Chapter 16 separated data returns from the right to observe. Once data enters competition, the question is not only who is seen, but also why each person keeps increasing visible investment. A single participant who purchases courses, updates certificates, and extends hours online may be protecting opportunity under given rules; when everyone does the same simultaneously, the number of positions and the capacity for life do not necessarily grow. What is locally reasonable can accumulate into collective loss. The enterprise first screened positions with a basic certificate; once candidates had broadly obtained it, the platform added advanced badges, activity scores, and a "continuous growth index." Ning, afraid of falling behind, enrolled in night training; others made the same choice. Two years later, candidates' study time had doubled, the enterprise's screening speed had not durably improved, and care, rest, and non-certified practice had been squeezed out. No single party commanded this escalation.
This chapter does not treat competition, excellence, or educational investment as waste. Competition can discover capability, motivate learning, and allocate scarce opportunity; some added training does improve production and choice. What it critiques is the point at which rules convert absolute capability into relative position, visible proof into admission qualification, and systemic uncertainty onto each individual. Harsh exploitation may be equally present: platforms, employers, and intermediaries can profit from continuous escalation, and may even have reasons to keep the threshold moving; demonstrating such interests does not amount to asserting a unified scheme, nor does it supply a plan for accelerating the race.
Objects of Competition and Threshold Movement
A basic certificate that reliably tests safe operation allows everyone who attains it to be regarded as possessing minimum capability. Thresholds will be revised as technology and risk change, but the reasons come from the task. Such a certificate permits many people to qualify at once, and added learning can enlarge overall capability. The focus of audit is validity, accessibility, and updating, not opposition to standards. An absolute standard can also be set too high or be irrelevant to the position. Actual work, the consequences of error, and alternative proofs must be compared; a qualification cannot be exempted from scope checks because of its name.
When positions are fewer than qualified candidates, institutions must rank. Ranking provides limited selection, yet it means that one person's position depends on others' performance. If all improve together, absolute capability rises while the structure of ranks remains unchanged. If basic opportunity recognizes only rank, participants will keep investing to avoid relative decline. A relative scale is not necessarily unjust. Sports competitions, scholarships, and limited positions do have a ranking function. Rules should state whether the scarcity is real, how strongly the ranking is tied to the task, and whether those who do not attain the top ranks still retain basic capability and other entrances.
Ning has actual project experience, yet she ranks below those with more certificates because she lacks a daily activity record. Display can lower information costs, and it can also let activities that are easy to upload, count, and standardize displace capability that is hard to see. People then invest time in proof, not necessarily in the task itself. Organizations cannot directly observe all real capability and need proxy indicators. The questions are whether proxies can be corrected by other materials, how quickly they decay, and who bears the labor of display. Equating the invisible with the nonexistent systematically favors those with the most time to cultivate a record.
The continuous growth index expands from learning records into a vocabulary of personhood: those with high scores are called disciplined, those with low scores are said to lack motivation. Limited predictions acquire total qualification, and competition shifts from the task to who the person is. Failure then not only loses a position but damages self-explanation and relational reputation. Personality assessment sometimes carries real performance information, yet it must be bounded in object and time. One bout of low activity must not follow the subject into housing, health care, or an entire career; correction and fresh starts require an institutional place.
Threshold Movement Converts Incremental Capability into Positional Defense
When old certificates no longer distinguish candidates, the enterprise may need more specific task tests. If the updating corresponds to changes in the work, the added training has real value. One cannot infer manipulation merely because the threshold has risen. Institutions should publish task changes, transitions for old qualifications, and the incremental validity of new standards. If the position is unchanged and certificates keep being added only because most people qualify, the standard's function shifts from confirming capability to manufacturing scarcity. The difference lies not in the height of the threshold but in whether it answers a new question about the task.
When Ning enrolled in the course she need not have believed the content mattered; she only feared that everyone else would have it. Each person's worry is reasonable within relative ranking, while the collective outcome is increased expenditure with the positional structure unchanged. No participant can stop alone, because whoever exits first bears the loss first. This is a coordination problem, not weakness of will. Urging individuals to let go of comparison does not change the rules. Common capability thresholds, restricting irrelevant proofs, and randomized or rotating treatment of equal candidates can lower positional defense while preserving necessary competition.
After obtaining the advanced badge, Ning must stay active, or her earlier investment appears to depreciate. The platform launches a new tier, and those who have already invested find it harder still to leave. Sunk cost, identity, and qualification continuity act together, so the escalation extends itself. Past value need not be written down to zero, but it cannot automatically prove that the next investment is worthwhile. Qualifications valid over long periods, separating learning outcomes from platform activity, and allowing proofs to be exported let existing capability persist when one declines the new product. Providers have the right to update their services; they may not unilaterally rewrite the historical meaning of what has already been attained.
Learning notifications, leaderboards, and employer views arrive at night, and Ning's rest and care are converted into time spent preserving qualification. Formally the course is voluntary, yet key opportunity rewards continuous presence. Competition therefore allocates not only positions but also who may possess time exempt from evaluation. Some professions genuinely require continuous learning; the study should then be counted into work and institutional cost, with reasonable cycles set, instead of assigning the whole burden of renewal to the individual's available margin. Non-working time is not a public inventory awaiting development.
Collective Loss, Costs, and Beneficiaries
Each added certificate may slightly raise one person's own opportunity while relatively devaluing everyone else's existing certificates. Others follow, average investment rises, and positions and wages do not move in step. The product of competition may consist mainly of maintaining position rather than enlarging total choice. Judgment requires comparing capability, output, distribution, and investment, not merely counting enrollments. "Arms race" is a structural analogy: it does not mean the participants are hostile, nor does it equate education with weaponry. It points out that when gains depend on relative position, individuals cannot solve collective escalation through moderation alone.
The "involution perpetual motion machine" preserves the sting this loop requires: everyone keeps increasing investment while total opportunity does not grow proportionally, and whoever exits is punished first. But it is not a physical perpetual motion machine, and it cannot classify all effort, competition, or professional upgrading as involution. The diagnosis holds only when new investment mainly maintains relative position, thresholds keep moving because everyone follows, and coordinated exit lacks a realistic path; if task capability and collective gains keep growing, a more accurate description is needed.
Unable to observe long-term capability, firms rely on badges, check-ins, and public projects; job candidates rationally produce these signals. Signals are initially correlated with capability; later more people learn how to satisfy the form, and institutions add new signals. Both sides invest more, and information quality may not improve. This adaptation need not be cheating. The rules reward a certain visible behavior, and people act according to the rules. Governance should reduce task-irrelevant display, spot-check actual capability, and admit long-term results and diverse paths, rather than accusing participants of insincerity.
Uncertain about future position requirements, firms have platforms encourage everyone to purchase several skill sets in advance. When demand shifts, the cost of obsolescence falls on learners; the successful are displayed as having invested correctly, the unsuccessful are said to have planned poorly. Individuals do bear responsibility for planning, but institutions that command demand and returns should also bear proportionate informational and transitional obligations. Firms cannot be required to guarantee the future. They can be required to distinguish forecast, plan, and promise, to state the limits of their data, and to avoid packaging internal experiments as certain career channels. Public education, in turn, supplies a transferable foundation, reducing the bet on any single forecast.
Ning obtained a position through the advanced course. The success is real, and it does not prove that all the investment was effective. When the platform displays only those hired, the time, debt, and opportunity costs of the unhired vanish. Overall evaluation needs the denominator, alternative paths, and long-term outcomes. Conversely, many people failing to obtain positions does not prove the course worthless; they may have gained usable capability or relationships. Gains should be kept in separate columns; a single outcome, hired or not, cannot stand for the whole.
Which Costs the Rules of Competition Make Visible
Course fees are easy to count; night hours, equipment, substituted care, emotion, and switching costs do not enter easily. A free course may still demand continuous data and display. When individual choices and public programs are evaluated, the principal resources should be listed rather than a zero price used to prove that no extraction occurred. Not all time should be monetized. The purpose of recording is to compare allocation and sustainability, not to fold learning, relationships, and joy into a single net value.
Someone with savings can try several directions; someone carrying care duties or debt may lose housing and health margin with one wrong choice. The same rules of competition thus grant different numbers of trials. Looking only at final scores mistakes buffer for pure capability. Support can provide subsidy, time, and second chances according to need while keeping task standards intact. Differentiated help does not lower the demands on the subject; it prevents failure from becoming permanent exit.
While Ning attends night classes, family members shoulder care and community services fill the transport. Her certificate is recorded as individual investment, while its supporting conditions stay off the books. Competitors are not isolated individuals; rules distribute costs along relationships. Acknowledging support does not mean the family owns Ning's career decisions, nor does it license institutions to treat care as unlimited and free. Public childcare, study time, and predictable scheduling can formalize part of these costs. Whoever benefits from improved capability should participate in reproducing its conditions, rather than only the finally visible being rewarded.
Intensive training raises test scores, while sleep, pain, and recovery appear only afterward. A thought experiment cannot ground medical causation from this; real judgment requires reliable studies and individual assessment. Institutions can still record workload, entrances to rest, and exit, without waiting for confirmed harm before acknowledging risk. Health tools may likewise be used to demand that learners continuously prove their condition. Support should minimize data use and must not convert care into a new qualification surveillance.
Who May Profit from Continuous Escalation
Platforms earn revenue from courses, certification, and recruitment services, and the courses may genuinely help learning. The revenue relation does not prove that the platform deliberately maintains anxiety, but it will influence what the product prefers to display. Publishing completion rates, principal outcomes, advertising relations, and qualification validity periods lets participants judge the interests at stake. Demanding that platforms have no commercial motive is unrealistic. The questions are whether revenue depends on ever-shortening qualification lifespans, hidden common entrances, or exaggerated employment promises, and whether risk falls mainly on learners.
Firms demanding more certificates reduce their own testing and training costs. If the qualifications are valid, this may improve matching; if large amounts of irrelevant preparation are borne by candidates, firms gain convenience while social investment rises. The hiring side should demonstrate that high-cost conditions are task-related and consider on-the-job training and multiple forms of proof. Legal and market environments differ; specific duties need real verification. This chapter states only the cost relations; a unified regulatory conclusion cannot be derived from a thought experiment.
A certification body that both interprets job trends and sells new badges has a conflict of interest. It may provide professional value, and it may also have an incentive to describe ordinary renewal as urgent obsolescence. Recommendations, advertising, and qualification decisions should be separated, and forecasts should retain uncertainty and historical accuracy. A harsher scenario is several parties tacitly maintaining mutually recognized thresholds, leaving learners no other entrance. Judgment requires evidence of contracts, decisions, and competition; collusion cannot be asserted from rising prices or similar products. Governance focuses on interoperability, alternatives, and disclosure of interests; it does not teach how to close a market.
Those who have already obtained scarce qualifications want them to hold their value. This interest is understandable, yet it may block more effective alternative proofs. Professional standards require experience, and renewal should not be monopolized by incumbent members either. Independent publics, employers, learners, and affected parties participating together lower the risk of qualifications becoming a closed club. Lowering thresholds must not come at the price of safety. What is reviewed is task relevance, evidence, and alternatives; openness is not presumed to be always correct.
Data Acceleration and the Difficulty of Exit
Seeing herself drop two places daily, Ning adjusts her courses frequently; given capability feedback only at reasonable intervals, she might attend more to long-term learning. Real-time ranking has entertainment and informational value for certain contests, but when connected to basic opportunity it amplifies short-term reactions. The frequency of visibility should be proportionate to decision cycles and to the subjects' burden. Hiding all rankings may in turn render screening a black box. Institutions can publish rules and bands without continuously displaying everyone's relative position. The object of transparency is power, not a demand that subjects be infinitely visible to one another.
The system tells everyone that "similar candidates are increasing a certain skill"; a single recommendation may be accurate, but once all recipients change together the original prediction fails and the threshold rises again. The objects a model observes respond to the model. Platforms should explain how recommendations affect the aggregate and avoid disguising relative trends as guarantees. Learners cannot compute this feedback alone. Independent research and public statistics can compare investment, positions, and returns before and after recommendations, without exposing individuals' sensitive data.
When Ning moves to a new trade, her old activity level still lowers her recommendations. Competitive records flow across tasks, turning one exit into long-term punishment. Data should be limited in use and time, allowing subjects to demonstrate change with new material. Records of genuine fraud or safety incidents may be retained longer, with scope kept proportionate. Deleting history cannot cancel all third-party risk either. Correction, layered access, and time limits are more reliable than total forgetting or permanent memory.
The platform says the ranking is generated by data; the employer says it looks only at results; the certifier says it merely supplies input; each party can deny the decision. Whoever chooses the target, the features, the thresholds, and the consequences for opportunity bears the corresponding account. Complex supply chains need explicit entrances for suspension and appeal. A model may discover real differences without automatically deciding how those differences should distribute positions. Factual prediction and value choice remain the responsibility of the subjects who hold the authority.
Why Exiting the Competition Is Difficult
Even seeing the collective escalation, Ning cannot demand that everyone else stop simultaneously. Declining the new badge may mean being filtered out first. This marks the limit of moral persuasion: the problem requires rule coordination, basic thresholds, and alternative entrances. An individual may still choose to exit according to her own values, but she should not be required to prove the critique correct through sacrifice.
If ranking decided only a single prize, the cost of exit would be limited; if it simultaneously affects income, insurance, and housing, competition acquires a master switch over life as a whole. Different resources should avoid depending on the same revocable score, and basic security should not be conditional on continuous display. Scarce positions still must be allocated. Separating non-hiring from personhood and other rights already reduces the cross-domain consequences of a single failure.
Restricting irrelevant certificates, agreeing on study periods, or recognizing multiple qualifications may reduce waste, and may also be used by insiders to keep newcomers from competing. Rule-making must include newcomers, different paths, and independent assessment, with periodic checks on whether it is the task or the position that is being protected. The actual legal boundaries of anti-competitive and antitrust practice need concrete verification. Jointly reducing ineffective investment does not license price or opportunity to be manipulated by a closed body.
Care, play, community activity, and uncertified creation may not raise ranking, yet they sustain body, relationships, and meaning. If all time must prove vocational increment, these activities lose their place. Chapter 21 discusses the conditions of non-market life further; here we note only that they need not justify their existence through competitive value, nor are they thereby exempt from evaluation of their concrete consequences.
Excellence, Coordination, and Vested Position
A certificate confirms a safety threshold; position selection then uses limited, task-related material. Those not selected retain their qualification and need not repurchase every proof because of one competition. Layering reduces the swallowing of historical capability by ranking outcomes.
Institutions can cap application materials, recognize equivalent paths, and lengthen validity periods, lowering the display race. Participants may still pursue depth, creation, and excellence on their own; the point is only that irrelevant display does not remain the sole entrance. Enlarged choice does not mean that all differences disappear.
What platforms, employers, learners, families, and public institutions each gain and each bear should appear in project evaluation. If efficiency gains turn only into higher targets and shorter qualification lifespans, the common margin does not grow. The next chapter follows the distribution directly, once efficiency has improved.
Before a new indicator is added, its task increment must be demonstrated; when an old indicator fails, it is deleted; when investment rises while matching, capability, and income no longer improve, the institution triggers review. Stopping is not a promise never to change; it means that change must be justified anew, rather than continued automatically by competition.
Returning from Positional Defense to Capability Growth
Joint Coordination May Also Protect Vested Positions
Existing members may restrict new qualifications, raise industry entry, or block substitute technologies in the name of reducing ineffective competition. Jointly reducing waste is justified only when task standards, the rationale for quotas, and member mobility are verifiable. Newcomers, career changers, and atypical paths must participate in rule evaluation; those already benefited cannot define what counts as superfluous investment. Openness does not mean abolishing safety and professional training either. The audit requires that thresholds lead back to the task, the consequences of error, and incremental validity — not a presumption that the lower, the fairer.
Ning is willing to study in depth and draws meaning from competition; institutions should not forbid her for the sake of average burden. Restrictions should target qualification bundling, irrelevant display, and forced presence, not mandate identical investment from everyone. Subjects may pursue excellence, provided the surplus investment is not quietly converted by platforms into everyone's new minimum line. Distinguishing voluntary surplus from de facto threshold requires observing how employer demands, rankings, and peer expectations shift. Calling something "optional" in words does not answer.
A single general examination created to reduce the number of certificates may lower duplication, and may also render practical work, care experience, and local knowledge invisible. Equivalent proofs, task modules, and appeals keep coordination from compressing everything into a single personhood. Multiple paths cost more than one examination, yet remain more bearable than endless personalized surveillance.
Several platforms sharing the same data, certification, and recruitment entrances present apparent competition behind a common gate; a single public platform with open, portable, and accountable rules is not necessarily more closed. One should examine the control nodes, the switching costs, and the availability of substitutes rather than count brands. Real legal judgment further requires specific market evidence.
Long-Term Distribution and Non-Competitive Gains
Reducing irrelevant proofs may save time, yet firms convert the savings into more unpaid trial work, and the loss merely changes shape. Evaluation should track income, stability, training responsibility, and life margin, not treat fewer application materials as complete success. Rule changes should cover the next interface into which costs may migrate.
Winners indicate that the path may work; the unhired indicate the distribution of investment and the alternatives. Those who exited, never enrolled, or could not participate because of care duties also bear on the judgment of accessibility. Samples cannot be drawn only from a platform's existing users, or those unable to enter vanish once more.
Once adult competition is constrained, families may move certificate preparation earlier into children's education. The threshold does not disappear; it merely shifts forward in time. Education and recruitment rules need joint observation, protecting child development and family margin, rather than consign all future uncertainty to ever-earlier investment. This does not mean children should not learn or families cannot plan. The questions are whether basic opportunity requires ever-earlier proof of relative position, and whether failure allows a fresh start.
A certificate that once served only as positional defense may regain real capability value after a technological shift. Institutions should review by task, outcome, and cost, not turn the "involution" label into a talisman that exempts one from considering innovation. Critique itself must also admit counterevidence.
Making Non-Competitive Gains Visible
Ning has mastered tools, built relationships, or completed personal works; even without a position, these are real gains. Public evaluation should let these values exist, while not using them to mask failed employment promises. Different gains belong in separate columns, so that success narratives do not substitute for one another.
Many tasks depend on sharing, maintenance, and mentoring, while individual ranking may reward monopolizing visible results. Team evaluation, joint works, and peer accountability can supplement, and they also produce acquaintance bias requiring sourcing and appeal. Cooperation does not abolish individual contribution; it brings the conditions of joint production into view.
Once capability sufficient for the task is reached, a subject may choose not to continue displaying and still retain the qualification. The sufficiency line should update with risk, not merely because higher scores appear in the market. It leaves excellence still pursueable, while the common entrance for all is not turned into an endless race.
If society still has no more time for rest, care, innovation, and trial after capability improves, efficiency gains may be absorbed by new thresholds. Evaluating competition looks not only at the winners' output but also at whether the community retains space that is not immediately ranked. This question leads directly to the next chapter's account of where the gains go. The Linchuan Project separates basic capability certificates from position ranking, restricts irrelevant badges, and admits project results, practical tests, and different learning paths. Qualifications remain valid for reasonable periods; platforms may not rewrite attained capability through continued activity; firms publish the task basis of high-cost conditions and share the burden of necessary training. Real-time ranking becomes periodic feedback, and those who decline personalized competitive prompts may still use the common entrance. These changes will not make positions less scarce, nor give every investment a return. They make it easier for participants to judge whether they are building capability, contesting position, or preserving qualification, and they let the community see the moment when total investment rises while opportunity does not. Competition remains, yet failure need not become, across domains, a total failure of life and personhood.
Reform should also prevent losses from migrating into invisible private screening. When public certificates shrink, employers who select through connections, family background, or unappealable profiles lower the formal burden while raising arbitrariness. Task-related tests, records of decisions, privacy boundaries, and multiple forms of proof must be built in step. Removing one indicator does not abolish the responsibility of evaluation; it re-limits evaluation to objects that can be accounted for. Collective loss is the hardest to recognize because it is formed of many locally reasonable decisions. The absence of a unified manipulator does not mean the rules have no beneficiaries; platforms earning revenue, firms saving on screening, and winners maintaining thresholds may all keep the escalation going. Critique requires evidence of interests and decisions; it must not write the participants as fools, nor package structural outcomes as natural competition. The next chapter asks where the margin went once efficiency rose: when matching, automation, and production speed genuinely improve, why the time, risk, and returns thus saved do not necessarily return to those who supplied the data and the labor.