Chapter 12: Endless Feeds, Autoplay, and the Abolition of Stopping Cues

Overview

There is a moment that most people who have used the modern internet recognize: you look up from your phone and realize, with a small shock, that considerably more time has passed than you thought. You sat down to check something quickly. You checked it. And then, without deciding to continue, you continued — through a cascade of videos, posts, articles, and advertisements — until some external interruption (a notification from a different app, a knock at the door, hunger, the dimming of your screen) broke the spell and returned you to the awareness of where you were and how long you had been there.

This experience is not accidental, and it is not primarily a failure of individual self-regulation. It is the predictable result of a set of deliberate design choices made by engineers and product managers at the world's largest technology companies, choices aimed at removing the natural stopping points that would otherwise interrupt engagement. The technology of the endless feed — infinite scroll, autoplay, seamless content transition — represents one of the most consequential engineering interventions in the history of media, and its primary effect is the systematic elimination of what psychologists call stopping cues.

This chapter examines the psychology of stopping cues and their role in regulating behavior; the history and psychological rationale of infinite scroll and autoplay; how these designs exploit the Zeigarnik effect and circadian biology; the documented consequences of stopping-cue removal for time perception, sleep, and wellbeing; and the emerging design alternatives proposed by researchers, advocates, and — occasionally — the platforms themselves. The argument is not that technology is inherently addictive or that the internet is harmful. The argument is that specific, identifiable design choices have specific, identifiable effects on human behavior, and that understanding those effects is prerequisite to evaluating and potentially changing them.

Learning Objectives

• Define stopping cues and explain their role in regulating natural behavior across multiple domains
• Describe Aza Raskin's invention of infinite scroll and the psychological logic of his later regret
• Analyze how infinite scroll eliminates stopping cues from the feed experience and the behavioral consequences that follow
• Explain the Zeigarnik effect and how autoplay exploits it to sustain engagement across episodes
• Compare how Netflix and traditional television handle the stopping-cue problem differently
• Apply research on time distortion to explain why subjective time underestimation occurs during scrolling
• Assess the role of circadian disruption in the "3am scroll" phenomenon
• Evaluate specific design alternatives to stopping-cue removal, including their practical and business-model implications
• Critically assess the tension between Tristan Harris's human-technology critique and Marcus Webb's product development perspective

12.1 The Psychology of Stopping Cues

Human behavior is organized around natural endpoints. The chapter ends; we put down the book. The episode finishes; we turn off the television. The plate is empty; we stop eating. These endpoints — stopping cues — are signals embedded in the structure of activities that indicate completion and permit the behavioral transition to the next activity.

Stopping cues serve multiple psychological functions. They provide closure — the satisfying sense that a task or experience is complete. They create natural moments of transition at which the evaluation question "should I continue or stop?" can be raised. And they provide information about how much of an activity has occurred — the progress cue that makes it possible to regulate consumption.

The importance of stopping cues becomes apparent when you consider what happens in their absence. Research in behavioral economics on food consumption has documented a robust "unit bias" effect: people consume approximately one unit of whatever they are consuming, regardless of how large the unit is. Popcorn bucket size, plate size, and container size all influence how much people eat, not through any effect on hunger or satiety, but through their effect on the perceived completion point. Remove the visual cue that marks "one serving" and consumption expands to fill whatever container is provided.

Brian Wansink's food research (though some of his specific findings have been challenged in subsequent replications) illustrated this principle memorably with a "bottomless soup bowl" experiment: participants eating from bowls that were secretly refilled from below consumed approximately 73% more soup than those eating from normal bowls, while reporting equal levels of satiety. The absence of the visual stopping cue (the empty bowl) allowed consumption to continue well past the point it would otherwise have stopped.

The parallel to digital media consumption is direct and has been explicitly recognized by the designers who removed stopping cues from the digital environment.

12.1.1 Natural Stopping Points in Traditional Media

Traditional media was rich with stopping cues. A newspaper had a front page, sections, and a back page; you could see where you were within it at a glance. A book had chapters with defined lengths, visible through the page-fraction remaining in your right hand. A television episode ended with closing credits that announced completion. A scheduled television broadcast had a grid — the program listing — that structured time and provided natural transition points between programs.

Even radio, which is broadcast continuously and has no physical stopping cue, was structured into clearly demarcated segments (songs, news breaks, shows) with auditory signals marking transitions. The structural organization of traditional media created a scaffolding of natural stopping points that regulated consumption without requiring deliberate effort from the consumer.

12.1.2 What Happens Without Stopping Cues

Without stopping cues, the default human tendency is continuation. Kahneman's System 1 / System 2 framework (2011) is relevant here: deliberate evaluation of whether to continue an activity requires System 2 effort, while continuation of an ongoing activity is the System 1 default. Stopping cues create natural transition moments at which System 2 evaluation is prompted; without them, System 1 governs, and the default is to keep going.

This is not a failing of human cognition but a feature of it. In environments where stopping cues are provided by the natural structure of activities, System 1 continuation serves us well: we finish the book chapter, reach the natural pause, and evaluate. The trouble arises when the environment is engineered to eliminate those natural pauses — when every ending is immediately followed by a beginning, and no transition point is allowed to emerge.

12.2 Infinite Scroll: The Invention and the Inventor's Regret

In 2006, Aza Raskin was working at a startup called Humanized. He was thinking about a problem that many web designers were wrestling with: the friction of pagination. When websites divided content across multiple pages, users had to click through to the next page — a small interruption that broke the flow of browsing. Could the content simply continue scrolling, automatically loading new material as the user approached the bottom of the screen?

Raskin implemented infinite scroll in a matter of a few hours. The technical solution was elegant: a simple script that detected when the user was near the bottom of the visible content and triggered the loading of additional content. No click required. No page break. No stopping cue.

The feature was rapidly adopted across the web. Twitter implemented it. Facebook implemented it. LinkedIn, Pinterest, Instagram: essentially every major social media platform adopted infinite scroll as a standard design pattern, and it remains ubiquitous today. Raskin did not patent the idea; it spread freely and became a foundational element of the modern web experience.

Years later, reflecting on what he had created, Raskin offered a striking assessment: "It's as if they took behavioral cocaine and sprinkled it on the interface." He estimated that infinite scroll, across all the platforms that had adopted it, was responsible for approximately 200,000 hours of human attention consumed daily that would not have been consumed if the natural stopping cue of page loading had been preserved. He had, in the span of an afternoon, created a mechanism that collectively rerouted the attention of millions of people.

Raskin became a prominent advocate for design ethics and a co-founder of the Center for Humane Technology alongside Tristan Harris. His story represents a particular narrative in the history of technology: the inventor who comes to regret not what they intended but what they enabled; who recognizes, with the benefit of hindsight and research, that the elegant technical solution to a design problem was, at scale, a social and psychological intervention of unintended magnitude.

12.2.1 Why Infinite Scroll Works Against Users

The behavioral economics of infinite scroll is straightforward. Every time a user reaches the bottom of a content list and must actively choose to load more — to click "next page" — there is a moment of decision. The user must actively opt into continued consumption. Many users, at many decision moments, would choose not to continue. The friction of the click provides the space for this decision.

Infinite scroll eliminates this decision point. Content continues automatically. The user who scrolls to what would have been a page boundary simply continues, without any signal that a natural completion has been reached or any prompt to evaluate whether continuation is desired. The choice to continue is never made; continuation is simply what happens by default.

This is a design choice in service of engagement metrics. More time on platform means more content viewed, more advertising revenue, more data generated for algorithmic training. The user's experience of time well spent, by contrast, tends to be worse after prolonged infinite scroll sessions than after equivalent time spent with natural stopping cues. The design optimizes for engagement at the expense of the experience quality it generates.

12.2.2 Maya and the Scroll

Maya is doing homework on a Sunday afternoon when she opens TikTok to take what she intends as a five-minute break. The first video is funny — she watches it twice. The next is about a skincare routine she's been curious about. The next is someone doing a dance she vaguely recognizes. The next is a news clip about something happening at her school district. The next is a cooking video. The next—

Forty-five minutes later, she puts the phone down and stares at her homework. She is not sure what happened to the forty-five minutes. She was not trying to spend forty-five minutes on TikTok. She was not, at any point, consciously deciding to continue. She was simply... in the scroll. And then she wasn't.

The homework remains undone. But more than the lost time, what she notices is a kind of attentional grogginess — as if her focus, which had been continuously pulled and released, pulled and released, across dozens of disconnected pieces of content, cannot immediately reassemble itself around a single, sustained task.

12.3 Autoplay and the Zeigarnik Effect

If infinite scroll works by eliminating the stopping cue between discrete items in a feed, autoplay works by eliminating the stopping cue between discrete episodes of longer content. The psychological mechanism is somewhat different, and it exploits a different feature of human cognition: the Zeigarnik effect.

Bluma Zeigarnik was a Russian psychologist working in the 1920s who documented a striking phenomenon: people remember interrupted or uncompleted tasks better than completed ones. This effect — later named after her — reflects the mind's tendency to maintain a kind of cognitive bookmark on unfinished business, a low-level attentional pull toward the incomplete.

The Zeigarnik effect is adaptive in most contexts. Remembering that you left the stove on, that a task is unfinished, that a problem is unsolved: these are useful cognitive functions that prevent important things from being forgotten. But the Zeigarnik effect can also be exploited. Any narrative that ends on an unresolved note — the cliffhanger — exploits the Zeigarnik effect to generate the pull toward continuation: you need to see what happens next, not merely because you want to but because your cognitive system is tracking the incompletion and generating motivational pressure to resolve it.

12.3.1 The Cliffhanger as Technology

The television cliffhanger — the practice of ending episodes at moments of unresolved narrative tension — predates the streaming era by decades. It was understood by television writers as a retention tool: if viewers felt invested in what would happen next, they were more likely to return for the next episode, which in the weekly television schedule meant one week later. The Zeigarnik effect was being exploited for decades before autoplay existed.

What streaming autoplay adds to this existing exploitation is the elimination of the gap. In weekly television, the Zeigarnik effect had to work across a seven-day interval; viewers had to carry the cognitive bookmark for a week before the resolution was provided. In the streaming era, autoplay provides resolution — the next episode — within seconds. The cliffhanger effect is resolved before the viewer has consciously registered the ending.

Netflix's ten-to-fifteen second countdown before the next episode begins is the visual representation of this dynamic. It is long enough to show that a transition is occurring and short enough to ensure that most users will not make the deliberate choice to stop. Research on default-and-opt-out design across many domains consistently finds that most users accept defaults — the ten-second countdown frames continuation as the default and stopping as the deliberate choice requiring action.

12.3.2 Netflix and the Research It Did Not Publicize

Netflix has conducted extensive internal research on autoplay's effects on viewer behavior, and the broad findings have been reported in industry reporting if not in academic publications: autoplay significantly increases binge-watching rates relative to a non-autoplay condition. When viewers must actively choose to begin the next episode, many pause at episode boundaries to assess whether to continue, and a meaningful proportion stop. When autoplay provides the next episode automatically, almost all viewers continue at least through the opening of the next episode — and by that point, the Zeigarnik effect of the new episode's opening minutes has re-engaged them.

Netflix's internal decision to implement and maintain autoplay reflects its optimization objective: subscriber retention and content consumption per subscriber. By the metrics Netflix optimizes for, autoplay is successful — it increases consumption. Whether increased consumption aligns with the interests of users is a separate question that Netflix's internal research does not optimize for, and that Netflix does not publicly discuss.

The European Union's Digital Services Act has required platforms to provide alternatives to autoplay and has prompted Netflix to offer opt-out options for some markets. These regulatory interventions represent external pressure on a business decision that was, under purely internal logic, correct.

12.4 Time Distortion During Scrolling

One of the most consistently reported experiences of heavy social media use is time distortion: the subjective underestimation of time elapsed during engagement. Users who have scrolled for an hour report feeling that twenty minutes have passed. The subjective experience of time on social media runs at roughly half to two-thirds its actual rate, according to self-report research.

The cognitive mechanism of this distortion relates to attention and memory encoding. Time estimation depends significantly on the number of distinct experiences or events that can be retrieved from a period. Under conditions of high attention and active engagement with a structured, memorable sequence of events, time can feel expanded — the common experience of time "stretching" during intense physical or emotional experience.

Social media scrolling creates the opposite conditions. Content is highly varied — each item is different from the last — but the variation is shallow. Items do not build on each other narratively; there is no story, no arc, no structure that organizes the sequence into memorable episodes. The attention is continuously captured and released, captured and released, but the captured material does not encode into distinct episodic memories because it does not have the depth or structure that supports episodic encoding.

The result is that an hour of scrolling leaves fewer distinct memory traces than an hour of reading a book, having a conversation, or watching a structured film. With fewer memories to count, time estimation runs short. The hour feels like twenty minutes.

12.4.1 The Dissociative Quality of the Scroll

Users often describe the experience of prolonged scrolling in terms that suggest mild dissociation: "I wasn't really present," "I lost track of where I was," "I kind of zoned out." This language points to something real about the attentional state that scrolling produces. The continuous partial attention of the scroll — always engaged enough to keep scrolling, never engaged enough to be fully present — is a distinctive cognitive state that may be qualitatively different from other forms of media engagement.

Researchers have described this state using the concept of "flow" — Csikszentmihalyi's (1990) framework for the state of complete absorption in an activity. But the scroll is, at best, a degraded or counterfeit form of flow: it shares some phenomenological features (time distortion, loss of self-consciousness) while lacking others (the sense of purposeful progress, the accumulation of skill, the meaningful engagement with the activity's content). It produces the container of flow without its contents.

12.5 The 3am Scroll: Circadian Disruption and Endless Feeds

The abolition of stopping cues has particular consequences during the hours when the circadian system is signaling sleep. Traditional media provided natural stopping points that aligned reasonably well with bedtime: the late news ended, the late-night show finished, the program grid emptied out, and the television screen presented the expectation of sleep.

The endless feed has no such alignment with the circadian clock. There is no "end" to a social media feed at any hour. The content available at 2am is as rich and varied as the content available at 2pm. There is no emptying out, no signal from the platform that it is time to stop. The only signal is from the circadian system itself — sleepiness, the heaviness of eyes — and that signal is competing against an attentional engagement designed to override it.

Research by Charles Czeisler and colleagues at Harvard Medical School has documented in detail how light from electronic screens, particularly the blue-spectrum light emitted by most smartphone screens, suppresses melatonin production and extends circadian wakefulness. A person using their phone in bed at midnight is receiving both the behavioral pull of engaging content and the phototransduction signal that delays sleep onset.

The combination of these factors — no stopping cues from the platform, compelling content at any hour, sleep-suppressing light, and the behavioral momentum of the scroll — creates what has become a characteristic feature of contemporary adolescent and young adult life: the 3am scroll. Users who began checking their phones before bed, intending to spend ten or fifteen minutes, are still scrolling in the small hours, not because they decided to stay up but because no natural stopping point emerged to interrupt the continuation.

12.5.1 Consequences of the 3am Scroll

The consequences of chronic sleep disruption through late-night phone use are well-documented and serious. Adolescents require eight to ten hours of sleep for normal developmental functioning, and the National Sleep Foundation estimates that a substantial majority of American teenagers are chronically sleep-deprived. Sleep deprivation in adolescents is associated with impaired cognitive function, emotional dysregulation, increased risk-taking behavior, and compromised immune function.

Less obvious but equally important are the second-order effects on mental health. Emotional regulation — the capacity to modulate emotional responses appropriately — is significantly impaired by sleep deprivation. The chronically sleep-deprived teenager is not only cognitively impaired but emotionally dysregulated: more reactive, less able to manage distress, more vulnerable to the social comparisons and FOMO spirals described in Chapter 11. The endless feed's disruption of sleep thus amplifies its other psychological harms in a compounding cycle.

12.6 The Velocity Media Debate: Marcus Webb and Dr. Aisha Johnson

The question of whether to include autoplay in Velocity Media's video content player came to a head in a product review meeting attended by CEO Sarah Chen, Head of Product Marcus Webb, and Dr. Aisha Johnson, whose role as Head of Ethics was still finding its footing within the company's decision-making processes.

Webb's case for autoplay was standard product reasoning. Engagement data from beta testing showed that autoplay increased average session length by 34%. Return rates (users who opened the app again within 24 hours) were 12% higher for users who had experienced autoplay. "Every streaming platform has autoplay," Webb noted. "We would be at a competitive disadvantage without it."

Dr. Johnson raised the research. She had circulated a summary memo before the meeting that cited Zeigarnik's foundational work, the Netflix binge-watching research, and Gloria Mark's attention fragmentation studies. Her argument was not that autoplay was uniquely evil but that the company had made a public commitment to "time well spent" as a design principle, and that autoplay was structurally inconsistent with that commitment. "Time well spent means the user chose to spend the time," she said. "Autoplay is designed to make that choice for them."

Webb's response acknowledged the tension but reframed it: "Users who don't want autoplay can turn it off. We offer that setting. The majority of users, when given the choice, keep autoplay on. Is that not their preference?" Johnson's counter was direct: "When the default is autoplay and turning it off requires navigating a settings menu, keeping autoplay is not a preference — it's an absence of having made a deliberate choice. Default equals choice only if the default was chosen deliberately."

The exchange ended without resolution. Autoplay was launched with the standard default-on configuration and an opt-out buried in the settings. Internal data showed that fewer than 4% of users had changed the default.

This dialogue illustrates what philosophers of technology call the problem of embedded values. Every design choice encodes values — about what users want, about what outcomes matter, about the relationship between platform and user. The choice to make autoplay default-on encodes the value that engagement metrics outweigh the user's deliberate choice about how to spend their time. That value may be reasonable or it may not be; but it is a value, and claiming that design is neutral or merely responsive to user preference does not make it so.

12.7 Tristan Harris and the Center for Humane Technology

No single figure has been more influential in bringing the stopping-cue critique to public consciousness than Tristan Harris, a former design ethicist at Google who left to found the Center for Humane Technology in 2018.

Harris's core argument — articulated in his viral 2016 essay "How Technology Is Hijacking Your Mind," in multiple TED talks, and in the 2020 Netflix documentary The Social Dilemma — is that the attention economy has created a race to the bottom of the brain stem. Competing for human attention, platforms find themselves optimizing for the most primitive attentional engagement: novelty, variable reward, social validation, and FOMO. The result is that the interface of the modern smartphone is calibrated to the most evolutionarily basic human motivations, not to the full range of human values and intentions.

Harris's "time well spent" concept — the idea that the relevant metric for evaluating a digital experience is not how long it lasted but whether the user felt the time was well used — has influenced design conversations at major platforms and in regulatory settings. It reframes the question from "how do we maximize engagement?" to "how do we maximize satisfaction?" — and argues that the two objectives, while often aligned in the short term, systematically diverge over time.

The Center for Humane Technology has advocated specifically for the removal of autoplay and infinite scroll, for chronological feeds as an alternative to engagement-optimized algorithmic feeds, and for notifications redesigned to serve user needs rather than platform engagement objectives. Some of these recommendations have been partially adopted; all remain contested within the product development culture of major technology companies.

12.8 Design Alternatives to Stopping-Cue Removal

The critique of stopping-cue removal is most useful when paired with concrete alternatives. What would a social media feed or video platform look like if it were designed with stopping cues intact?

Time-Well-Spent Prompts: At regular intervals (fifteen minutes, thirty minutes), the platform could ask users a simple question: "You've been watching for thirty minutes. Are you enjoying this?" Research on similar interventions suggests that even simple awareness prompts can significantly reduce habitual overconsumption — not because users are forced to stop but because they are given the transition moment that stopping cues would naturally provide.

Natural Content Endpoints with Genuine Pause: Between episodes or after a defined viewing session, a genuine pause — not a ten-second countdown but an actual stop with a visible "start next episode" button — would restore the decision point that autoplay eliminates. This is how DVD players worked; the episode ended and play stopped until you chose to continue. Bringing this design back within the streaming context would represent a modest but meaningful reintroduction of stopping cues.

Usage Dashboards at Point of Use: Screen time dashboards (the kind implemented by Apple's Screen Time and Android's Digital Wellbeing) are typically accessed through settings menus, making them post-hoc feedback rather than point-of-use information. A dashboard visible within the app — showing time spent in the current session, cumulative daily use, and comparison to the user's own stated goal — would provide stopping-cue-adjacent information during rather than after use.

Chronological Feed with Defined End: A feed that displayed content chronologically and then actually ended — "You've seen everything posted since you last checked" — would restore the completion signal that the infinite feed has eliminated. This design existed in early Facebook and Twitter; its replacement with the infinite algorithmic feed was a choice that removed stopping cues in service of engagement optimization.

Sleep Mode Defaults: Platforms could automatically reduce functionality during late-night hours for users who have set sleep goals — not forcing cessation but making the default behavior quieter, with notifications suppressed and content loading slowed. This is technically straightforward and would align the platform's behavior with the circadian needs of a majority of its users.

12.9 The Economics of Stopping Cues

The design alternatives described above share a common feature: they would reduce engagement metrics. Users who are prompted to evaluate their consumption would sometimes stop consuming. Users who encounter genuine episode-end pauses would sometimes decline the next episode. Users who see a "you've seen everything" message would sometimes put down the phone.

This reduction in engagement is not hypothetical. It is the predicted, quantifiable business cost of reintroducing stopping cues. Platforms know this because they have tested alternatives; the engagement data on chronological vs. algorithmic feeds, on autoplay vs. manual play, on infinite vs. paginated scroll all exists internally and consistently shows that removing stopping cues increases engagement.

The economics of stopping cue removal thus represents one of the starkest examples of the structural conflict between platform business models and user wellbeing in contemporary social media. The case for removing stopping cues is straightforward and compelling in business terms. The case against it requires invoking considerations — user autonomy, sleep quality, time well spent, psychological health — that are real and important but that do not appear in the revenue and engagement metrics that govern platform decisions.

Regulation is one mechanism for introducing these considerations into the decision calculus. If autoplay and infinite scroll were subject to the same kind of regulatory scrutiny as, say, cigarette packaging or car safety features — if platforms faced legal liability for documented harms associated with stopping-cue removal — the business-case calculation would shift. Whether such regulation is feasible, appropriate, or politically achievable is a question beyond the scope of this chapter, but it is a question worth taking seriously.

12.10 Voices from the Field

"I invented the infinite scroll and now I can't stop using it. That tells you something. The tools we build should match our intentions. Infinite scroll doesn't match my intention when I sit down to browse. I want to check in, get what I need, and leave. Infinite scroll is designed to prevent that." — Aza Raskin, co-founder, Center for Humane Technology

"Netflix is competing with sleep. And we're winning." — Reed Hastings, CEO of Netflix, earnings call, 2017 (quoted widely in technology press)

"The ten-second countdown is the most manipulative element on any interface I'm aware of. It's genuinely brilliant, in a disturbing way. You've made stopping an active choice requiring effort. That's a profound design achievement in the service of an ethically questionable objective." — Tristan Harris, Center for Humane Technology

"We tell ourselves the user has control. They can turn off autoplay. They can set screen time limits. But when the default is maximum engagement and every alternative requires deliberate action, we are not giving users control. We are giving them the appearance of control while optimizing against their expressed preferences." — Dr. Aisha Johnson, Velocity Media Head of Ethics (internal memo, used with permission)

Summary

The endless feed represents one of the most consequential and least-examined interventions in the history of media design. By systematically eliminating the stopping cues that regulate natural behavior — through infinite scroll, autoplay, seamless content transition, and the absence of any signal that marks completion or prompts evaluation — platforms have created an attentional environment without natural endpoints.

Aza Raskin's invention of infinite scroll in 2006 removed pagination's brief stopping point; its adoption across all major platforms extended its reach to the entire social media ecosystem. YouTube and Netflix's autoplay features exploit the Zeigarnik effect to sustain engagement across episodic content by making continuation the default and stopping the deliberate choice requiring action.

The consequences include chronic time distortion (subjective underestimation of time spent), sleep disruption through late-night scroll behavior that overlaps with circadian rest periods, and the attentional grogginess documented by users who emerge from long scroll sessions unable to fully engage with other activities.

Design alternatives exist — time-well-spent prompts, genuine episode-end pauses, usage dashboards at point of use, chronological feeds with actual endpoints — but their adoption faces the headwind of engagement economics: stopping cues reduce consumption, and reduced consumption hurts revenue. Structural change in this domain is unlikely to come primarily from individual willpower or platform goodwill; the economic logic runs strongly in the other direction, and only regulatory or cultural pressure from outside the platform ecosystem is likely to shift it.

Discussion Questions

1. Aza Raskin describes himself as the inventor of infinite scroll who cannot stop using it. What does the creator's inability to resist the mechanism he created tell us about the relationship between design knowledge and behavioral control? Does knowing how something manipulates you protect you from its effects?

2. The chapter compares the stopping-cue problem in digital media to portion-size research in food consumption. How apt is this analogy? What aspects of the food environment parallel do and do not translate cleanly to the digital environment? What additional insights does the food analogy generate?

3. Netflix's Reed Hastings said the company competes with sleep — and, at the time, claimed they were winning. What are the ethical implications of a business explicitly identifying human biological needs as the competition to be defeated? How does this framing compare to how cigarette or alcohol companies have historically framed their relationship with human biology?

4. Dr. Johnson argues that "default equals choice only if the default was chosen deliberately." Is this argument persuasive? When is accepting a default a meaningful exercise of choice, and when is it not? What criteria would distinguish genuine preference expression from default acceptance?

5. The design alternatives described in Section 12.8 would all reduce engagement metrics. How should a public company weigh this business cost against the user wellbeing benefit? What mechanisms — governance structures, regulatory requirements, stakeholder pressure — might make this tradeoff more favorable to wellbeing?

6. The chapter mentions that the Zeigarnik effect was exploited by television cliffhangers decades before streaming autoplay existed. Does the existence of pre-digital FOMO mechanics and stopping-cue exploitation change the ethical evaluation of their digital equivalents? Why or why not?

7. If you were designing a social media platform from scratch with the explicit goal of maximizing time well spent (rather than time spent), what would the feed look like? How would you handle content continuation? What would you keep from current design, and what would you change?