In his Something Big Is Happening post, Matt Shumer compared AI to February 2020 - the quiet phase before people realized the world had already shifted. The comparison captures the lag in perception, but it still implies that a synchronized awakening is coming. It isn’t.
There will be no single moment when everyone agrees that work has changed. What has shifted is more structural: the hard part of knowledge work is no longer what most institutions believe it is.
For decades, execution was the constraint. Writing production systems, drafting contracts, building financial models, producing research, preparing policy — all of it consumed scarce cognitive effort. Because execution was expensive, coordination structures evolved to protect it. Planning cycles, layered approvals, consensus rituals, cross-functional ceremonies — these were rational control systems in a world where implementation capacity had to be carefully rationed.
That assumption is weakening across domains.
AI systems now assist in drafting, modeling, analyzing, coding, synthesizing, summarizing, and designing. In any field where work primarily happens on a screen, the marginal cost of iteration has dropped sharply. Execution is not free. But it is dramatically cheaper.
Coordination did not become cheaper at the same rate.
That asymmetry is the real shift.
The Constraint Migrated
Most institutions still operate under assumptions that once made perfect sense.
- Execution is scarce.
- Coordination prevents waste.
- Collective alignment must precede implementation.
But when iteration becomes cheaper and reversible, the bottleneck moves. Waste appears less in implementation inefficiency and more in misalignment, delay, and over-coordination.
Today, friction most often shows up as unclear intent, prolonged alignment cycles, approval chains calibrated to an outdated cost curve, and defensive behavior under uncertainty. This is not a software problem. It applies to legal departments, consulting firms, financial institutions, research organizations, public agencies — anywhere structured thinking and digital output dominate.
Individuals feel the shift before institutions do. A lawyer compresses research time dramatically. An analyst runs ten scenarios instead of two. A consultant produces structured drafts in hours rather than days. A manager synthesizes data instantly.
Yet the coordination cadence remains unchanged.
That creates structural tension.
Engels’ Pause
Between roughly 1790 and 1840, Britain experienced explosive productivity growth during the Industrial Revolution. Mechanized production and steam power radically increased output. National income rose, but real wages for ordinary workers barely moved for decades.
This period is often called Engels’ Pause, associated with Friedrich Engels, who documented the living conditions of industrial workers in England. Later economic historians quantified the paradox: productivity climbed, yet median living standards stagnated.
The economy had transformed; the distribution mechanisms had not. Technology moved first, institutions lagged, and compensation adjusted last. There was no synchronized awakening - the shift propagated unevenly.
Some adapted early. Others resisted. Tension accumulated quietly. The defining feature of Engels’ Pause was decoupling: productivity and lived experience moved at different speeds. That is the relevant analogy.
Execution capacity in knowledge work is shifting rapidly. Compensation models, legitimacy structures, and coordination rituals remain calibrated to execution scarcity.
Capability has changed; institutional logic has not.
Two Realities Inside One Organization
Inside many organizations today, two realities coexist.
The formal reality revolves around planning cadences, distributed ownership, layered approvals, and consensus-driven decisions.
The informal reality increasingly includes AI-assisted execution, rapid private iteration, asymmetric individual throughput, and quiet automation of previously manual tasks.
Some individuals operate on a new cost curve. Others remain calibrated to the old one. The language to describe this inversion is still emerging. Explaining that “the hard part changed” sounds abstract until you experience it directly.
This is why it can feel like Cassandra syndrome. The warning is not about catastrophe. It is about a constraint shift that is unevenly recognized.
Consensus as a Risk Multiplier
Consensus culture evolved under a specific constraint: execution was expensive and largely irreversible. When building something required significant coordinated effort, broad alignment before action reduced risk. Under those conditions, consensus was rational; but when execution becomes cheap and reversible, the risk profile changes.
AI-augmented individuals - centaur units - can prototype, test, refactor, and validate at speeds that previously required distributed team effort. Exploration becomes low-cost. Iteration becomes cheap. Reversal becomes manageable.
In that environment, requiring pre-emptive consensus before experimentation does not reduce risk. It delays learning.
Centaur work introduces asymmetry. One individual, properly augmented, can generate architecture, implementation, and domain-aligned output at a level that previously required coordinated team effort.
Consensus rituals are designed to preserve symmetry. Shared cognitive ownership. Distributed responsibility. Collective visibility before action.
When asymmetric capability emerges inside a symmetry-preserving culture, tension is inevitable. The coordination model interprets asymmetry as risk. Normalization follows. More discussion. Architectural debates. Re-implementation under shared visibility. However, equilibrium calibrated to an outdated cost structure becomes drag.
- In a high-cost execution world, alignment before action minimizes waste.
- In a low-cost execution world, action before alignment often produces clarity faster.
If consensus remains a prerequisite for experimentation, organizations will systematically underutilize centaur leverage.
The Real Danger Is Overcorrection
History adds a sharper warning.
Engels’ Pause did not end in a neat institutional adjustment. Industrial transformation generated political reactions. One ideological response to early industrial dislocation was articulated in The Communist Manifesto, co-authored by Engels. It emerged from a context in which productivity had surged while workers’ conditions lagged behind.
When capability and distribution decouple long enough, ideological pressure builds.
The modern risk is not a replay of 19th-century communism. History does not run on templates. But the structural dynamic is familiar. If asymmetric productivity gains accumulate without visible institutional redesign, political energy will seek an outlet. This is where the real overcorrection risk lives.
Late recognition of the inversion may trigger blunt responses. Headcount reduction becomes the visible signal of adaptation. Efficiency becomes the narrative. Organizations shrink without redesigning coordination. Fear rises. Defensive behavior spreads.
Under those conditions, calls for centralization grow. Not necessarily in the language of Marx. More likely in the language of stability, fairness, governance, and safety. Centralized AI procurement. Standardized decision engines. Expanded compliance layers. Top-down monitoring justified as risk mitigation.
A neo-Soviet outcome would not look like 20th-century state socialism. It would look like technocratic centralization powered by AI.
In that scenario, the centaur unit is reframed. Not as an autonomous amplifier of intent, but as a controllable production node. Asymmetric capability is politically flattened rather than structurally integrated. Autonomy is reduced in the name of coherence.
That is the dystopian branch; and it does not require ideological zeal. It only requires mismanaged asymmetry plus institutional panic.
The destabilizing variable is not AI capability; it is delayed cultural adaptation followed by centralized reaction. If execution cost has fallen, the correct move is not to suppress asymmetry. It is to redesign legitimacy structures so that asymmetric capability remains accountable and aligned with intent.
Without redesign, external coordination layers will grow by default. History suggests that when institutions fail to update voluntarily, structure is imposed.
Collective Realization
Organizations are abstractions composed of individuals interpreting signals under uncertainty. Many individuals already sense that effort and output are no longer tightly coupled. They feel that something fundamental has shifted.
Yet collectively, behavior remains largely unchanged.
Executives are not omnipotent drivers of change. Structural shifts rarely occur through command alone. They occur when enough participants internalize that the facts on the ground have changed.
This moment is not about control; it is about updating the shared mental model before the system forces the update through crisis.
What This Actually Means
This is not an extinction event for knowledge work; it is the end of a scarcity assumption. When execution becomes cheaper, scarcity migrates toward clarity, alignment, and verification. Institutions that continue behaving as though implementation is the primary constraint will slow themselves down. Institutions that redesign around intent and feedback loops will compound advantage quietly.
Engels’ Pause demonstrates that structural economic shifts can decouple productivity from lived experience for decades. The question is not whether capability has increased; the question is whether coordination logic will adapt before political pressure demands centralization.
Consensus culture, once a risk mitigator, can become a risk multiplier if it does not evolve. The divergence ahead will not be driven by technology alone; it will be driven by whether institutions integrate asymmetry intelligently - or attempt to suppress it after the fact.