SWITCHED ON
The daily technology series nobody asked for but everyone needed
Stolen Focus: Digital Addiction and the Attention Economy
The most valuable resource in the twenty-first century is not oil, not data, and not compute. It is human attention. An entire industry has been built to capture it, hold it, and sell it. The cost is being paid by the people whose attention was taken.
There is a consensus emerging that we have built an attention economy that is corroding the capacity for the kind of sustained, focused thought that produces the most valuable human cognitive work — and that the people building it knew this was happening and continued anyway, because engagement was the metric and engagement was the business.
Yesterday we stepped out of the smart home and into the smart city — Sidewalk Toronto's collapse as a lesson in data governance, London's surveillance camera density and the live facial recognition debate, Barcelona's data sovereignty model, and the argument that a city is a political community not a platform, and that conflating the two produces governance failures that serve technology companies rather than residents. Today we are going somewhere deeply personal and considerably less institutional. This episode is about what the technology ecosystem we have examined across thirty-seven episodes of this series is doing to the individual human capacity to think. Not to feel, not to connect, not to vote or work or age — those have all been covered. To think. Specifically, to think at length, in depth, without interruption, on one thing at a time. The attention economy, digital addiction, and the case that the design choices embedded in the products billions of people use every day represent one of the largest-scale and least-discussed cognitive interventions in human history.
01 — The Design Isn't Accidental
The features of digital products most associated with compulsive use — infinite scroll, variable reward notifications, autoplay, like counts, streaks, pull-to-refresh — were not arrived at accidentally through neutral design processes. They were developed deliberately, tested iteratively against engagement metrics, refined based on what increased time-on-platform, and implemented because they worked. The people who built them understood what they were doing. Several of them have since said so explicitly.
Aza Raskin, the designer who invented the infinite scroll — the mechanism by which content feeds continue indefinitely as you swipe downward, eliminating the natural pause that a page bottom would create — has publicly estimated that infinite scroll costs humanity approximately 200,000 hours of collective attention per day. He invented it to solve a pagination problem and describes the experience of watching it spread across the internet as one of regret. Tristan Harris, a former design ethicist at Google, has spent years documenting the persuasive technology techniques deployed by major platforms — drawing explicitly on the psychological literature on variable reward schedules, social validation loops, and reciprocity triggers that were developed in the context of slot machine design and addiction research, and applying them to social media feature design.
The slot machine comparison is not a metaphor. The variable reward schedule — the unpredictable intermittent reinforcement that makes gambling more compulsive than predictable reward — was a known psychological finding before social media existed. The architects of the engagement-optimised feed applied it deliberately. The comparison is a design specification, not an analogy.
The business model that drives this is the attention economy: platforms that are free to users generate revenue by selling advertising, which means their commercial interest is in maximising the time users spend on the platform and the frequency with which they return. Every feature that increases engagement — even engagement that users themselves describe as compulsive, regretted, or contrary to their stated preferences — serves the business model. There is no feature team at a major social platform whose job is to ensure users spend the appropriate amount of time on the platform and no more.
02 — What the Interruption Research Shows
The cognitive science literature on interruption and attention recovery provides the empirical foundation for concerns about digital distraction that are often expressed in anecdotal terms. The findings are more consistent and more alarming than the common cultural framing of distraction as a personal discipline problem suggests.
Research by Gloria Mark at the University of California Irvine, tracking knowledge workers in their actual work environments over many years, found that the average time before workers were interrupted or self-interrupted had fallen from approximately two and a half minutes in 2004 to approximately forty-seven seconds by 2020. The time required to fully resume deep focus after an interruption — not just to return to the task but to restore the mental context required for complex cognitive work — averages approximately twenty-three minutes. The arithmetic of these two figures, applied to a working day punctuated by frequent notifications, is not encouraging for the quality or quantity of deep cognitive work being produced.
The research on multitasking — the common belief that frequent task-switchers become better at switching and can effectively process multiple streams simultaneously — consistently finds the opposite. Heavy media multitaskers perform worse on tasks requiring sustained attention, are more easily distracted by irrelevant stimuli, and have more difficulty filtering out information they are trying to ignore than light multitaskers. The cognitive habit of frequent interruption and task-switching does not train the brain to handle it better. It degrades the capacity for sustained focus that it repeatedly interrupts.
03 — The Addiction Framework and Its Complications
The use of addiction language to describe compulsive digital behaviour is contested and the contestation is legitimate, not merely defensive. Clinical addiction involves tolerance, withdrawal, significant life disruption, and continued use despite clear harm — criteria that are met by some heavy digital users and not by others who describe their use as compulsive but manageable. Applying addiction language too broadly risks both trivialising clinical addiction and pathologising ordinary consumer behaviour in ways that serve neither understanding nor those who are genuinely struggling.
What the more careful framing offers is the concept of problematic use — digital behaviour that is experienced as inconsistent with the user's own stated preferences and values, that the user repeatedly attempts and fails to control, and that interferes with relationships, work, or wellbeing. Rates of problematic smartphone use in population surveys vary significantly by methodology and definition, but consistently find that a meaningful minority of users — typically ten to twenty percent in well-designed studies — report symptoms that meet the threshold of problematic use by the most conservative definitions. For adolescents, rates are typically higher. The evidence that these symptoms are associated with poorer wellbeing outcomes, independent of other factors, is reasonably consistent even as the causal direction remains debated.
The policy implications of the addiction framing matter. If problematic digital use is understood primarily as an individual self-regulation problem, the response is individual — screen time limits, digital detoxes, app blockers, personal discipline. If it is understood as a product design problem — the predictable consequence of applying deliberate compulsion-engineering techniques to products used by billions of people — the response is systemic: design standards, advertising model reform, regulatory requirements, and producer liability for harms. The evidence supports both framings simultaneously, but the systemic framing has received considerably less policy attention than its causal significance warrants.
04 — What Sustained Attention Is Actually For
The argument for protecting sustained attention is not primarily about productivity in the narrow economic sense, though the economic case is real. It is about the kinds of thinking that are only possible under conditions of sustained, uninterrupted engagement with a single subject, and what is lost when those conditions become difficult to achieve.
The psychologist Mihaly Csikszentmihalyi documented the state he called flow — the experience of complete absorption in a challenging task, characterised by effortlessness, loss of self-consciousness, and deep satisfaction — as one of the primary sources of human meaning and fulfilment. Flow requires uninterrupted focus for an extended period. It is difficult to achieve in an environment structured around frequent interruption and the habitual checking of notifications. The erosion of the conditions for flow is an erosion of one of the primary mechanisms through which humans derive meaning from their activities, not merely a productivity loss.
The capacity for sustained reading — the kind of deep engagement with complex text that builds knowledge structures across extended periods of concentration — is documented to be declining in populations with high digital media use. Maryanne Wolf, a cognitive neuroscientist at UCLA who studies reading, has described a pattern she calls the "bi-literate brain" — the ability to read deeply for extended periods, which develops through practice and can be lost through disuse, in tension with the skimming, scanning, hyperlink-following reading patterns that digital environments reward. The concern is not merely that people read less but that the cognitive circuits that enable deep reading are being used differently, and that the capacity for a specific kind of sustained intellectual engagement is declining as a population-level skill.
05 — What Can Actually Be Done
The individual interventions are real and the research on them is more positive than the structural critique might imply. Turning off non-essential notifications reduces interruption frequency significantly and is one of the most consistently recommended and evidence-supported behavioural changes for improving focus. Establishing phone-free periods — particularly the first hour of the morning and the last hour before sleep — has documented effects on both sleep quality and cognitive performance during the day. Deliberately scheduling distraction — specific periods of social media use rather than continuous ambient availability — exploits the same variable reward mechanism by making the timing more predictable, reducing its compulsive pull. These are not solutions to the structural problem. They are genuine individual mitigations.
The structural interventions that would matter more: design standards requiring opt-in for the most compulsion-engineering features, advertising model regulation that reduces the commercial incentive for maximising engagement at any cost, age-appropriate design requirements for products used by minors, and producer liability frameworks that create legal accountability for harms from deliberately compulsive design. Several of these are in various stages of regulatory development in the UK, EU, and US. None are yet implemented at the scale and with the enforcement that the evidence of harm would justify.
The deeper reframing that the evidence calls for is treating attention as a public health issue rather than a personal productivity issue. The systematic engineering of compulsive behaviour in products used by most of the adult population and virtually all of the adolescent population is a population-level intervention in human cognitive capacity. It deserves a population-level response — the same kind of regulatory, educational, and design-standard response that we apply to other consumer products whose design choices affect public health. The comparison to tobacco is imperfect and often overstated. The comparison to any other industry that systematically deployed known psychological manipulation techniques against its users and faced no regulatory consequence for doing so is harder to make, because there isn't one.
Tomorrow we are moving from the psychology of individual attention to the sociology of online identity — specifically the phenomenon of cancel culture, accountability, and what the internet has done to the question of whether people and institutions can change, be forgiven, and move on. A genuinely difficult topic that deserves more nuance than it typically receives from either direction. See you then.
Switched On is a daily technology series covering the ideas, systems, and arguments shaping the digital world. Opinionated. Witty. Occasionally wrong. Always worth the argument.
