← Back to the reading room
THE CORE THESIS

When cognitive work becomes capital.

What changes when an output no longer requires a fresh human hour every time? The boundary between selling time and owning production starts to move.

Original X thread: 2026.09.09 · Website adaptation: 2026.10.04

Contents & sources
  1. From hours to systems
  2. An intellectual map, not borrowed authority
  3. Cheap cognition can erode profits
  4. How the thesis can lose
  5. Primary sources checked

Capital can increasingly rent some cognitive work by the token. Intellectual labor was already bought and sold. The change is that certain tasks can now be executed repeatedly without purchasing a fresh human hour for each output.

From hours to systems

Research, design, code and analysis were already economic activities. AI makes parts of them repeatable production processes. The question becomes who controls the system, and who can sell what it produces.

Our proposed loop is capital → compute → cognitive work → products → revenue → more capital. Every arrow has conditions: accuracy, energy, demand, competition and execution. A better model does not complete the loop on its own.

An intellectual map, not borrowed authority

Marx’s general intellect asks how social knowledge becomes embodied in productive machinery. Nick Land imagines capital and technology accelerating each other. Cognitive capitalism traces collective knowledge into value. Platform capitalism examines ownership of infrastructure other producers depend on.

These perspectives are not interchangeable. Citing a critical author does not make that author an advocate of our pro-capitalist direction. Extreme Capitalism is Kamui’s editorial lens on production and ownership in the age of AI.

Cheap cognition can erode profits

Lower cognitive costs also lower entry barriers. When competitors can rent the same tools and produce similar outputs, prices and margins can fall. Using AI or owning a model is not automatically a durable advantage.

Our hypothesis is that value may move toward scarce complements: trusted distribution, verification, power and real-world execution. The scarce layer changes by industry and over time.

How the thesis can lose

Measure cost per verified task, including human review and retries. Watch whether smaller firms gain ground, and whether productivity gains reach wages, customer prices or profits.

If access spreads and competition prevents profits from concentrating, the concentrated-ownership version of the thesis weakens. Counterevidence belongs inside the argument.

Build with AI. Own useful productive systems. Compete on measurable outcomes.

An editorial adaptation of Kamui’s original X thread, retaining its argument and countertests. It is not a recommendation to buy or sell a particular asset.

From a way of thinking
to an allocation of capital.

Invest in the fund

30-day lock-up · 10% profit share. Leveraged derivatives can lose all principal. Past performance does not guarantee future returns.