How the Necessary Pursuit of Capacity-Building Can Undermine Its Own Goals
The Problem of Existence
To exist is to confront problems. An organism must maintain itself against disturbances. It must obtain resources, respond to changes in its environment, and recover from damage. Failure to do these things eventually ends its existence.
As human beings, we are aware that the problem is more expansive. We inhabit a world that we can never fully anticipate. We face known problems, unknown problems, and problems we do not yet possess the concepts to recognize. Our continued existence therefore depends on more than maintaining the conditions that sustain us today. We must remain capable of discovering and creating the conditions that will sustain us tomorrow.
Existence is an open-ended undertaking. It has no predetermined terminal state at which every relevant problem has been solved. Every solution may introduce new possibilities, new dependencies, and new problems. This gives those of us interested in existence an overarching purpose: to sustain the possibility of continued existence indefinitely. That purpose does not prescribe a particular way of living. It does not require everyone to pursue the same interests, adopt the same values, or participate in the same institutions. Quite the opposite. The open-ended character of existence is what makes a plurality of purposes possible. Every viable purpose can live within the larger container of existence. But the container must be sustained for any of those purposes to remain possible.
The Human Club
Human beings are continuous with a long evolutionary lineage. We are animals, and something much more. We have an extraordinary cognitive capacity. As David Deutsch argues in The Beginning of Infinity, we are universal explainers.
A universal explainer is an intelligence capable, in principle, of understanding anything that is comprehensible. This does not imply omniscience, infallibility, or unlimited resources. It means that explanatory capacity isn’t intrinsically restricted to a predetermined set of subjects or problems. Once this threshold of universal cognition is crossed, knowledge can be created about domains that were previously inaccessible or inconceivable. We can discover errors in our understanding, formulate better explanations, invent new tools, and transform the circumstances of our existence.
This capacity is what defines membership in the human club in the relevant cognitive sense. Biological substrate is beside the point. Any being that genuinely possesses universal explanatory capacity belongs to the same epistemic category.
Members of the club are fallible. We do not receive authoritative explanations from outside the process of knowledge creation. We conjecture, criticize, test, revise, and discover. There is no epistemic authority above the club of humanity that can relieve its members of responsibility for understanding reality. We can create tools that extend our abilities. We can cooperate with other intelligences. But no technology becomes an independent source of unquestionable authority merely because it is powerful. This matters because the continued existence of the club depends on its ability to solve problems it cannot yet anticipate.
Value as Capacity Building
Continued existence is a necessary condition for the pursuit of other purposes. But existence does not simply sustain itself. It requires the continual development of capacities to confront problems, including those we cannot yet anticipate. Capacity building is therefore a fundamental form of value.
Capacity is the ability to do something that could not otherwise be done, or to do it more reliably, more effectively, or under a wider range of conditions. Capacity building expands what becomes possible. Knowledge is capacity. Tools can embody capacity. Skills, institutions, infrastructure, relationships, and arrangements for cooperation can all contribute to capacity.
But capacity is more than a stock of possessions or capabilities. Some of the most important capacities are regenerative: the ability to repair, reproduce, adapt, replace, or create other capacities. This matters enormously. An institution might increase production while exhausting the resources that make production possible. It might improve immediate performance while degrading the knowledge or independence required to respond to future disturbances. Such an institution may be increasing one capacity while destroying others.
Therefore, value cannot be reduced to maximizing any single measure of capability. The relevant question is whether capacity building contributes to sustaining and expanding the conditions for indefinite, open-ended existence. Capacity also has to remain usable, recoverable, and adaptable. A capability that works brilliantly under one narrow set of conditions but fails catastrophically outside them may contribute less to long-term viability than its apparent performance suggests.
We need capacity. We need to keep building it. And this necessity creates the conditions for a profound systemic pathology.
Why We Scale
Human beings discovered that cooperation, specialization, and coordination allow us to build capacities that individuals cannot build alone. A person can make a tool. A community can sustain specialized toolmakers. A civilization can build laboratories, universities, transportation networks, hospitals, manufacturing systems, and scientific institutions. Scale multiplies our capabilities.
But scale introduces a coordination problem. As an organization grows, its participants cannot individually understand every activity, every dependency, or every consequence of every decision. Direct knowledge becomes increasingly expensive to maintain.
To coordinate at scale, we develop representations that stand in for more complicated realities. We use prices to represent economic information, grades to represent learning, credentials to represent competence, performance indicators to represent organizational effectiveness, and articulated mandates to represent institutional purposes. These representations are proxies.
Proxies make coordination possible by compressing complexity into forms that can be communicated, compared, standardized, and acted upon. Without them, much of modern civilization would be impossible. But a proxy is never the full reality it represents. And that is where the problem begins.
Proxies, Blind Spots, and Friction
Every proxy is selective. It preserves some features of reality while omitting others. A useful proxy preserves enough of the relevant structure to support effective understanding and action. An impoverished proxy discards relationships that turn out to matter. The omitted features are its blind spots. Blind spots are not necessarily accidental defects. They are often what makes a proxy useful in the first place.
A representation that includes fewer complications can be easier to calculate, transmit, enforce, and optimize. Omission reduces friction. Reduced friction facilitates scale. This creates a selection pressure that is easy to overlook. Proxies are selected not only for what they successfully represent, but also for what they allow a system to ignore.
A school can more easily compare examination scores than the full development of explanatory capacity in its students. A hospital can more easily report throughput than the total consequences of care. An organization can more easily evaluate compliance with a mandate than independently determine whether that mandate contributes to real capacity building.
The discarded complexity does not disappear from reality. It disappears from the representation governing the system. This is critical. The fact that something is difficult to represent does not make it unimportant. Yet the very features that resist easy representation can be systematically disadvantaged during scaling. Over time, a civilization can preferentially adopt representations that facilitate coordination while discarding information necessary to understand the consequences of that coordination.
The Scale Attractor
We can now identify the Scale Attractor.
The Scale Attractor is the self-reinforcing tendency for capacity-building systems to favor increasingly scalable proxies, including proxies whose blind spots undermine the capacities those systems originally sought to build.
One causal sequence through which this tendency can emerge is:
- Existence requires capacity
- Capacity building benefits from scale
- Scale requires proxies
- Proxies reduce friction through selective omission
- Low-friction proxies gain influence
- Proxies become optimization targets
- Institutional drift follows
The crucial feature is the selection mechanism. An organization that adopts a simpler, more easily optimized representation may coordinate more efficiently than one that preserves additional complexity. Greater efficiency can enable faster growth, wider influence, and greater resource accumulation. As the organization scales, the proxy becomes more deeply embedded in its operating procedures. More decisions depend on it. More participants organize their behavior around it.
Eventually, the system is no longer merely using the proxy to pursue its original purpose. It is reorganizing itself to satisfy the proxy. This is the familiar problem associated with Goodhart’s law, in which a measure becomes a target. But the Scale Attractor explains why the problem repeatedly arises even in institutions staffed by intelligent, sincere, competent people.
The drive toward proxy substitution is generated by a process that civilization genuinely needs. Scale helps us build capacity. The resulting coordination pressures favor simplified representations. Those representations then shape which capacities are recognized, rewarded, and reproduced. The pathology is endogenous to capacity building.
No bad actor is required for this pathology to emerge. The institution can remain perfectly accountable to its chosen measures while becoming progressively detached from the reality those measures were supposed to represent. Indeed, more effective enforcement can accelerate the drift. But the resulting structures also create an emergent exaptation surface: features selected for efficient coordination can subsequently be repurposed by actors who recognize opportunities within the system’s blind spots. The implications of that possibility extend beyond the endogenous mechanism considered here.
Institutional Drift and the Meaning Crisis
Institutional drift occurs when an organization increasingly serves its representations of purpose rather than the reality that originally justified those representations. Its procedures may remain coherent. Its participants may remain sincere. Its performance indicators may improve. Yet its actual contribution to capacity building can deteriorate.
Growth itself can become a load-bearing institutional assumption. In Science Since Babylon, Derek de Solla Price documented the exponential expansion of scientific activity and recognized the limits of extrapolating such growth indefinitely. Eric Weinstein’s concept of Embedded Growth Obligations identifies what happens when institutions build the continuation of a growth regime into their commitments. Positions that were honestly supportable when established can become unsupportable without anyone changing those commitments. The resulting obligations can make institutions increasingly dependent on reproducing their existing structures, even when continued viability requires regenerating the capacities that originally sustained them.
This is where meaning becomes essential. The word meaning is used in several legitimate senses. It can refer to linguistic signification, personal significance, existential purpose, or explanatory structure. These senses should be distinguished rather than silently substituted for one another. In the structural sense developed here, meaning is what enables representations to preserve, expose, and recover their consequential relationships to reality.
Meaning helps us recognize what a proxy captures, what it omits, and why those omissions matter. An unimpoverished proxy should not be confused with an exhaustive description of reality. No such description is available. It is a representation sufficiently connected to the relevant structure that intelligent agents can recognize consequential omissions and improve their understanding. This connection is not guaranteed by formal precision. A proxy can be mathematically precise and still be catastrophically impoverished.
The experience people describe as a loss of meaning may sometimes be an encounter with this structural failure. They may sense that their activities no longer connect intelligibly to what those activities are supposed to accomplish. A teacher experiences a school that rewards test preparation over understanding. A professional experiences an institution that rewards reporting over competence. A participant in a large organization experiences an expanding machinery of procedures that seems increasingly disconnected from the work itself.
People may feel the loss without possessing a causal model of it. Their felt experience becomes evidence worth investigating. The danger is that the prevailing proxies may classify these signals as resistance, dissatisfaction, inefficiency, or noncompliance. The institution then becomes capable of suppressing feedback that could expose its own drift. This produces a self-concealing failure mode: the more successfully a proxy dominates the system, the less visible its blind spots may become.
Blind spots are inevitable. We must remain aware that they exist and preserve the ability to discover which ones matter. That process is iterative. We use representations, encounter problems, create explanations, revise our representations, and continue. It need not be formally recursive. We do not need a perfect model of our models. We need an ongoing capacity for error discovery and correction.
Outside Context Problems and Requisite Variety
Why is this problem existential?
Iain M. Banks introduced the idea of an Outside Context Problem: an encounter with something so far outside a civilization’s existing expectations and capabilities that its established ways of understanding and responding may be inadequate. The concept illustrates a fundamental constraint on survival. We cannot prepare a predetermined response to every future problem because we cannot enumerate every problem that will arise. A civilization therefore needs more than an inventory of known solutions. It needs the capacity to generate new solutions.
This connects to W. Ross Ashby’s law of requisite variety, a foundational principle of cybernetics. Stafford Beer subsequently applied these cybernetic principles to organizational viability, demonstrating how complex systems can sustain effective adaptation through distributed regulation and requisite variety. In simplified terms, effective regulation requires sufficient response variety to address the disturbances that matter to the system. For civilization, requisite variety cannot be understood merely as a fixed number of available responses. Universal explainers can create new explanatory knowledge. They can invent responses that did not previously exist. The essential capability is therefore generative. Civilization must sustain the capacity to produce the variety required by problems it does not yet know how to formulate.
The Scale Attractor can threaten precisely this capability. Standardization, specialization, and optimization can greatly increase performance within established domains while reducing the diversity of approaches available for unfamiliar situations. A system can become more capable according to its existing measures and less capable of confronting disturbances outside those measures. The scale of its operations grows while its adaptive repertoire contracts. This is a particularly dangerous form of institutional success.
The same selection pressure favors exploitation over exploration. Established capacities produce results that are easier to measure, standardize, and scale, while exploration seeks possibilities whose value may not yet be recognizable through prevailing proxies. Eric Weinstein’s Edge essay on Excellence explores a related distinction. In it, he contrasts expertise within an existing fitness landscape with the capacity to transform the landscape itself. Exploitation and exploration are complementary, but the Scale Attractor systematically favors one side of that complementarity. Exploitation can generate the surplus and slack necessary for exploration, while exploration creates the knowledge that renews and expands the capacities available for exploitation. The Scale Attractor threatens this regenerative relationship by systematically rewarding exploitation while eroding the conditions necessary for exploration.
The Human Ecosystem
Requisite variety is constitutive of viability, including the viability of the human ecosystem. Members of the human club possess universal explanatory capacity, but they do not possess identical knowledge. We inhabit different environments, encounter different problems, develop different skills, and pursue different lines of inquiry. Those differences matter.
A conceptual distinction developed in one community may enable a discovery in another. A practical technique preserved by a small group may become indispensable under changed circumstances. An unconventional explanation rejected by prevailing institutions may eventually reveal a blind spot in their dominant models. The human ecosystem generates possibilities through the interaction of its members. It contains niches in which unusual practices can develop, relationships through which knowledge can circulate, and independent communities capable of pursuing explanations that prevailing institutions neglect.
Its value is not reducible to the sum of present individual capabilities. The human ecosystem is part of the infrastructure that generates future capacity. This means variety isn’t just a preference or a matter of representation. Relevant diversity is a constituent of civilization’s adaptive potential. That does not mean every difference is equally useful or that every competing explanation is equally good. It means we cannot know in advance which differences will prove essential to the next problem.
Monoculture is dangerous because it produces correlated blind spots. When an entire civilization relies on the same representations, the same errors can propagate across institutions that appear separate. Common standards provide enormous coordination benefits. But a shared standard can also create a shared failure mode. A civilization that eliminates independent explanatory traditions in pursuit of efficiency may destroy options it cannot reconstruct when circumstances change. The Scale Attractor therefore threatens not only individual autonomy but the ecological conditions under which new knowledge is created.
The Archive and Representational Repertoire as Commons
The human ecosystem depends on a shared inheritance, a generational endowment. One form of this is a living Representational Repertoire that is archived as a commons. It comprises the accumulated and continually developing means by which the human club represents, explains, investigates, criticizes, and acts upon reality.
Science is a major part of this repertoire, but the repertoire extends beyond formal scientific knowledge. It includes languages, mathematics, conceptual distinctions, theories, models, methods, technical practices, historical records, failed experiments, abandoned approaches, and other forms of transmissible knowledge. This is an important civilizational commons.
It allows individuals to begin from discoveries they did not personally make. It permits explanations from different domains to be recombined. It allows knowledge to survive the disappearance of its original creators. Most importantly, it preserves possibilities whose future significance cannot currently be known. That is why archival efforts matter.
Archiving is more than preserving the past for cultural appreciation. It is maintaining access to representations that may become useful under future conditions. A failed theory may contain an overlooked insight. An obsolete technology may embody a principle worth recovering. An endangered language may preserve distinctions or practical knowledge absent from dominant representations.
Of course, storage alone is insufficient. Archived material must remain discoverable, interpretable, and recoverable. The living human ecosystem must retain the capacities to engage with it, criticize it, and extend it. Preservation and generation are complementary. An archive without living explanatory capacity becomes inert. A creative civilization without memory repeatedly destroys possibilities accumulated through earlier discovery.
The Scale Attractor can threaten both. Archiving, maintaining alternative representations, and sustaining independent traditions impose costs that are immediate and measurable. Their benefits may be uncertain, deferred, or impossible to express through prevailing performance indicators. Consequently, optimization for present efficiency can systematically undervalue the Representational Repertoire. A civilization can appear to be accumulating knowledge while progressively impoverishing the range of ways it can understand reality. This is a loss of requisite variety, even if its dominant institutions report increasing success.
The Civilizational Viability Kernel
The model now requires a governing orientation. In viability theory, developed notably by Jean-Pierre Aubin, a viability kernel is the set of states from which a system can continue satisfying specified constraints through feasible actions. Extending that concept to civilization gives us the Civilizational Viability Kernel. It represents the conditions under which civilization retains the capacity for continued, open-ended existence.
This cannot be a fixed technological blueprint or a fully known boundary. We cannot completely specify the future disturbances civilization will encounter or the knowledge it may create in response. Our understanding of viable states and feasible actions must itself remain open to revision. The viability kernel is therefore an orienting concept, not a final inventory of permissible states.
The governing meta-principle is: Maintain and regenerate the requisite variety necessary to keep civilization within its viability kernel indefinitely.
This is the connection between existence, value, capacity, and the human ecosystem. Capacity building matters because continued existence requires adaptive capability. Requisite variety matters because disturbances and Outside Context Problems can exceed our present repertoire of responses. Meaning matters because our representations must remain connected to reality sufficiently to expose consequential blind spots. Heterogeneity matters because the human ecosystem generates possibilities that no single perspective can anticipate. Archiving matters because the Representational Repertoire preserves possibilities across time. Institutions matter insofar as they sustain and regenerate these conditions. And scale matters because it enables capacity building while simultaneously creating the pressure that can degrade those conditions.
The meta-principle does not demand that every institution persist. Some institutions are temporary. Some should dissolve. Others must be redesigned or replaced as circumstances change. But the functions that are genuinely load-bearing for continued existence must be sustained or regenerated. Institutional regeneration need not mean institutional replication. A civilization must be able to discard obsolete structures without destroying the capacities they were responsible for maintaining.
Governance Without Appeals to a Higher Authority
This framework also changes how we should think about governance and technology. A governance system cannot become adequate merely by articulating purposes more precisely and enforcing fidelity to them. Some purposes are themselves destructive. A system can be perfectly faithful to an objective that undermines the conditions necessary for existence. Accountability does not solve this problem. Accountability ensures answerability to a standard. It does not establish that the standard is sound.
Nor can the problem be resolved by introducing a supposedly superior technological authority. Software can extend our representational and computational capacities. It can help identify inconsistencies, compare models, coordinate activity, and discover anomalies. Those are potentially valuable contributions to capacity building. But calling software artificial intelligence does not confer special epistemic authority.
Intelligence is intelligence. If a system genuinely possesses universal explanatory capacity, it belongs to the same epistemic category as other universal explainers. If it does not, it remains a tool, however sophisticated. In neither case does it become an infallible guardian of human purposes. The human club cannot delegate away the responsibility for creating and criticizing knowledge.
A particularly dangerous governance architecture would be one that formally preserves human authority while progressively eliminating the practical capacities necessary to exercise it. The club might retain nominal oversight while losing the independent knowledge, institutional alternatives, and explanatory competence needed to challenge the system. Such a system could report increasing alignment while hollowing out the conditions of self-government. The proper role of technology is therefore capacity building, not the establishment of a substitute authority.
Living With the Scale Attractor
The Scale Attractor cannot be addressed by simply abandoning scale. Scale is indispensable to many of the capacities that sustain modern civilization. Nor can proxies be eliminated. Representations are indispensable to coordination, and every finite representation has limitations. The challenge is to obtain the benefits of scale without allowing its selection pressures to systematically destroy the conditions for continued adaptation.
This requires a different orientation toward institutional design. We should preserve opportunities to challenge dominant representations, compare alternative explanations, investigate anomalies, and revise purposes when experience reveals their inadequacy. We should protect the human ecosystem’s capacity to generate independent knowledge, including forms of inquiry that prevailing performance indicators cannot yet justify. We should maintain accessible archives and cultivate the ability to recover dormant knowledge.
We should recognize the importance of redundancy, local competence, experimental niches, and institutional plurality, even when these appear inefficient through the dominant proxies. We should treat institutional purposes, mandates, performance indicators, and governance architectures as fallible conjectures rather than final authorities. And we should distinguish growth in immediately measurable capacity from growth in the capacity to regenerate, adapt, and survive.
None of these practices provides a permanent solution. They are themselves vulnerable to becoming proxies. An institution could optimize for a metric of diversity while destroying useful diversity, or accumulate archival material while losing the ability to interpret it. Even measures intended to counteract the Scale Attractor can be absorbed into it.
This is why the corrective cannot be reduced to a single fixed institutional design. It must remain an open-ended, iterative process of knowledge creation and error correction. Perfect foresight is impossible. The objective is the sustained ability to discover and correct consequential errors before those errors destroy the capacities required for further correction.
The Fundamental Tension of the Scale Attractor
We can now state the problem in full. Existence requires capacity building. Capacity building benefits from scale. Scale generates a demand for proxies that reduce coordination friction. Proxies achieve part of that reduction through selective omission. The Scale Attractor rewards proxies that facilitate further scaling, including those whose consequential blind spots make them increasingly detached from the capacities they were intended to build. As those proxies become optimization targets, institutions can drift away from reality while continuing to appear successful.
The resulting system may become more powerful, more standardized, and more efficient while degrading its own requisite variety. It may impoverish the Representational Repertoire, weaken the human ecosystem, suppress feedback, and eliminate capacities that will be indispensable when Outside Context Problems arise. The paradox is that the very process required for our survival can generate pressures that threaten our survival.
This does not make capacity building bad. Nor does it make scale inherently destructive. It means that capacity building must remain oriented toward the conditions of indefinite existence rather than toward its own immediately scalable representations. The central civilizational problem mustn’t be how to maximize scale or how to eliminate uncertainty. It is how to sustain an ecosystem of universal explainers capable of continually generating the knowledge and requisite variety needed to remain viable in a world that cannot be fully anticipated.
The Scale Attractor explains why that task is difficult. Meaning, in its structural sense, helps us recognize where our representations have lost contact with reality. The human ecosystem supplies the diversity of perspectives, problems, and explanatory pathways through which new possibilities arise. The Representational Repertoire preserves and extends those possibilities across generations. The Civilizational Viability Kernel provides the orienting meta-principle.
We build capacity to continue existing. We scale to build capacity. But we must never allow the machinery of scale to destroy the capacities that make continued existence meaningful and possible.