The third ledger opens with a form of occupation that aggregate clock time cannot show: how many times attention is summoned, and whether each ending is truly an ending. Entertainment can connect a family and form rituals, or it can occupy shared time; these everyday consequences must be judged separately, within concrete relationships. For now this ledger follows only the chain of revenue: how a minute of staying is metered, predicted, and sold; who can change the feedback structure; and who bears the cost of what cannot end. Attention here is not a moral weakness but a resource that is organized, priced, and relieved of its risks by transfer.
Time is a necessary condition for attention but not its complete measure. To possess three hours does not mean that all three can be given to one kind of thinking; continuous completion, standing availability, frequent switching, and emotional vigilance leave different afterstates. Nor is concentration the only value. Care, collaboration, and emergencies rightly require interruption; exploratory work may gain new connections through switching; and mind-wandering in rest can also restore. Writing every dispersal down as damage would force life onto a single production line.
The fictional learning plan continues. On Tuesday Zhou Yao has three consecutive hours for her portfolio; on Thursday she again has three hours, but scattered across ten gaps: platform notifications, delivery calls, family inquiries, and ad-hoc work confirmations arrive at any moment. The minutes she finally invests are nearly the same; she completes less on Thursday, and in the evening she keeps revisiting unfinished items.
This contrast of tasks is set by the author and does not measure the size of real-world losses. Task switching carries time costs, as experimental research supports: the experiments of Rubinstein, Meyer, and Evans found that switching costs increase with rule complexity and can be reduced by task cues. This cannot be extrapolated into the claim that every interruption damages performance, still less into the whole causal chain linking waiting, anxiety, and platform revenue; on this basis the chapter proposes the conditions that need to be compared separately.
Minutes, Interruptions, and the Notification That Did Not Happen
"Attention drilling" can name arrangements that continuously extract attention, but this sharp metaphor must carry a strict scope of judgment: a platform is not mining an in-brain deposit of unlimited meterage, and analysis cannot infer manipulation from duration of use alone. The name describes only an observable relation — a system repeatedly manufactures opportunities to respond; its revenue comes mainly from prolongation and re-entry, while the costs of interruption, recovery, and the inability to end fall mainly on the user. If a service allows genuine ending, if its revenue does not depend on continuous recall, or if the user chooses long engagement under adequate conditions, then not all concentration can be counted as drilling.
The timing tool records that Zhou Yao had the design file open for one hundred eighty minutes, but it does not distinguish reading notifications, restoring versions, searching for materials, and actually editing. Total duration is real information, not the whole cause of task completion. Attention has objects, continuity, entry costs, and exit consequences. The same minutes deliver different capacities in different relationships. Zhou Yao is not required to work continuously, yet she must respond to shifts within half an hour. Formally idle, her action must remain reserved for potential notifications. She cannot begin an uninterruptible long task or complex undertaking. Chapter 7 already separated execution time from time control. This chapter asks further: during waiting, even with no task at hand, how do vigilance and uncertainty change available attention? Some people can read, do housework, or rest while waiting; the loss is not uniform. Personal habits, the character of the task, and the reliability of notifications all change the outcome. One therefore cannot price every on-call minute at the same rate, nor treat the minutes as fully free because some people can make use of them. Accounting must state the requirements of control and the realistic alternatives.
Zhou Yao receives a simple inquiry and answers it in two minutes. Returning to her portfolio, she must find the layer again, remember the reason for the change, and check whether it was saved. Notification duration and task-recovery duration are different quantities. Recovery cost is not a fixed constant per instance. Familiar tasks, clear bookmarks, and more natural interruption points may lower it; emotional conflict and complex reasoning may raise it. Zhou Yao's mind simultaneously holds the goal, the materials, the unfinished steps, and the possible errors. Interruption pushes part of these relations out of current reach; on return they must be rebuilt from records or memory. Tools can save state and cue the next step, reducing reconstruction. They may also generate more notifications and versions, making the technical solution a new source of the problem. Judgment must compare net outcomes. The interrupted person sometimes receives useful information and changes the original plan; not every deviation can be counted as loss. The relevance, timing, and decision rights over an interruption shape its value. An organization that logs only the two minutes of message handling hands the reconstruction to the individual. A more complete time record would at least mark task switches and recoveries, without claiming to measure the whole mind.
Zhou Yao waits for an ad-hoc work confirmation; the phone does not ring all hour. She still checks it several times, fearing that missing the message means losing the shift. The occupation comes from anticipation, not from the number of actual messages. This vigilance may be personal habit, or it may be formed by short windows, unclear rules, and the consequences of past missed calls. Intent cannot be read directly off checking behavior. If the notification were guaranteed to arrive at three in the afternoon, Zhou Yao could schedule continuous tasks before then; "could come at any moment" turns the whole period into a background of potential response. Longer response times, fixed windows, and confirmation mechanisms can restore schedulability. Emergency services and genuine care cannot always fix their times; they require commensurate compensation, rotation, and exit. Platforms may also exploit uncertain arrival to sustain continuous checking, so that users keep returning even when there is nothing new. Finding intentional design requires evidence from interfaces, rules, and behavior; it cannot be inferred automatically from heavy use. The critique stops at mechanism and responsibility: who controls arrival, who profits from returns, whether the subject can switch off, and how large the consequences of missing out are. It does not describe how to optimize uncertain stimuli.
Ten similar, predictable small tasks may be easier to recover from than a single sudden conflict. Counting only frequencies misses content, timing, and emotion. Switching among different simple tasks sometimes maintains vitality; switching among several complex goals may increase errors. Concrete conclusions require task materials. Zhou Yao's self-chosen break, turning to sketching, preserves a rhythm of ending and returning; an external notification demanding immediate response leaves less control. Same action, different position of the subject. Self-chosen engagement can also be shaped by platform inducement or habit, so analysis cannot end merely because the finger clicked voluntarily. Conversely, being prompted does not mean the subject has no choice. Switchable notifications, batching, and clear priorities increase control. Work or care roles may still require timely response; the conditions must be knowable at the time of commitment. To evaluate attention costs, one records the object of the switching, its predictability, the control rights, and the consequences, rather than searching for one universal number.
An email about debts takes three minutes to read, yet it keeps Zhou Yao thinking about next month's payment. The task has closed; the relation continues in attention. Life problems carry real importance, and worry cannot all be labeled inefficiency. Emotion may also signal a need for action, not noise to be cleared away. Unclear bills, appeals with no condition of ending, and eligibility that changes at any time keep the subject in permanent preparation. The institution counts only the processing window, not the uncertainty of the waiting. Solving the problem, stating deadlines, and providing status come closer to the cause than demanding that individuals manage their emotions. Personal support can still help, but it does not substitute for institutional responsibility. Some worries stem from misunderstanding, and clear information relieves them; some risks are real, and calm does not remove them. Attention analysis must not diagnose reasonable vigilance as a personal defect. Real mental-health judgments require professional evidence. This book only states that unfinished relations can occupy the conditions of planning; it claims no unified disease causation.
Feedback Revenue and the Value of Attention
The platform earns advertising or subscription value from views, clicks, and dwell time. Zhou Yao's attention thereby enters the business model, which is not in itself illegitimate. Content production and services need revenue. The questions are whether the revenue depends on continuous occupation, whether users can understand the exchange, and whether the losses are borne by private and public life. Long use may be deep learning, or it may result from confusing interfaces, repeated searching, or the absence of any way to exit. When operators treat dwell time as success, other outcomes must correct it. Short use is not necessarily better either. High-quality creative work, games, and relational exchange may require long investment. Duration shows exposure only; it does not determine value. One harsh inference is that a system can convert recovery, anxiety, sociability, and work all into entry points for return, so that every fluctuation of life generates new tradable attention. The critique should analyze revenue structure and exit, not treat the emotionally vulnerable as a target list. Commercial revenue can align with user value — useful content, for example, brings long-term subscription. The question is whether the metrics allow "task completed, no further use needed" to count as success.
Completion progress, peers' responses, and the occasional new opportunity make Zhou Yao willing to return. Feedback aids learning, and rewards can sustain difficult tasks. Uncertain feedback may induce repeated checking, but a single neurotransmitter word is not enough to explain all behavior. This chapter only proposes behavior-level relations to be tested: whether the way feedback appears changes return frequency, whether exit costs increase, and who collects the revenue. Neural mechanisms, effect sizes, and individual differences require the corresponding research; they cannot be derived from interface description. Zhou Yao enjoys the platform interaction; the pleasure is real value, and the platform's gain is also real. Both can hold at once. Critics who call all pleasure false decide the correct life on the subject's behalf; platforms that use pleasure to prove there is no difficulty of exit overstep just the same. Feedback design should make goals, frequency, and shutoff understandable; high-stakes services in particular should not mix eligibility notifications with entertainment stimuli in one unclosable feed. Real addiction and clinical diagnosis require professional standards. This chapter infers no illness from heavy use, repeated checking, or liking.
Some treat continuous reading and long plans as maturity, and chat, short-form content, and games as degradation. That turns modes of attention into ranks of subjects. Short forms can spread knowledge, sustain relationships, and provide recovery; complex lives also need rapid switching. Value depends on object, choice, and consequence. An organization that demands Zhou Yao work long stretches without interruption and forbids contact with those she cares for may also be appropriating attention. Continuity is not automatically legitimate because it is efficient. Autonomous attention includes the capacity to enter, sustain, switch, and exit. A person can choose immersion and still answer to important relationships, without being permanently occupied by one goal. Deep-work advice suits some tasks; it cannot replace labor boundaries and public design. Teaching individuals to practice concentration does not repair indefinite on-call availability and high-stakes notifications. A published critique should avoid the language of civilizational rank. The structure of attention is shaped jointly by technology, labor, family, and individual difference; no user of a given medium can be presumed inferior.
Zhou Yao gradually learns to handle simple tasks in the gaps and to save cues for complex work. Her performance improves, which shows that people can form new strategies. Adaptation is a real capacity and should not be described only as victimhood passively endured. The question is how gains and costs are distributed. Zhou Yao organizes lists in her own time; family members answer calls for her; the platform sees stable responsiveness. Private and relational investment makes system requirements look cheap. Adaptation can also enlarge future freedom. New strategies carry over into other parts of life, and not all learning can be counted as dispossession. The questions are whether the capacity is portable and whether it arises from an assumable choice. When an organization shortens response windows because its members have adapted, the recovered margin is extracted once more. The "recovers-then-raises" pattern discussed in Chapter 3 reappears in the domain of attention. Institutions should record the conditions that help stabilize performance, not only the final efficiency. The adapted still have the right to demand reasonable design; being able to bear a burden does not create an unlimited obligation to carry it.
After her fragmented day, Zhou Yao needs quiet, sleep, or low-demand activity. Modes of recovery differ by person; entertainment and social contact can also restore, not only doing nothing. Institutions cannot precisely own or measure each person's recovery. Quantifying all rest creates new performance pressure. Whether an activity restores can be observed in whether the subject is afterwards more able to choose, complete, and respond, not only in momentary pleasure or heart-rate data. Real measurement still requires professional methods. Recovery may also conflict with duty. Zhou Yao needs rest; family members need care; public arrangements cannot guarantee everyone an ideal environment at every moment. Sharing, rotation, and minimum boundaries are more workable than moral command. A platform's relaxing content may genuinely help, or it may keep Zhou Yao inside the same notification environment. An exit away from the entry point and the closing of high-stakes alerts are among the conditions of recovery. Chapter 12 takes up rest specifically. Here the chapter only confirms that attention costs extend past the end of the task and cannot be settled by the current minutes.
Costs Pushed Downward and the Boundaries of Observation
Facing courses, debts, work, and care all at once, Zhou Yao would need to read far more material to compare options. The more careful the choice required, the less attention remains for comparing. Complex menus formally expand options, while their cognitive and time costs may hand the advantage to the default path. This connects to the menu coercion of Chapter 4. Clear summaries, recommendations, and defaults reduce the burden and let people act. If providers hide alternatives or rank them by their own revenue, simplification decides on the subject's behalf. Full disclosure can also become an information flood. Layered explanation, key risks first, queryable details, and independent comparison are more accessible than demanding that all terms be read at once. A subject can delegate decisions to professionals or trusted relations; delegation is not the abandonment of final boundaries. The helper's interests and authority need to be visible. Insufficient attention is not automatic evidence that consent is invalid. One looks at the consequences of the decision, the information design, the time pressure, and the alternatives, avoiding the after-the-fact cancellation of commitments by appeal to any difficulty.
The platform demands instant confirmation; when Zhou Yao is busy, a family member reminds her; the work is fragmented, and the care schedule is rearranged by someone else. Individual response rates look good; the relational network carries the coordination. This connects to the transfer of care responsibility. System efficiency may rest on families maintaining shared calendars, memory, and emotional containment. Family members' willingness to help has real value and cannot all be converted into unpaid exploitation. If providers continuously rely on such help, it should be acknowledged in service design and cost. Those without support networks are more easily judged unreliable, and evaluation then writes relational difference into personal capacity. Appeals should allow an account of conditions rather than demand the disclosure of entire family lives. Public spaces, care services, and stable schedules can reduce private backfill. They carry budgetary costs, and an attention critique does not automatically generate unlimited rights from them. Accounting down to the final bearer makes the price of convenience more real, and prevents highly responsive people from being used as proof that everyone can adapt.
Wanting to understand interruption costs, an organization might require Zhou Yao to install detailed tracking that records every click and pause. Measurement itself adds attentional and privacy burdens. Not measuring at all, though, lets experience be dismissed as subjective. Voluntary logs, task outcomes, sampled interviews, and system events can serve as materials, each with its limits. If the aim is to adjust notification windows, reading the content of private messages is unnecessary; recording arrival times, response requirements, and task interruptions may suffice. Self-report conveys experience but cannot alone prove the whole causal chain; behavioral data show patterns but cannot read meaning. The two correct each other. Participation in measurement should not become a condition of eligibility; withdrawing from the study should not affect basic service. Otherwise subjects must hand over more attention data in order to receive help. Data retention and use need limits. A system built to reduce occupation must not then use the traces to evaluate personality and sharpen stimulation.
Zhou Yao sorts notifications into urgent, batchable, and for-reference-only; her course sets stable deadlines; the ad-hoc work provides response windows. Important events will still interrupt her, but she no longer has to stay on call for every possibility. Design cannot rest on individual settings alone. If a missed notification means lost income, an off switch carries no substantive freedom. Providers must adjust consequences and alternatives. A learning platform could count a user who completes the course while using it less as a success, rather than a decline in continuous hours. The conflict between service goals and business revenue needs open governance. Organizations preserve continuous blocks for complex work and schedule responsive windows for care and collaboration. Different tasks are not governed by one ideal of attention. Recovery time, switching, and private backfill enter the discussion of resources, without demanding that all mental life be precisely converted. Visible does not mean ownable. Zhou Yao retains trial, exit, and rollback. Once the structure of attention has changed, past consent does not automatically cover the new high-frequency demands.
Public Conditions and the Responsibility to Respond
Responding constantly to care and deliveries at home, Zhou Yao goes to the community reading room and gains two hours of stable work. The space did not strengthen her will; it changed the conditions of interruption and equipment. Libraries, schools, transit, and quiet places can all become attention infrastructure. They provide not only floor area but predictable rules, lighting, network, and the possibility of staying without consuming. Children's activities, exchange, and collective creation need sound; demanding that all public space serve deep focus excludes other uses. Times and zones can be divided, and limited resources acknowledged. Private venues may offer excellent environments; public entry reduces the dependence of capacity on purchase alone. Both require maintenance labor; quiet should not be imagined as naturally given. Spatial rules can also over-govern bodies and expression. Protecting attention should not turn into requiring everyone to hold the same posture; reasonable adjustment follows the task and the shared impact. Attentional capacity forms jointly from individual strategy and public conditions. Teaching self-discipline alone misses the places; building spaces alone cannot decide for the subject how to use them.
Zhou Yao wants the course not to notify her instantly; her collaborator needs to confirm that she has received an important change. Communication requires response; complete freedom from interruption is not a universal right. Relationships can agree on response windows, urgency categories, and backup contacts. A clear commitment makes non-response a predictable boundary rather than a sudden disappearance. If every message is labeled urgent, the categories lose meaning and the subject returns to continuous standby. Senders should bear the cost of misuse; receivers cannot indefinitely ignore genuinely urgent commitments. Different roles carry different obligations. On-call duty, care, and ordinary sociability cannot use the same response standard; pay, rotation, and exit follow the degree of control. A platform's do-not-disturb mode helps, but if high-stakes messages travel only through the same channel, switching off becomes unaffordable. Backup channels should be few in number and clear in purpose. The right not to respond is not a refusal of all relational responsibility; it lets the subject know when reachability is required and when attention can be returned to the life at hand.
The same notification arrangement can have different consequences for different people. Novices need more context reconstruction; the practiced recover faster; pain, care duties, or language differences can also change the conditions of availability. These differences cannot predict individuals directly from identity, nor be used to disqualify. Real adjustment requires the subject's own material and professional grounds. Offering text and voice options, clearer pacing, or more recovery time can widen participation. If an organization then sorts people permanently into low-capability categories, support turns into a restriction of opportunity. Children's modes of attention in particular cannot be evaluated solely by adult standards of production. Education requires the practice of sustained engagement, and also activity, relationship, and developmental difference, with concrete boundaries left to professional research. Bodily information is sensitive material. Improving notifications need not collect a complete health profile; adjusting to a specific need is narrower than a total label. Recognizing difference deprives the uniform minute of its monopoly, without implying that any subjective feeling can dictate shared arrangements. Consequences for all parties still need negotiation.
The platform has repeatedly marked ordinary recommendations as important; Zhou Yao gradually ignores all alerts, and one critical deadline is missed. The problem of attention is not only quantity but the relation between signals. Accurate classification lets the subject check less; mislabeling shifts the cost of judgment onto the receiver. Successful sending does not mean effective arrival. When the platform explains why it labels, allows adjustment, and corrects after error, users gradually form expectations. If every failure is attributed to personal settings, the relationship cannot learn. Users also misjudge priorities; senders cannot be required to guarantee that everyone understands. Important matters can have confirmation and reasonable repetition, not replacement by unlimited bombardment. An organization that profits from open rates may tend to expand the important label. Whether this happens requires data and documentary evidence; structurally, service-necessary notifications and promotional metrics should be separated. When the signal is trustworthy, the notification that did not happen occupies less vigilance. Reliability is itself an attention resource, though it is hard to state directly in minutes.
Compensation, Ending, and Performance Evaluation
Because frequent ad-hoc notifications kept Zhou Yao from finishing her course, should payment be made per interruption? The question is not easily answered. Attention has no uniform market price, and outcomes depend on the individual's tasks. But the impossibility of precise pricing does not mean the costs do not exist. Labor agreements can compensate standby, night response, and cancelled arrangements, or provide recovery time, in forms that vary with the task. Higher pay can make a hard arrangement an assumable choice, yet it does not automatically authorize intrusion on health, safety, and basic rest. Real floors require law and professional norms. Conversely, once a subject accepts explicit pay and response duties, she bears the commensurate performance. One cannot declare all attention costs void after the fact when the outcome is unfavorable. The platform user's exchange is more complex: free content, data, and attention are interconnected. Clear exit, low-interference options, and revenue transparency may be more suitable than payment by the minute. Public-service notifications, though not profit-driven, should still reduce irrelevant interruption. Different motives do not erase the consequences of design. Compensation must also ask whether costs shift onto the family. Paying only Zhou Yao's response time leaves the ledger incomplete if a caregiver absorbs the rearrangement; this does not mean that all relational labor must be commodified.
The aim of attentional boundaries is to restore choice and sustainable commitment, not to turn every human gaze into a sellable asset. What can be priced enters the contract; what cannot be converted can still ground a limit.
Zhou Yao promises to respond promptly to the course discussion this month, committing to help the team finish the project. After the project ends, the group chat and reminders continue; temporary reachability becomes default long-term reachability. Shared tasks require attentional commitment; quitting too early passes costs to peers. A commitment should include its object, its duration, its handover, and its emergency exceptions. That Zhou Yao once responded at high frequency does not mean she permanently prefers that rhythm. New requirements from the organization need fresh statement; historical behavior cannot substitute for consent. Nor can she suddenly close all contact and refuse handover right before a key delivery. Attentional autonomy and the duty of performance hold together in time. Ending notifications, archiving the group, and naming the new point of contact free the mind from keeping vigil over a relation already completed. Technical closure should follow the completion of real responsibility. Endable commitments make periods of high investment a choice rather than an expanding norm. Attention can then return to other parts of life and to the next shared plan.
The next chapter takes up measured performance and invisible labor. Output in fragmented environments is lower, and evaluation that looks only at results deletes attention conditions, maintenance, and coordination from the causal account. Interruption, vigilance, context reconstruction, unfinished emotion, and recovery time mark the different positions from which attention costs are compared. If an arrangement turns all these fluctuations into entry points for return and then uses dwell time to prove value, it may overlook the switching costs the subject pays. Whether such a transfer occurs requires usage records and the parties' material; "voluntary clicking" by itself can neither prove it nor rule it out. Nor does it prove that all short content, entertainment, notifications, or heavy use are harvesting. Feedback can support learning, switching can serve care, adaptation can create capacity, and long concentration can also be dominated. Attention cannot be priced by the minute alone, because the minutes do not record who controls the object, when exit is possible, how one recovers after the task, and who, in the end, bears the cost.