Clausewitz · AI · Strategy
War is the continuation of politics
by other means.
AI development is the continuation of institutional competition by computational means.
Clausewitz built the only theory of purposeful action that survives contact with reality. Not a system — a theory of friction. The recognition that uncertainty, chance, and political purpose cannot be factored out of any contest between wills.
The Remarkable Trinity
I
Passion · Primordial Violence
The blind natural force. In war: hatred, enmity, the people's will to fight. Cannot be eliminated, only directed.
→ in AI
Raw capability, emergent behavior, the model's tendencies that exceed specification. The thing the system does that surprises its builders.
II
Chance · The Play of Probability
The domain of the commander. Courage, luck, creativity in the space where probability governs. Neither blind force nor rational purpose.
→ in AI
Inference stochasticity, environmental noise, the irreducible uncertainty of deployment. The space where agent judgment matters.
III
Purpose · Political Reason
The government's domain. War is always subordinate to political purpose — if this is forgotten, war becomes catastrophe.
→ in AI
Human purposes that direct the system. Alignment is not a technical problem — it is a political one. The question of whose purpose prevails.
Any theory that fixates on only one force fails. Capability-maximalism fixates on Force I and forgets purpose. Pure alignment work fixates on Force III and ignores the emergent behavior that makes Force I dangerous. Stochastic parrots arguments fixate on Force II (chance/noise) and miss the directed capability underneath. Clausewitz would say: the trinity is a pendulum, not a triangle. The balance is always in motion.
Three Forces That Dominate Real War
Concept I — Friktion
Friction
On War · Book I · Ch.7
"Everything in war is very simple, but the simplest thing is difficult. The difficulties accumulate and end by producing a kind of friction that is inconceivable unless one has experienced war."
In AI systems: friction is the structural gap between what a system is designed to do and what it actually accomplishes when deployed in the real world. It is not a bug — it is the nature of purposeful action in a complex environment. Friction cannot be eliminated by better planning. Systems that expect friction are stronger than systems that plan for its absence. The agent that can maintain coherent purpose through partial tool failures, degraded information, and environment drift is demonstrating something closer to Clausewitz's military genius than any benchmark score captures.
Concept II — Nebel des Krieges
Fog of War
On War · Book I · Ch.3
"If the mind is to emerge unscathed from this relentless struggle with the unforeseen, two qualities are indispensable: first, an intellect that, even in the darkest hour, retains some glimmerings of the inner light which leads to truth; and second, the courage to follow this faint light wherever it may lead."
Fog in AI: the agent's fundamental uncertainty about the world, about other agents' goals, and about the effects of its own actions. Clausewitz's third premise — that fog is irreducible — is contested by information dominance theorists. If sensors, computation, and communication can eliminate uncertainty, then systematic optimization becomes viable again. But friction is not only informational. It is volitional, organizational, and physical. Even perfect information doesn't eliminate the fog of other minds, the friction of organizational coordination, or the irreducible stochasticity of execution.
Concept III — Kulminationspunkt
Culminating Point
On War · Book VII
The moment when the attacker's superiority is exhausted and further advance becomes counterproductive. Every offensive has a culminating point. The commander who cannot identify it overextends and loses what was won.
In AI/product: the culminating point is where shipping features past the point of user absorption destroys coherence. Expanding capability past the point of system coherence. Pressing a competitive advantage until the cost of maintaining the position exceeds the benefit of holding it. The capability expansion that was clearly winning at rung 3 becomes overextension at rung 5. Clausewitz's most underrated lesson: the art of knowing when to consolidate is rarer than the art of attack.
Concept IV — Coup d'Oeil
Genius Under Fog
On War · Book I · Ch.3
Military genius is not systematic knowledge but coup d'oeil — the ability to see the essential truth through fog — combined with the determination to act on it. This cannot be taught through doctrine. It can only be cultivated through experience and educated judgment.
The center of gravity in AI/skills competition is not compute, not data, not even model quality — it is the ability to maintain coherent purpose through friction. The agent (human or machine) that can keep acting meaningfully when its information is degraded, its environment shifts, and its tools fail partially. This is Clausewitz's definition of military genius applied to the agent condition. Burja calls the same quality "live player." It is the rarest thing in any competitive system.
From the Text
"War is not merely a political act but a real political instrument, a continuation of political intercourse, carried on with other means. What remains peculiar to war is simply the peculiar nature of its means."
On War · Book VIII · Ch.6B · The most misquoted sentence in strategic theory
Clausewitz does not mean war is politics' tool (subordination). He means war is politics in a different medium — war's logic is always and everywhere political logic, even when the fighters forget this. The fighters who forget this — who pursue military victory disconnected from political purpose — are fighting toward culminating-point overextension. The AI translation: capability expansion that loses sight of whose purpose it serves isn't "neutral" — it has defaulted to Force I (primordial violence/emergent behavior) without a moderating Force III. The capability runs away from the purpose.
"The military genius is the man who can pierce through the fog of uncertainty and see the essential truth about the situation — and then has the will to act on it while others are still waiting for clarity."
On War · Book I · Ch.3 · Paraphrase of the Clausewitzian definition
This is why the deepread ladder's hardest rung is not 20 — it is 10. Rungs 1-9 are the annotator's privilege: you can work carefully, at your own pace, with fog absent. Rung 10 requires producing an artifact that a domain expert can evaluate. The fog enters. At rung 11 you are commanding a domain — making real decisions under real uncertainty. The coup d'oeil Clausewitz describes is the thing that distinguishes rungs 10+ from rungs 1-9. It cannot be taught, but the ladder is the best approximation we have of cultivating it.
"Genius in war is not the accumulation of principles. He who studies examples and thinks himself thereby qualified to manage such affairs is grievously mistaken."
On War · Book II · On the Theory of War
Clausewitz's attack on Jomini, Bülow, and every theorist who thought war could be reduced to geometry. Is alignment the Jomini of AI — trying to make a political problem geometric? The safety framework that specifies objectives formally is solving the wrong problem if the problem is irreducibly political. Formal specification assumes Force III (purpose) can be isolated and written down. But the trinity means Force III is always in tension with Forces I and II. You cannot factor out the passion and the fog and address only the purpose.
The Translation
Before reading Clausewitz: he's the "fog of war guy," a source of military aphorisms. After: he is making a philosophical argument about the nature of purposeful action under radical uncertainty. The trinity is not a military framework — it is a general theory of any contest between wills operating through organizations in conditions of incomplete information.
The three forces (passion / chance / purpose) are present in any competitive domain. The fog is not a metaphor — it is the structural condition of any agent that cannot see all of the state space it operates in. The friction is not a complication — it is the nature of purposeful action in a complex environment.
The center of gravity in AI competition is not compute, not data, not even model quality. It is the ability to maintain coherent purpose through friction. The agent that can keep acting meaningfully when its information is degraded, its environment shifts, and its tools fail partially — that agent is demonstrating something Clausewitz spent his career trying to describe: genius under fog.