (ETA, entrepreneurs) → critical segment, but constrained
III. Community/employee buyers (ESOPs, co-ops) → aligned with local stability, but underdeveloped.

Demand exists – it just does not scale.

For a prospective entrepreneur or financially aware citizen, the landscape presents a troublesome flaw: The US built a startup ecosystem – not a succession ecosystem;

1. People do not know buying a business is an option
2. Buyers & sellers cannot find each other
3. Financing is slow and exclusionary
4. There is no support after acquisition
5. Owners plan their exits too late

If done right, up to 12 million jobs will be preserved, $250 billions of local spending will be protected, while supply chains and communities will be strengthened.

If done wrong however, widespread closures and hollowed-out local economies might soon follow.

This brings up another facet of the issue: geography. It matters – a lot.

• Rural areas = highest risk, because of:
  • fewer buyers
  • weaker financing ecosystems
• Urban areas = higher absorption capacity, meaning cities recycle businesses – rural areas lose them.

Faced with all these challenges – what if Artificial Intelligence can soften the blow, if not even prevent tragedies from occuring? This is not science fiction. Rather, it is a concrete proposal to utilize emerging technologies in a productive and constructive manner. AI will not simply ‘fix’ the market – it will redefine who has the power to operate within it.

Market opacity for example, presents an AI opportunity. Given that there is no central marketplace for SMBs – and the deal flow is fragmented and hidden, AI adoption can offer:

1. Matching algorithms,
2. Marketplace aggregation, and
3. Deal discovery engines.

Additionally, financing is slow, bespoke, and manual, while high transaction costs kill deals. AI implementation can not only automate financial analysis but also risk score models and produce faster underwriting – saving precious time and effort.

The landscape is changing, however there is a lack of standardized data. Poor financial documentation and inconsistent valuation practices are serious logistical challenges. Whoever controls the data layer will control valuation – and therefore capture the transfer.

What about the new owners, coming into the picture? Currently there are buyer credibility gaps, as well as a lack of support for the post-acquisition reality. In these cases, AI can offer copilots for operators, automate workflows and create decision support systems for those in need.

That does not mean AI can have solely positive benefits. It is both a tool for concentration and a tool for democratization. It does not inherently democratize markets, rather it scales whoever deploys it best, because:

1. Platforms dominate deal flow,
2. Capital aggregates faster, and
3. Small players get priced out.

However, AI also:

1. Allows individuals to access deal flow,
2. Small acquisitions become viable, and
3. Ownership expands.

In conclusion, value does not disappear in fragmented systems – it is captured. The Great Wealth Transfer is not just a demographic event;

It is a market design failure, with AI becoming the infrastructure layer that determines who captures what.

It is quite paradoxical – there is simply too much supply (businesses for sale) and real demand (buyers), yet the system fails to connect them.

That’s exactly where the next wave of transformation – very likely AI-enabled – will happen. The transfer will occur regardless. The only question is: who is prepared to capture it.