I have been thinking about the Internet in much the same way that someone might think about a house they once loved, returned to after twenty years, and discovered that the doors had been removed. Not merely unlocked. Removed. The furniture is still there, there are familiar voices somewhere in the building, and every room retains a trace of what made it special. But strangers are walking through it now. Some are inspecting the walls. Others are shouting their opinions about the people who live there. A few are recording everything. The extraordinary thing is that the residents have been persuaded this is simply what a house is supposed to be.
That is increasingly how our digital lives appear to me. I find myself becoming resensitised to something I had been trained to regard as completely ordinary: the obscenity of the access we have granted to one another, to strangers, and ultimately to enormous companies whose business models depend on our willingness to remain available. We have given people access to our time, our photographs, our private relationships, our political thoughts, our movements, our insecurities and, perhaps most bewilderingly, the right to pass judgment on all of these things in public. We have acquired a corresponding expectation that we must watch, respond, compare, defend and explain. Somewhere along the way, access became an entitlement and participation became an obligation.
I remember an earlier relationship with the Internet. The 1990s version was neither innocent nor perfect, and I have no interest in pretending that an old computer and a dial-up modem constituted some lost moral paradise. There were ugly corners, obnoxious people and risks that were poorly understood. Yet the arrangement felt intelligible. You went online to do something. You checked an email, visited a website, downloaded something, learned about a subject, or entered a chat room that was recognisably a destination. The chat room might have been strange, crude or even frightening, but it had a threshold. It resembled walking into a particular club, knowing the kind of people who congregated there, and deciding whether to stay.
Today the club has been annexed to the street, the supermarket, the family dinner and the workplace. Its patrons have been invited into every other setting, and the basic assumption is that everyone should welcome their participation. A photograph of a birthday becomes a public occasion for evaluation. A person's appearance becomes material for strangers. A tentative thought becomes a permanent statement that can be recovered, circulated and litigated years later. An artist shares something made with care, and the first response is often a row of comments from people who have confused their ability to react with some inherent right to be heard. There is an astonishing social experiment concealed inside this arrangement, and we have been conducting it on ourselves for years.
Researchers Alice Marwick and danah boyd described one aspect of this problem as 'context collapse'. Different audiences that would normally occupy different social situations are pressed together into a single online space. Friends, relatives, colleagues, acquaintances, critics and strangers encounter the same expression, although they have entirely different relationships with the person expressing it. Ordinary human beings adjust how they speak depending on where they are and whom they are with. That flexibility is part of social intelligence. A platform that collapses those distinctions then encourages the person to perform an oddly flattened version of themselves for everybody simultaneously. [1]
It is worth thinking about how profoundly unnatural that arrangement is. Imagine opening your front door on a Saturday morning and finding a queue of people who have decided they are entitled to inspect your living room. Some know you; many do not. They comment on the colour of the walls, your choice of breakfast, your parenting, the way you speak to your partner, and whatever they believe your possessions reveal about your character. They assure you that you can always ignore them. But they keep arriving. A few photograph your reactions. Others summon their friends. Eventually a business offers to help you manage the queue, provided you agree to let it measure everything that happens in your house.
We would recognise this as an absurd, invasive and unhealthy situation in physical life. Digitally, we have given it an interface, a notification sound and a friendly name. 'Engagement' is a particularly revealing piece of vocabulary. It is capable of describing an affectionate exchange between friends and a torrent of hostile comments with equal satisfaction. The technology does not necessarily care which kind it receives. Even where platforms attempt moderation, the underlying arrangement grants an audience and a reaction mechanism to almost everything. Experimental work on uncivil online discussions finds that different forms of hostile commentary can produce anger, anxiety and disgust, and can affect people's willingness to participate. Our discomfort with the comment section is not purely a matter of temperament. [2]
I have come to regard the universal comment section as one of the Internet's strangest inventions. There are places where discussion is desirable, even wonderful. A devoted community discussing an obscure film, a technical forum helping someone repair a machine, or a group of friends sharing ideas can demonstrate the generosity of the medium. But the assumption that every piece of information and every personal expression should arrive with a permanently open amphitheatre attached to it is extraordinary. We never agreed as a civilisation that every restaurant meal, family photograph, creative work, announcement or article required a public hearing. We drifted into that convention because the machinery rewarded it.
The result has been an oversharing of access and an oversharing of obligation. We are expected to be visible, and then expected to manage the consequences of being visible. We are expected to provide updates, and then expected to digest everybody else's. We are expected to have opinions about things we would never otherwise have encountered, to know why someone we have never met is being denounced, to understand a dispute that erupted while we were asleep, and to return to the same environment the next morning to discover which version of the dispute has survived. It is a peculiar way of spending the limited span of a human life.
This is where the subject connects to something I have written about elsewhere, the freedom of manufactured obligations and the machinery of FOMO. Fear of missing out is usually presented as a personal insecurity, but its daily force is also constructed by the systems that supply endless things to miss. The feed creates an event, makes it feel urgent, distributes a consensus about it, and then informs you that your silence has consequences. Even leisure begins to acquire the texture of unpaid administrative work. You are keeping up with your friends, keeping up with strangers, keeping up with culture, keeping up with politics, keeping up with the latest outrage, keeping up with the people who are keeping up with all of it. The list cannot be completed because incompletion is part of the design.
For companies, this arrangement has had an additional benefit. Users provide the material, reveal their preferences, disclose their relationships and monitor their own standing among others. They voluntarily report the details of their lives while the service measures what they do with them. I think of it as a culture of self-surveillance: a society in which we have been persuaded to participate in our own observation, and to interpret the act of making ourselves legible to a platform as evidence of connection. The United States Federal Trade Commission's 2024 examination of major social-media and video-streaming businesses described extensive collection and retention of personal information, including information obtained beyond activity on the platforms themselves. The commercial surveillance was not an eccentric suspicion. It was a documented business practice. [3]
The language used to normalise that practice deserves scrutiny. Companies promise personalisation, relevance, connection and community. Each of these words can describe a genuine benefit. I have benefited enormously from the Internet, and I am still grateful for the things it allows me to learn, create and share. But the sales language conceals a remarkable transfer of power. A system observes what attracts us and then becomes increasingly skilled at arranging what we see. We come to think of this arrangement as our chosen environment, even when its architecture has been designed by people whose objectives are not our own. The problem is particularly clear when a person's attention is treated as the resource to be extracted, maintained and repeatedly returned to the service.
There is a difference between choosing to learn about the world and living inside an apparatus that chooses the world's next interruption for you. I want to be able to read about a subject because I am interested in it, to think about it without being immediately shown the reactions of a thousand other people, and to decide when that encounter is finished. I want the ability to close a book, walk outside, have a conversation, look after my family, or simply exist without supplying an audience with evidence that I am doing so. These sound like modest ambitions. Their modesty is precisely what makes the situation so strange.
The comparison mechanism may be the most intimate part of the problem. 'Comparison is the thief of joy' is a line often attributed to Theodore Roosevelt, although researchers tracing the quotation have found no reliable evidence that he said it. Its actual origin remains uncertain. The sentiment does not depend on a famous signature. [4] Human beings have always compared themselves with their neighbours, their friends and the apparently more fortunate. Social platforms have industrialised that tendency, turning a local habit into a continuous encounter with curated fragments of other people's lives. A person is no longer comparing an ordinary Tuesday with a neighbour's ordinary Tuesday. They may be comparing their own unedited interior life with hundreds of exceptional moments selected for display.
There is now research to accompany that intuition. A 2026 meta-analysis of twenty studies found a modest positive association between problematic social-media use and social comparison. Such an association cannot tell us that comparison always causes problematic use, or that the effects are identical for everybody. It does suggest that the two often travel together. [5] I cannot think of a compelling reason to invite a machine to supply me with fresh occasions for envy, insecurity or inadequacy every time I am bored. Why would I build such a mechanism into the centre of my daily life? Why would I accept it as the price of belonging?
The language of addiction is tempting here. People speak about the heroin or cocaine of the scrolling feed, and I understand what they are reaching for: the experience of returning to something you did not particularly enjoy, simply because you have acquired the impulse to check it again. It would be careless to pretend that every social-media habit is medically equivalent to substance dependence. The research is more complicated, and patterns of use vary enormously. Nevertheless, compulsive checking, difficulty disengaging and the deliberate cultivation of habitual return are legitimate concerns. We should be able to identify an unhealthy relationship with a product without waiting for a metaphor to become a diagnosis.
The more persuasive evidence, for me, concerns what happens when people stop. In a randomised trial published in 2022, researchers assigned 154 participants either to continue using certain social platforms or to take a one-week break. The break group reported better wellbeing and reductions in anxiety and depression over that short period. A larger study published in 2026, analysing Facebook and Instagram deactivation during the 2020 United States election period, also found small improvements in measures of emotional wellbeing. These are bounded findings, not a promise that every person who logs off will become happy. But they challenge the assumption that continuous participation is automatically good for us. [6][7]
My own experience is much simpler than a research paper. I stepped away. The anticipated disaster did not occur. My friendships did not evaporate. I did not become an exile. My interests did not disappear. My family remained my family, and the people who mattered remained capable of contacting me. The bubble popped and I was still here. I am alive. I have a life. And the longer I remain outside the mechanism, the more absurd it seems that I ever treated the mechanism as a condition of having one.
Something shifts in the quality of attention. You stop anticipating the commentary that will follow an experience. You stop imagining an invisible audience standing beside you, ready to convert something you have seen into a verdict on your taste or a measurement of your social standing. You are permitted to enjoy a film without contributing to its discourse. You can have an opinion that is still forming. You can meet a friend and leave the conversation where it belongs, between the two of you. Even a silence can remain a silence rather than a failure to respond. This is what I mean when I talk about retaining one's faculties. Attention, memory, judgment and curiosity are parts of a life, not raw materials that must be continually submitted for processing.
I do not want to romanticise withdrawal as a universal solution. Some people rely on online communities for disability access, isolation relief, professional opportunities, solidarity and genuine friendship. Others use social media to organise politically, reach audiences previously denied to them, or find help they cannot find locally. Those benefits are real, and no serious argument about the harms of platforms should erase them. Neither should we confuse abstention with a moral ranking of individuals. Someone who chooses to remain online is not necessarily unfree, and someone who leaves is not necessarily wise. What interests me is the recovery of the choice itself, including the possibility that a connection can be sustained without allowing a company to supervise its every expression.
There are signs that people are beginning to renegotiate that choice. Pew Research Center reported in 2025 that 44 per cent of American teenagers said they had cut back on social media, compared with 39 per cent when asked in 2023. Its 2026 research on news consumption found that about half of American adults felt worn out by the amount of news, and six in ten said they had reduced their overall news intake at some point. These figures do not show a universal exodus, and social platforms remain immensely popular. They do suggest that fatigue and deliberate disengagement are substantial experiences, rather than eccentric attitudes held by a few people who have simply failed to adapt. [8][9]
I find myself increasingly interested in the return of longer forms of attention. An essay that takes a quarter of an hour to read. A letter to a friend. A book that does not stop to ask whether I agree with its previous paragraph. A conversation that wanders until it reaches something worth saying. The expansion of paid newsletters and direct reader subscriptions has given writers another way to reach an audience without relying entirely on the open social feed. This does not establish that the world's attention spans are recovering, and newsletter platforms can reproduce social-media incentives of their own. It does demonstrate a continuing appetite for deliberate, sustained relationships with writers and subjects. [10]
In another series of essays, I have been thinking about the idea of picking back up where we left off. This belongs to that same line of thought. There are parts of ordinary human experience that did not cease to exist when the smartphone arrived. They were merely interrupted, repeatedly, by a new set of demands. Reading without checking something. Making something without immediately exhibiting it. Being unavailable without apology. Enjoying a private life that has no public counterpart. Taking an interest in something because it is beautiful, strange, funny or intellectually satisfying, and not because an algorithm has identified a profitable way to keep us occupied with it. All of these things remain available to us.
What makes the present moment interesting is that the cultural faith in the people running these systems has begun to show visible strain. The recent scandal involving Nvidia's Jensen Huang, one of the most influential executives in the artificial-intelligence industry, brought revelations of admitted personal misconduct, questions about corporate oversight and his resignation as chief executive in November 2026. The associated national-security allegations remain matters of investigation rather than established conclusions. An individual's conduct should not be used as proof that an entire technology is invalid. Still, the episode punctures the aura of inevitability and extraordinary judgment that has gathered around some of the industry's most powerful figures. A person can be brilliant at building a business and remain entirely fallible in the rest of life. [11]
A broader question of power is emerging as well. In a Pew survey conducted in June 2026 and published in November, 69 per cent of American adults said social-media companies had too much power and influence in politics. That is a direct expression of public unease about institutional authority. [12] The Democratic victories in the American midterm elections of 3 November 2026 also revealed an electorate willing to rebuke a governing political establishment. Cost of living, economic conditions and dissatisfaction with the administration were central to that result, so it would be dishonest to present the election as a referendum specifically on algorithms. What I take from it is a wider possibility: the narratives circulated by powerful institutions are not the only things people are capable of perceiving. Voters still encounter the world they live in, and sometimes that experience overrides the stories being sold to them. [13]
There are institutional signs of a rebalancing, too. Under the European Union's Digital Services Act, very large platforms must provide an option for recommender systems that does not depend on profiling. Australia has required age-restricted social platforms to take reasonable steps to prevent under-sixteens from holding accounts since December 2025. These policies raise difficult questions about effectiveness, enforcement, privacy and young people's rights. They are neither perfect nor sufficient. But they represent a departure from the old presumption that whatever a platform could normalise should simply become the new social arrangement. [14][15]
I suspect that this is the beginning of a rubber-band movement. Technology is not going away. Nor should it. I have no desire to return to the Middle Ages, to abandon the Internet, or even to reconstruct the 1990s in artificial form. There is too much knowledge, creativity and utility in the systems we have built. The error was allowing one particular commercial arrangement, a system of permanent exposure, engineered urgency and algorithmic supervision, to pass itself off as the natural destiny of human communication. We can keep the extraordinary capabilities while asking much more serious questions about what they are permitted to demand of us.
That rebalancing will not happen all at once. The major platforms retain enormous audiences, and their incentives are deeply embedded in business, entertainment, journalism and politics. The alternatives are imperfect. Some people will leave, others will limit their use, and others will continue to find valuable communities there. The change I am hoping for is quieter and more fundamental: a restoration of social permission to be absent. Permission to read without posting, watch without commenting, disagree without performing, have friends without maintaining a public friendship display, and let experiences end without turning them into content. Permission to disappoint an algorithm that has no legitimate claim on our attention in the first place.
I think we should become sensitive again to the strangeness of what we have accepted. We should be able to look at a comment section attached to a stranger's private life and ask why it exists. We should find it peculiar that an application expects to know where we are, what unsettles us, whom we envy and which emotional provocation will keep us looking. We should question why a platform has been allowed to convert our relationships into measurements, our presence into inventory and our hesitation into another opportunity for a notification. To be resensitised is to recover the capacity for ordinary human astonishment at arrangements that should never have become ordinary.
There is, I think, something almost comical about the threat that accompanies leaving. You will miss everything. You will lose touch. You will become irrelevant. You will cease to participate in the world. Yet the world remains outside, quite unconcerned with the dashboard reporting your engagement. Friends can still write. Someone can telephone. A letter can arrive. A meal can be enjoyed, a song heard, a child held, a thought pursued, a book finished. Whatever else the platforms have accomplished, they have been extraordinarily successful at convincing people that life takes place inside their account. It does not.
Once you experience that distinction directly, it is very difficult to unexperience it. I cannot return to seeing the demand for constant access as harmless simply because it is familiar. I cannot regard compulsory comparison, public performance and digital self-surveillance as the inevitable cost of being alive in the modern world. My life has become more fully my own by stepping back, and the evidence of that freedom is not a theory. It is the existence I continue to lead. We can remain connected without being continuously exposed. We can remain informed without being permanently interrupted. We can make things, love people, have opinions, change our minds and belong to the world without agreeing to be converted into a product. I have been picking back up where I left off. There is nothing obsolete about that life. I suspect more people are beginning to remember it.
Research notes
[1] Alice E. Marwick and danah boyd, “I Tweet Honestly, I Tweet Passionately: Twitter Users, Context Collapse, and the Imagined Audience,” New Media & Society (2011). Source
[2] Mustafa Oz and Scott Greeves, “Examining the Impact of Incivility and Intolerance on Participation in Social Media Discussions,” First Monday (2024). Source
[3] US Federal Trade Commission, “A Look Behind the Screens: Examining the Data Practices of Social Media and Video Streaming Services” (2024). Source
[4] Quote Investigator, “Quote Origin: Comparison Is the Thief of Joy” (2021, updated 2025). Source
[5] “Relationships Between Problematic Social Media Use and Social Comparison: A Meta-Analysis” (2026), PubMed Central. Source
[6] Jeffrey Lambert et al., “Taking a One-Week Break from Social Media Improves Well-Being, Depression, and Anxiety,” Cyberpsychology, Behavior, and Social Networking (2022). Source
[7] Hunt Allcott et al., “The Effect of Deactivating Facebook and Instagram on Users’ Emotional State,” American Economic Journal: Economic Policy (2026). Source
[8] Pew Research Center, “Social Media and Teens’ Mental Health: What Teens and Their Parents Say” (2025). Source
[9] Pew Research Center, “Americans’ Complicated Relationship With News” (2026). Source
[10] The Wall Street Journal, “The Year the Newsletter Business Reached a Fever Pitch” (2025). Source
[11] Reuters, “Nvidia CEO Jensen Huang Resigns in Blow to AI Boom He Helped Foster” (12 November 2026). Source
[12] Pew Research Center, “Majority of Americans Say Social Media Companies, News Outlets Have Too Much Power in Politics” (6 November 2026). Source
[13] Associated Press, “The Latest: Voters Give Democrats Control of Congress in Rebuke to Trump” (November 2026); Reuters, “Soaring Diesel Prices Helped Usher in Democratic Sweep” (5 November 2026). Source
[14] European Union, Digital Services Act, Regulation (EU) 2022/2065, Articles 27 and 38. Source
[15] Australian eSafety Commissioner, “Social Media Age Restrictions” (implementation from 10 December 2025). Source
