There’s an essay doing the rounds — you may have seen a version of it, there are actually thousands — that begins with a person staring at a screen, re-reading a message for the twentieth time, cursor blinking in the reply box like a pulse. The essay is tender. It tells the reader that their spiralling is not madness but memory, that their hypervigilance is an old survival pattern, that they are not broken, not too much, not hard to love. It ends with a permission slip: to care, to question, to breathe.
It is beautifully written. It’s also, in a way its author almost certainly never intended, a corporate document.
Because here’s what the essay never asks: why does the reader know the message was read? Why do they know it was read at 11:47pm? Why can they see that the other person is “typing…” — and then, agonizingly, that they have stopped? None of this information exists in nature. A letter didn’t tell you it had been opened and set aside. A landline didn’t report that someone had looked at it ringing and decided against answering. Every one of these small cruelties is a design decision, argued over in meetings, A/B tested against alternatives, and shipped for one reason above all others: it made you come back.
The overthinker rehearsing imaginary conversations in the shower believes the problem lives inside them — in their childhood, their attachment style, their wiring. The essay agrees, and consoles them for it. What neither will say is that a trillion-dollar industry has spent two decades building machinery around that wiring the way a casino builds architecture around a gambler. The read receipt is not a feature. It’s a hook. The “last seen” timestamp is not information. It is bait. Silicon Valley did not invent anxious attachment, but ours is the first civilisation in history to have monetised it at scale, and the numbers suggest business is excellent.
And here is where it gets really interesting. A second industry has grown up alongside the first, and its product is the language of consolation. The therapy-content economy: the infinite scroll of gentle reminders, trauma explainers, permission slips, and affirmations, delivered — note this carefully — through the very platforms manufacturing the distress. The machine agitates you, and then sells you, in the adjacent slot, the vocabulary for enduring the agitation. You are not too much. You are healing. Your spiral is valid. Now refresh the feed.
To question any of this is to trespass. Suggest that the wholesale therapeutic reframing of ordinary human difficulty might be making us more fragile rather than less, and you will be told you lack compassion. Suggest that “validation” has become a product category, that trauma has undergone the same inflation as the currency, that an entire generation is being taught to interpret every silence as abandonment and every discomfort as harm — and you will be informed, politely or otherwise, that you are part of the problem. The therapeutic worldview has achieved what every orthodoxy dreams of: it has defined disagreement with itself as a symptom.
I am going to trespass anyway. The tender essay is right about the pain — the pain is real enough; I have felt it, and so have you. But consolation without diagnosis is just anaesthetic. And nobody hands out free anaesthetic unless they are planning to keep operating.
The Machinery
Let’s begin with a gesture you have made, conservatively, a hundred thousand times: the pull-to-refresh. Thumb on glass, a downward drag, a moment of suspense, and then — perhaps — something new. The man who invented it, a developer named Loren Brichter, built it for an early Twitter app and has since expressed something close to remorse about it, because he understands better than most what he made. It is a slot machine lever in everything but appearance. The psychological mechanism it exploits — variable ratio reinforcement, the schedule of unpredictable reward — is the single most compulsion-forming pattern behavioural science has ever identified. B.F. Skinner discovered it in pigeons in the 1950s: a bird rewarded unpredictably will peck a lever with a desperate persistence that a bird rewarded reliably never shows. Reliability breeds calm. Uncertainty breeds obsession. Every casino on earth is a monument to this finding, and so is every “smart” phone.
This is not an accident of engineering, and the industry has stopped pretending otherwise. Sean Parker, Facebook’s founding president, admitted as much in 2017: the platform was consciously built to exploit “a vulnerability in human psychology,” to deliver “a little dopamine hit” at unpredictable intervals, and the men who built it — “me, Mark, Kevin Systrom, all of these people” — understood this and did it anyway. Chamath Palihapitiya, who ran Facebook’s growth team, went further, confessing “tremendous guilt” over building “short-term, dopamine-driven feedback loops that are destroying how society works.” These are the architects themselves, reading out the blueprints.
Nor did the knowledge arrive by accident. Stanford has run a Persuasive Technology Lab since the late nineties, where a generation of founders and product managers were formally trained in the science of engineered compulsion, in techniques for creating what the discipline calls, without apparent embarrassment, “behaviour change.” Its alumni fanned out across the Valley. The industry’s own handbook, Nir Eyal’s Hooked, lays the method out in four cheerful steps — trigger, action, variable reward, investment — and was for years handed to product teams the way sales teams are handed scripts. The goal, stated openly, is to move the user from external prompts to internal ones: to reach the point where the trigger is no longer a notification but a feeling. Boredom. Loneliness. Uncertainty. Anxiety. When your own discomfort becomes the push notification, the product has achieved what the trade calls habit formation and what any honest observer would call something less polite.
Now look again at the “read” receipt through this lens. Ask the product question: what does it optimise? Communication managed perfectly well without it — better, if anything, since messages once carried a merciful ambiguity that gave both parties room to breathe. What the “read” receipt optimises is return visits. It converts a closed loop into an open one. Before it existed, you sent a message and the matter left your hands, like a letter into a postbox. After it, you sent a message and acquired a small live wound to monitor. Seen, 11:47pm. No reply. The “typing…” indicator refines the torture to an exquisite pitch: it shows you the other person composing, reconsidering, abandoning. It’s a window into hesitation itself. No communication purpose on earth requires you to watch someone decide not to answer you. But an engagement metric requires it very much, because the anxious user checks the thread forty times before bed, and each check is an impression, and impressions are inventory, and inventory is revenue.
This is the plain economics that the entire wellness conversation tiptoes around. For all their talk of connection, the platforms are advertising companies: attention is the product, advertisers are the customers, and you are the crop. And here’s the uncomfortable arithmetic: a calm, secure, contented user is a low-yield crop. She checks her phone only when she has a reason to. The anxiously attached user — the one rehearsing conversations in the shower, the one decoding the two-minute delay — is high-yield. He checks compulsively, scrolls to self-soothe, returns to the platform in precisely the way a gambler returns to the table, chasing relief, never winnings. If you were designing for revenue and revenue alone, you would design for him. You would design, in fact, to produce more of him — to nudge the secure toward vigilance, because vigilance monetises.
Did they know? They knew. When Facebook’s own internal research into Instagram’s effects on teenage girls leaked in 2021 — findings that the product made body-image issues worse for a substantial fraction of the girls who used it, that teens themselves blamed the app for rising anxiety and depression, in the researchers’ own slides, in the company’s own words — the company’s response was to dispute the reporting rather than change the product. The parallel that keeps suggesting itself is tobacco, and it is worth stating precisely why it holds: the manufacturer’s own scientists documented the harm internally while the company’s public posture remained one of denial, and the product’s profitability depended on the very property — compulsion — that caused the harm. That is an industry in which the harm is the business model.
The Consolation Trade
If the story ended there, it would be a familiar one — Big Tech behaving like Big Tobacco, nothing we haven’t seen before. But something stranger has happened in the last decade, and it’s the part nobody wants examined. A second economy has grown up inside the first, symbiotic with it, and its product is the language of psychological consolation.
You know its voice the moment you hear it. Sans-serif font on a pastel background. Gentle reminder: you are not responsible for other people’s reactions to your boundaries. Your anxious attachment is not a flaw, it’s an adaptation. Normalise needing reassurance. It’s the voice of the essay we began with — warm, lowercase, infinitely forgiving — and it now constitutes one of the largest content genres on the platforms I have just described. Therapy-speak has become an aesthetic, an idiom, a career path. There are accounts with follower counts in the millions doing nothing but this, and behind the largest of them, sponsorship deals, merchandise lines, and app partnerships — frequently, and one must admire the elegance of it, partnerships with subscription mental-wellness apps, so that the content which names your wound can route you directly to the storefront that rents out the salve.
Let me be a little cautious here, because this is the point where the argument is most easily caricatured. My claim is not that therapy is a racket. Actual therapy — the private, effortful, frequently uncomfortable kind, conducted over months with a trained human being who is permitted to challenge you — has evidence behind it and has saved lives, and nothing here disputes that. The claim is that therapy-content on social media is to therapy what a slot machine is to an income: a superficially similar object with a precisely inverted function. Think about the differences. Therapy is private; content is performed. Therapy is tailored to one person’s actual history; content is calibrated to the broadest possible resonance, which means it must describe everyone, which means it describes no one. Therapy aims at change, and change is uncomfortable, and a good therapist will therefore sometimes make you baulk; content aims at engagement, and nothing engages like flattery, and so the genre converges — inevitably, algorithmically — on the message that you are fine exactly as you are and everyone else is the problem. Above all: therapy is designed to end. You are meant to graduate. Content is designed never to end. You are meant to subscribe. One is medicine. The other is a lozenge shaped like medicine, and the difference between them is the difference between treatment and retention.
And because the genre is governed by engagement rather than accuracy, its vocabulary has undergone a systematic inflation. The psychologist Nick Haslam gave the phenomenon a name in 2016 — “concept creep” — and documented it with some rigour: the steady horizontal and vertical expansion of harm-related concepts, so that words which once denoted severe and specific conditions come to cover ever-milder and ever-broader experience. Watch how certain words travel. Trauma, a term once reserved for the aftermath of events that overwhelm the psyche’s capacity to cope, now routinely describes an awkward childhood dinner. Gaslighting, which named a deliberate, sustained campaign to make a victim doubt her sanity, now means “disagreed with me.” Toxic means unpleasant. Boundaries, a genuinely useful clinical concept about what you will and won’t accept, has been quietly repurposed into a mechanism for issuing demands with therapeutic immunity. Emotional labour, coined by the sociologist Arlie Hochschild to describe the management of feeling as a job requirement — the flight attendant’s mandatory smile — now means being asked a favour by a friend.
This inflation is currency debasement, however compassionate its intentions. Every time trauma is spent on a minor unpleasantness, the word buys less for the person who needs it whole and intact — the refugee, the assault survivor, the soldier — and the culture’s finite reserves of urgent concern are diluted across an ever-expanding claimant class. Worse still, the debased vocabulary re-describes its user to himself. Language does more than label experience; it is the lens through which experience arrives. Teach a young person to file every discomfort under harm, every conflict under abuse, every unreturned message under abandonment, and you have done something far more consequential than validate their experience: you have restructured it. You have taught them, with the best intentions and the softest fonts, to experience ordinary life as a sequence of injuries. There’s a name in the clinical literature for the interpretive style that treats ambiguous events as threats and discomfort as danger: it is called catastrophising, and cognitive therapy was invented precisely to treat it. We have built an economy that trains it.
And here we reach the orthodoxy’s most heavily defended core belief. Because you are not permitted to say any of this. The therapeutic framework has performed the neatest trick available to any belief system: it has pre-classified its critics. Object to concept creep and you are “invalidating.” Suggest that resilience might be teachable and you are “victim-blaming.” Note that an identity organised entirely around one’s wounds might be a cage with excellent upholstery, and you will be told your observation is itself a form of harm — at which point the conversation is over, because harm cannot be debated, only apologised for. A framework that converts disagreement into evidence of the disagreer’s pathology has ceased to be a framework and hardened into a faith. And like most faiths, it deserves more respect for its intentions than for its immunities.
The Closed Loop
Now braid the two strands, because separately they are stories and together they are a structure.
The platform creates the unease. The content economy then explains it as a problem with you. Your attachment style. Your childhood. Your nervous system. Your trauma. Read ten thousand posts about the anxiety of waiting for a reply, and every single one will trace the pain back to your personal history. None will trace it back to the “read” receipt itself.
The app was deliberately designed to make you check, monitor, and lie awake at 11:47pm staring at your phone. But in these posts, the app is treated as a neutral backdrop — just weather, while your suffering is framed as a private storm. The design goes unmentioned. The designers go unmentioned. The only things left in the picture are you and your past.
We have seen this manoeuvre before. It’s worth naming the precedent exactly, because the mechanics are identical. In the early 2000s, BP — an oil company — hired the advertising firm Ogilvy & Mather and popularised a phrase you now use without thinking: the personal carbon footprint. The company built a calculator so you could measure yours. The genius of it was jujitsu: take a structural problem, produced overwhelmingly by an identifiable industry that profits from its continuation, and convert it into a private moral journey for the individual. Once the frame took hold, the argument was over before it began. The question was never again “what should be done about the industry?” but “have you done enough?” — and a public busy auditing its own recycling does not audit refineries.
The anxiety economy runs the same play with arguably better production values. Structural problem: an attention industry engineered for compulsion, documented internally, denied publicly. Reframe: your relationship with your phone. Your screen-time goals. Your digital hygiene. Your healing journey. The moment the frame lands, the “read” receipt has the same status the refinery had — regrettable weather — and the entire weight of the problem transfers to your discipline, your childhood, your self-work. And the self-work, conveniently, is a subscription. The meditation app, the therapy-adjacent chatbot, the workbook, the course, the retreat. The pattern completes itself with a symmetry that would be funny if it were fiction: the arsonist has gone into the fire-extinguisher business, and the extinguishers are sold on a monthly plan, and — the final flourish — they are advertised to you through the very feed that started the fire, targeted by an algorithm that knows you are anxious because it made you that way.
There is one more turn of the screw. In the old confidence tricks, the mark eventually ran out of money. In this one, the mark never runs out, because the product being extracted — attention, engagement, data — regenerates daily, and because the consolation economy has solved the problem of churn by converting the wound into an identity. You are no longer a person who is sometimes anxious. You are An Anxious Person, an over-thinker, an empath, a survivor with an attachment style, and your feed is curated to your diagnosis, and your diagnosis is your community, and your community is on their servers. Healing has become a hobby that must never be completed, pursued in the very casino that necessitates it. The loop is closed. Both industries feed, and each provides the other’s alibi: the platform can point to the flourishing wellness conversation as evidence of its social value, and the wellness economy requires the platform’s damage as its perpetual market. I think it’s the most elegant machine our civilisation has built, and it runs on the oldest fuel there is — the human fear of being abandoned.
The Cost
What does the loop actually extract? Attention, obviously — the hours are countable and the numbers are obscene. But the deeper extraction is harder to see because it takes the form of an absence.
Start with the epidemiological record, and let’s handle that one honestly. Beginning around 2012 — the year smartphone ownership crossed half of the population and the front-facing camera met the algorithmic feed — the indicators of adolescent mental health in the Anglophone world bent sharply and simultaneously in the wrong direction. Rates of depression, anxiety, self-harm admissions, and suicide among teenagers, especially girls, rose after two decades of relative stability, and the researchers who have assembled this record most prominently, Jean Twenge and Jonathan Haidt, note that the timing, the gender skew, and the cross-national synchrony are difficult to explain by any candidate other than the one in everyone’s pocket.
Now the steelman, because credibility is worth more than a tidy narrative. Serious critics — Candice Odgers, Amy Orben, Andrew Przybylski among them — reply that the correlational studies show small effect sizes, that some of the rise reflects better screening and reduced stigma and changed diagnostic thresholds, that causation is genuinely hard to isolate, and that we have panicked before, over novels and radio and television, and been wrong. These objections are real and should be weighed rather than waved at. But note what the strongest version of the sceptical case concedes: at best, it downgrades the verdict from “proven cause” to “prime suspect with motive, means, opportunity, and a confession from the founding president.” And note, too, an asymmetry in the burden of proof. The platforms deployed a persuasion architecture on two billion minors without a single safety trial, and the demand for definitive causal evidence is now issued to the critics, as though the null hypothesis were that an engineered compulsion machine is harmless until the professors prove otherwise. Tobacco enjoyed the same courtesy for forty years. We don’t have to extend it twice.
But the ledger’s largest entry is not diagnostic at all, and no epidemiology will capture it. It’s the slow atrophy of two capacities that turn out to be the same capacity: the tolerance of uncertainty, and the ability to love.
Consider what every intimate bond actually requires. To be known by another person is to hand them the means of your injury and trust them not to use it. There is no version of authentic connection without exposure, no exposure without uncertainty, no uncertainty without the ache the tender essay described so well — the pause, the silence, the unanswered question. The capacity to bear that ache without resolving it instantly is not incidental to love. It is love’s keel. And it’s precisely this capacity that the loop dismantles from both ends at once. The machine trains you out of tolerating uncertainty by promising its abolition — the read receipt, the location share, the last-seen stamp all whisper that ambiguity is a solvable problem, that with enough monitoring you need never sit in the unknown again. And then the consolation economy finishes the job by reclassifying the ache itself as pathology: your discomfort in uncertainty is your trauma response, your activation, your wound. Between the two of them, the ordinary weather of intimacy — the pause before a reply, the opacity of another mind — is rebranded as both technically unnecessary and psychologically dangerous. Is it any wonder the surveys find the rising generations lonelier than any cohort ever measured, dating less, touching less, deferring every milestone that requires exposure to another person’s freedom? We told them the unknown was a threat and sold them instruments for its surveillance. They believed us. The instruments work. And now nothing is unknown, and nothing is intimate, and the reply box glows at 11:47pm like the last lit window on an empty street. That’s the cost. Not the hours. The keel.
Defection
So what’s to be done? We must first clear away the answers the racket itself prefers, because a system this elegant doesn’t leave its exits unguarded.
The approved remedies are by now familiar: the digital detox, the screen-time dashboard, the greyscale mode, the mindfulness app, the phone-free Sunday. Notice what every one of them has in common. Each preserves, intact, the frame we previously identified — the problem is your discipline, your hygiene, your relationship with the product — and each therefore belongs to the loop it appears to oppose. The screen-time dashboard is built by the phone’s manufacturer. The mindfulness app is sold through the store that takes its thirty percent. The detox ends, as all detoxes end, in relapse, with the added interest of self-blame. These are the fire-extinguisher subscriptions again. A casino that hands out pamphlets on responsible gambling has simply purchased an alibi for its business model. The personal remedy, pursued personally, against an industrial adversary with a persuasion budget larger than most national school systems, is not a fair fight and was never meant to be one.
Which forces me to the structural answer, and to the first of this essay’s genuinely impolite proposals: regulate persuasive design as we regulate other engineered compulsions — as a problem of product safety rather than of speech. The distinction matters, and it disarms the reflexive objection. Nobody proposes censoring what anyone says on these platforms; the proposal is aimed at the machinery underneath, which says nothing and simply pulls. We already accept, without controversy, that a car must have seatbelts, that a casino may not admit children, that cigarettes may not be advertised to them, that a slot machine’s reward schedule is a licensed and inspected artefact. The variable-reward feed, the autoplaying next video, the streak mechanic that punishes a teenager for a day’s absence, the “read” receipt itself — these are reward schedules and pressure mechanisms, engineered by professionals to produce compulsion, and they are currently subject to less safety scrutiny than a toaster. A serious regime would treat them accordingly: mandatory disclosure of engagement-optimisation experiments run on users, as we require disclosure of drug trials; the right to use any communication platform with the compulsion architecture switched off, as a default for minors and an option for everyone; liability, actual and financial, when a company’s internal research documents harm that its public statements deny.
None of this is exceptional, and none of it waits on invention. Australia has legislated a minimum age of sixteen for social media accounts. The European Union’s platform law prohibits interfaces designed to deceive or manipulate, and bans advertising targeted at minors. Britain’s design code for children requires the most protective settings to be switched on by default. And China — whatever one makes of its motives — caps the domestic version of TikTok at forty minutes a day for children under fourteen, a limit the same company has never volunteered anywhere its regulators didn’t demand it. The blueprints exist on four continents. What the project still lacks is permission — withheld because the industry that would be regulated is also the industry through which we now conduct these arguments.
And one detail shows how narrow the gap between the possible and the actual has become. The read receipt can be switched off. Tonight, in about four taps: the setting exists on every major messaging platform, buried a few menus deep, switched on by default. On the largest of them it carries a toll — disable your receipts and you lose the ability to see anyone else’s — so the exit is framed as a loss and taken, the platforms know the figure precisely, by almost no one. A parent who does nothing else after reading this can do that before the phone goes back into a child’s hand. The buried setting proves the larger point: the compulsion architecture can be turned off. It was simply never meant to be found.
But regulation, even if it comes, only disarms the machine. It doesn’t rebuild what the machine dismantled. So my second proposal is cultural, and it’s the one that will get me called cruel: we must recover the unfashionable idea that some suffering is simply weight — and that carrying weight is how a person becomes strong.
Every previous generation of humans possessed this idea in some form. The Stoics built a philosophy on it; every religious tradition worth the name has a fast, a vigil, or a desert in it; every apprenticeship, every basic training, every rite of passage ever devised operates on the premise that the self is extended through voluntary contact with difficulty. From a distance this can look like masochism. It was load-bearing knowledge — the understanding that the capacities a person most needs, such as patience, courage, and the tolerance of the unknown, are built exactly and only the way muscle is built: against resistance. The therapeutic orthodoxy did not refute this knowledge. It simply reclassified the resistance as harm, and the training as trauma, to the extent that within a generation the entire technology of human strengthening became unspeakable — surviving only in gyms, the one arena where we still permit ourselves to believe that stress makes you stronger, provided the only thing strengthened is tissue.
I will state it plainly. The unanswered message is a weight. The ambiguous silence is a weight. The discovery that another person is free — free to pause, to hesitate, to leave — is the heaviest weight in ordinary life, and there are only two things a civilisation can teach its young to do with it. It can teach them to carry it, in which case they grow the strength that every love and every friendship will eventually require of them. Or it can teach them to put it down instantly — to resolve it with surveillance, to medicate it with scrolling, to reclassify it as harm and demand its abolition — in which case the strength never develops, and the weight, when it finally arrives unrefusable, as it does in every life, in the hospital corridor and the empty house, falls on a person who has been carefully trained to bear nothing.
What does the teaching look like in practice? A kitchen, eleven at night, a fourteen-year-old staring at seen, 10:52pm and no reply. The parent’s whole art in that moment is to sit down beside the weight without lifting it — to resist diagnosing the silence, and to resist filling it, because the silence is exactly what the child needs to discover is survivable. Company in the dark, and no rescue from it. Every generation before ours understood this instinctively, at bedsides and gravesides; ours must relearn it against a machine whispering, from the same glowing rectangle, that the dark is optional.
The tender essay doing the rounds said: you are not broken. Fine. Agreed. But the sentence has a second half the genre never writes, and here it is: you are not broken — and you are also not finished, and the difference between consolation and care is that care is willing to say so.
None of this requires abandoning tenderness. The permission slip and the barbell are not opposites; every decent coach and every good parent has always issued both at once. What it requires is refusing the loop’s central bargain — comfort in exchange for the frame — and its two lies, told from its two tongues: the machine’s lie that uncertainty is a defect in the world, and the consoler’s lie that your difficulty carrying it is a defect in you. Neither is true. Uncertainty is the medium of freedom, yours and everyone’s, and the ache of it is nothing gone wrong. It is the feeling of another person mattering.
Return, then, to where we began: the figure lit by the screen at 11:47pm, cursor blinking in the empty reply box like a pulse. The tender essay saw a wounded person and offered a blanket, and there are nights when the blanket is the right and only thing. But look once more at the image, closely, because we misread it the first time. The cursor was never a pulse. A pulse belongs to a living creature there in the room with you. The cursor is a meter — and it has been running all night, billing someone, and it was never billing the person who didn’t reply.
Put the phone face down. Not as hygiene, not as a detox, not as another private discipline in the racket’s approved style — but as a small act of defection from a machine that has told you two lies at once and charged you for both. The silence on the other end of the thread is no verdict. It’s just the old, ordinary, survivable dark that every human who ever loved anyone has had to sit in, long before anyone found a way to sell tickets to it.
Sit in it. It will not kill you. It never did. That was always the secret, and it cost nothing, and no one could make a nickel from it — which is exactly why no one told you.