Who Gets to Authorize the Future

Why Incumbent Elites Monetize Decaying Scarcity Before a Regime Shift

The most dangerous incumbent is rarely the one whose power is fully secure.

It is the one that senses the old monopoly is ending, but still controls the passage into whatever comes next.

That is where many systems now seem to be.

Across intellectual life, media, education, finance, and parts of corporate power, the old elite still holds visible control over legitimacy, capital, certification, recognition, and large-scale deployment. But the underlying scarcities that once justified that control are weakening quickly. Writing is easier. Research is cheaper. Coding is faster. Interpretation is abundant.[1] New firms can be built with far fewer people. Individuals and small teams can now produce outputs that previously required much larger institutional machinery.[2]

The old order is not gone.

But the bottlenecks that sustained it are beginning to dissolve.

This is the real position of the present elite: not permanent rulers of a stable system, but the last owners of decaying chokepoints.

That distinction matters because it changes the nature of the conflict.

The issue is not simply whether elites will resist change. They always do, at least partially. The more important question is what incumbent classes do when they realize they can no longer fully stop the next order, but can still meter access to it.

They do not only defend.

They monetize the transition.

Every Order Is Built Around Bottlenecks

No elite rules in the abstract.

Every ruling class is downstream of a scarcity.

Aristocracies formed around land, force, lineage, and direct political privilege. Industrial elites consolidated around factories, transport, capital concentration, and administrative scale. Broadcast-era gatekeepers controlled distribution, publishing, narrative legitimacy, and the ability to decide what became publicly visible.

Intellectual elites did not merely possess ideas. They controlled the formats through which ideas became legible. Seriousness had a style. Entry had a grammar. Recognition required fluency in codes most people did not naturally inherit.

This is why elite power often appears more durable than it really is.

People mistake the social form for the underlying bottleneck.

But when the bottleneck changes, elite power does not disappear all at once. It becomes unstable. It begins to separate from the conditions that originally made it seem natural.

That is the phase many current elites are entering.

They still control the visible interfaces of authority. But they increasingly do so after the scarcity that legitimized their dominance has already started to erode underneath them.

The Weakening Elite Often Becomes More Extractive

There is a recurring pattern in regime transition.

An incumbent class often appears most aggressive not when its power is secure, but when it first senses that its monopoly is no longer guaranteed.

That is because a weakening elite is not yet a defeated elite.

It is often a more extractive one.

Once a ruling class realizes that the future may not preserve its old privileges intact, it begins to behave differently. It becomes less concerned with preserving every feature of the old order on principle and more concerned with converting existing influence into forms that can survive the transition.

It stops acting like a permanent sovereign and starts acting like a conversion specialist.

That conversion can take many forms.

A declining aristocracy may trade direct rule for financial position, bureaucratic influence, or protected status inside a constitutional order. A weakening religious authority may lose doctrinal monopoly but preserve power through education, administration, and moral certification. A fading media class may lose publication monopoly but seek new control through ranking, curation, trust infrastructure, or safety language.

The visible ideology changes.

The underlying behavior is older.

When a bottleneck breaks, incumbents do not simply vanish. They try to secure privileged placement at the next chokepoint.

AI Is Dissolving Several Old Bottlenecks at Once

This is what makes the present moment unusual.

Artificial intelligence is not dissolving only one elite scarcity. It is weakening several at the same time.

Writing quality is easier to generate. Research assistance is cheaper. Coding and prototyping are faster. High-level style is less monopolized by those who passed through elite institutional filters. The cost of making thought legible is dropping quickly.[3]

This does not mean intelligence has become equal.

It means the old taxes on legibility are collapsing.

That matters because much of modern gatekeeping depended not on exclusive intelligence in the deepest sense, but on control over recognized form. Institutions decided what counted as rigorous, serious, publishable, fundable, or credible. They did not create intelligence. They compressed it into sanctioned formats and rationed access to visibility.

As those formatting costs collapse, one of the quiet foundations of elite reproduction weakens.

The person outside the institution can increasingly sound institutionally fluent. The outsider can produce polished work, frame arguments in accepted language, synthesize sources rapidly, and iterate at speeds that previously belonged to large organizations.

This does not abolish incumbency.

But it does mean the old claim — that elite control reflects a necessary concentration of rare cognitive competence — becomes harder to defend in the same way.

When Legibility Becomes Cheap, Power Moves Elsewhere

The mistake is to think that collapsing one bottleneck automatically democratizes power.

It does not.

It relocates struggle.

When legibility becomes cheap, other layers become more important: trust, distribution, capital access, legal recognition, platform control, compute, regulatory clearance, and the ability to translate raw capability into socially recognized authority.

This is why the present transition feels democratizing and oligarchic at the same time.

Below, capability diffuses.

Above, control over deployment can harden.

A solo founder can build more than ever. A researcher outside the academy can publish more than ever. A writer without establishment backing can become more legible than ever. Small teams can produce outputs once reserved for firms, newsrooms, or institutions with hundreds of employees.[4]

But that same world can also intensify concentration in infrastructure, cloud dependency, model access, payments, compliance, distribution systems, and legitimacy shells. AI may dissolve cognitive bottlenecks while intensifying infrastructural ones.

This is already visible in AI itself. Capability is spreading, but compute, deployment, safety validation, and public trust remain concentrated in a small number of firms and institutional actors. Frontier labs now formalize staged deployment, internal safety thresholds, and escalating governance procedures around advanced capabilities, even as model capabilities continue to diffuse more broadly.[5]

The diffusion of capability does not automatically produce the diffusion of power.

It is now increasingly possible to think, write, build, and coordinate outside the old institutions. But it is still much harder to get those capacities recognized, financed, insured, distributed, and lawfully scaled without passing through incumbent shells.

That is why the transition feels simultaneously open and blocked.

The productive frontier is spreading faster than the legitimacy frontier.

The old monopoly is gone. The toll remains.

The incumbent no longer says, “You cannot enter.”

The incumbent says, “Entry still passes through me.”

The Old Monopoly Is Gone. The Toll Remains.

This is the more precise way to describe the current elite.

They may not be capable of permanently stopping the next stage.

But they can still slow it, tax it, certify it, moralize it, and convert it into channels they continue to control.

It no longer rests on total exclusion. It rests on metering.

The incumbent no longer says, “You cannot enter.”

The incumbent says, “Entry still passes through me.”

This is a different kind of power than old monopoly, but it can be highly profitable and politically durable for a time.

If AI makes writing abundant, incumbents can shift toward deciding what counts as trusted writing.

If technical production becomes cheap, incumbents can shift toward controlling licensing, contracts, audits, and deployment rules.

If one-person firms become more viable, incumbents can intensify control over finance, distribution, and acquisition.

The old monopoly weakens, but the toll survives.

Not All Elites Face the Same Future

The temptation is to speak of “the elite” as though it were one coherent class.

That is too simple.

Different elite layers face different kinds of erosion.

Intellectual elites are among the first to lose monopoly over legibility. The old prestige attached to style, credentialed seriousness, and obscure fluency weakens once explanation, synthesis, and rhetorical competence become cheap.

Managerial and bureaucratic elites may temporarily strengthen. When institutions can no longer trust older cognitive proxies, they often become more procedural, more verification-heavy, and more conservative. Under uncertainty, they retreat toward whatever remains auditable. This is already visible in education and hiring, where generative AI is pushing institutions toward tighter verification, stronger assessment-validity concerns, and more controlled evaluation formats rather than broader trust.[6]

Capital and infrastructure elites may remain powerful longest. Cheap cognition is not the same as cheap compute, legal shells, logistics, or large-scale coordination. The more intelligence diffuses, the more valuable the infrastructure that routes and scales it can become.

Legitimacy elites — political, legal, media, regulatory — may also gain importance during instability. When old standards weaken and new ones are not yet settled, the authority to certify what is real, safe, lawful, or responsible becomes especially valuable.

This is why the current transition should not be misread as simple flattening.

Some hierarchies are weakening.

Others are being rebuilt around scarcer layers.

The collapse of one elite form can become the strengthening condition for another.

The Historical Pattern Is Conversion, Not Simple Overthrow

This is where the analogy to aristocratic decline becomes useful, if handled carefully.

The most important historical pattern is not dramatic overthrow.

It is negotiated regime conversion.

In nineteenth-century Britain, the old landed elite did not simply disappear when industrial capital and parliamentary reform began to reorder the political system. Many adapted by moving into finance, administration, law, imperial governance, and the boards of the very institutions shaping the new order. The basis of dominance shifted, but elite position often survived by migrating into the new machinery of coordination rather than dying with the old one.[7]

That is closer to what many current incumbents may be doing.

They do not need to restore a lost monopoly over thought, writing, production, or entrepreneurial capacity. They only need privileged placement inside the successor system.

The goal is not always to stop the next order.

It is often to own its bridge.

Of course, not all conversion strategies succeed. Some elites will misread the transition, defend the wrong bottleneck, or discover too late that the privilege they were trying to preserve no longer maps to where power actually moved. That too is part of regime change.

The Final Human Bottleneck May Be Legitimacy

The coming transition is not blocked mainly by lack of capability.

Capability is already diffusing.

What lags is the reassignment of legitimacy.

Synthetic systems can increasingly support thinking, coordination, production, and judgment. But human institutions still control the visible grammar of permission. They decide what counts as recognized work, lawful deployment, accredited expertise, valid authorship, responsible scaling, acceptable risk, and legitimate authority.

That is why the incumbent elite matters so much in this interval.

It still holds the shells through which high-capability action must pass.

Its power is therefore neither fully substantive nor merely decorative.

It is transitional.

The current elite may be the last class whose dominance rests on controlling legacy human bottlenecks even as the real generative frontier begins moving elsewhere.

That does not make them omnipotent.

It makes them historically dangerous in a specific way.

A class that knows it cannot fully preserve the old regime, but still controls the interfaces into the new one, can distort the transition more effectively than a secure ruling class ever could.

Not by total prohibition.

By selective admission.

Why Incumbents Become More Moralizing Near Transition

One sign of an incumbent class losing structural confidence is that it often becomes more moralized in its rhetoric.

This is not because all incumbents are cynical.

It is because older justifications weaken.

When elites can no longer straightforwardly defend their role through obvious superiority, they increasingly defend it through stewardship language: safety, responsibility, trust, democratic norms, institutional seriousness, anti-chaos, public protection, quality control.

Sometimes these claims are real.

Sometimes they are cover.

Usually they are mixed.

That ambiguity is exactly what makes the moment difficult to parse.

The incumbent often is performing a real function. Civilizations cannot simply run on raw capability without coordination, trust, and stabilizing procedure. The problem is that incumbents also tend to overprice their own indispensability. They protect genuine order and legacy rent at the same time.

That is why current conflicts around AI feel so intense. Open capability rhetoric repeatedly gives way to certification, responsible-scaling language, deployment controls, and infrastructure concentration. The issue is not only safety versus recklessness. It is also who gets to define the legitimate path into the next order.[8]

What Happens Next

The likeliest future is not clean replacement.

It is prolonged struggle over which scarcities remain politically valid.

As more forms of human legibility, production, and cognitive performance become cheap, the pressure will intensify on remaining bottlenecks. Some incumbents will entrench around trust, regulation, capital, and deployment infrastructure. Others will lose relevance quickly. Some will successfully convert old influence into successor influence. Others will discover too late that the bottleneck they controlled no longer matters enough to preserve them.

The key point is that abundance at the capability layer does not guarantee openness at the deployment layer. Open models may improve, solo builders may proliferate, and small teams may ship faster than ever. But enterprise procurement, insurance, compliance, public-sector contracts, and large-scale distribution can still remain concentrated. Capability spreads. Admissibility lags.

That is why the central political question is not whether the future can be invented.

It can.

The question is who gets to authorize its reality.

Who Gets to Authorize the Future

The current elite may not be the final ruling class of the old world.

It may be something more specific:

the last class able to extract rents from bottlenecks that synthetic systems are rapidly dissolving.

Its function is not simply to defend the past.

Its function is to meter the passage into what comes next.

That is why the transition feels blocked even as capability expands everywhere.

Not because the future is impossible.

But because the future is arriving through systems whose productive logic is spreading faster than their legitimacy can be reassigned.

The decisive struggle, then, is not between elites and masses in the old populist sense. Nor is it simply between humans and machines.

It is between distributed new capability and the incumbent shells that still control legitimacy, capital, certification, and deployment.

The future will not be decided first by who can imagine it.

It will be decided by whoever still controls the terms under which imagination becomes admissible, deployable, and real.

Footnotes

[1] Stanford Institute for Human-Centered Artificial Intelligence, AI Index Report 2025, Stanford University, 2025.

[2] McKinsey & Company, The State of AI in Early 2025, 2025; Stanford Institute for Human-Centered Artificial Intelligence, AI Index Report 2025.

[3] Ethan Mollick, Co-Intelligence: Living and Working with AI, New York: Portfolio/Penguin, 2024; Stanford Institute for Human-Centered Artificial Intelligence, AI Index Report 2025.

[4] McKinsey & Company, The State of AI in Early 2025, 2025; Stanford Institute for Human-Centered Artificial Intelligence, AI Index Report 2025.

[5] Anthropic, “Responsible Scaling Policy Updates,” last updated February 24, 2026; OpenAI, “Our Updated Preparedness Framework,” April 15, 2025; Stanford Institute for Human-Centered Artificial Intelligence, AI Index Report 2025, chap. 6, “Policy and Governance.”

[6] Phillip Dawson, Margaret Bearman, Mollie Dollinger, and David Boud, “Validity Matters More Than Cheating,” Assessment & Evaluation in Higher Education 49, no. 7, 2024; OECD, Trends Shaping Education 2025, OECD Publishing, 2025; Ray A. Smith, “AI Is Forcing the Return of the In-Person Job Interview,” The Wall Street Journal, August 12, 2025.

[7] David Cannadine, The Decline and Fall of the British Aristocracy, New Haven: Yale University Press, 1990.

[8] Anthropic, “Responsible Scaling Policy Updates,” last updated February 24, 2026; OpenAI, “Our Updated Preparedness Framework,” April 15, 2025.