8.7 — Superhuman Crisis and Asymmetric Power

Fictional in its literal form, and it maps onto one of the most real and important problems there is.

The fictional scenario: an individual or small group acquires capability far beyond everybody else, and there is no mechanism to constrain them.

The real version: that has happened repeatedly, it is happening now, and the question of what to do about it is live.

The actual problem

Asymmetric power: what happens when a small number of actors can affect a very large number of people, and the affected people have no meaningful recourse.

And the historical instances are not hypothetical.

Nuclear weapons. **A handful of individuals, in a handful of states, hold the capacity to end civilisation. The constraint is a set of doctrines, treaties and procedures — not a law anybody can enforce.

Concentrated wealth and corporate power, where decisions affecting hundreds of millions are made by very small groups.

Surveillance capability, where states and companies can observe populations in ways individuals cannot detect or refuse.

Engineered pathogens, where the technical barrier to creating something dangerous has fallen substantially and continues to fall.

And artificial intelligence, where the question of what happens if capability concentrates rapidly is currently one of the most seriously discussed problems in technology policy.

In every case the structure is the fictional one: enormous capability, few holders, and no adequate mechanism of accountability.

What actually constrains power

Which is the useful content, because it is the same answer in every case.

**Fiction resolves this with a stronger individual. Reality has never resolved it that way.

Five real mechanisms.

Distributed capability

Power is constrained when others have comparable power.

This is the logic of deterrence, of competitive markets, and of federalism. It is unstable and it has prevented a great deal.

Institutions with teeth

Courts, regulators, inspectors, auditors, and treaties with verification.

Weak, slow, and the only thing that has ever reliably constrained concentrated power over time.

And the critical feature is verification rather than agreement. An arms control treaty with inspections works; one without them does not.

Transparency

Power operating in the open is constrained by reputation, politics and law. Power operating in secret is not.

Which is why disclosure requirements, free press and whistleblower protection are structural rather than decorative.

Internal refusal

And this is the most under-rated.

Concentrated power requires the cooperation of many ordinary people — engineers, officials, soldiers, administrators.

Historically, refusal at that level has stopped things: the officer who did not pass on the launch order, the engineer who reported the fault, the official who leaked the document.

Volume VII's 13.7 is the chapter on why people do not do this and what makes it possible: the line decided in advance, the ally found before the room, the question asked rather than the accusation made.

This is the mechanism most available to an individual and the one most people never think of as a mechanism at all.

And distributed knowledge

A population that understands what is happening is much harder to act against than one that does not.

The individual's position

Which is the honest part.

Against genuinely asymmetric power, an individual has very little direct leverage.

And the things that do work are unglamorous.

Do not participate in the harmful thing. The single most available action, and it is what internal refusal consists of.

Support the institutions that constrain it, which is boring and is what the mechanism actually runs on.

Maintain your own independence where you can — of information sources, of finances, of a single employer, of a single platform. Dependence is what converts asymmetry into control.

Understand the thing. A population that cannot evaluate a technology cannot govern it.

And find the others. Individual refusal is fragile; collective refusal is not.

What the fictional version gets wrong

Three things, and each is instructive.

It personalises. The threat is a person, so the solution is a person. Real asymmetric power is structural — it persists across the individuals holding it, and removing one changes little.

It resolves. The crisis has an ending. **Real ones do not end; they are managed, continuously, for decades. Nuclear weapons have been managed for eighty years and are still being managed.

And it assumes the powerful actor is malicious. Most real harm from concentrated power comes from indifference, from unexamined incentives, and from decisions made without the affected people in the room — not from villainy.

Which changes the response entirely: the answer to indifference is representation and accountability, not opposition.

The version that is actually approaching

Because it deserves naming honestly.

The capability question in artificial intelligence is the closest real analogue to this chapter's fiction, and it is genuinely unresolved.

The serious concerns are the same structural ones: rapid capability concentration, decisions affecting billions made by small groups, difficulty of verification, and the absence of institutions with jurisdiction.

And the serious responses are the same five mechanisms: distributed capability, institutions with verification, transparency, the ability of people inside to refuse, and a public that understands enough to govern.

This book takes no position on the technical questions. It notes that the structure of the problem is the one this chapter describes, and that the mechanisms that have historically worked are the ones listed above.

What this drill actually trains

Four things.

Recognising asymmetric power when you see it, including in ordinary settings — an employer, a landlord, a platform, an institution.

Knowing what constrains it, which is never a stronger individual.

Understanding that the most available lever is refusal from inside, and that it requires a line decided in advance.

And distinguishing structural problems from personal ones. The error the fiction makes — treating a structural problem as a personal one — is made constantly in real life and it is why so many attempted remedies fail.

What to do with this page

Asymmetric power is real, it is current, and the fictional framing personalises what is structural.

Nothing has ever been constrained by a stronger individual. Distributed capability, institutions with verification, transparency, internal refusal, and an informed public.

The most available lever to you is not participating, and Volume VII's 13.7 is how that is actually done.

Maintain independence — of information, of income, of platform. Dependence is what turns asymmetry into control.

And distinguish structural from personal. Getting that wrong is why most remedies aimed at a person achieve nothing.

Next: 8.8 — what the fictional drills actually trained, assembled.