Buying an answer. Owning the system.
A firm can become more productive and remain dependent on an AI provider and a marketplace. Creating value and capturing it are different questions.
Original X thread: 2026.09.11 · Website adaptation: 2026.10.04
Contents & sources
Imagine a small firm that rents AI to research products and sells through a marketplace. Research gets cheaper. The firm owns neither the AI service nor the route to its customers. A producer can also be a dependent customer.
Four lenses on one firm
e/acc asks how to expand useful experimentation. Technocapitalism examines control of research and knowledge. Cognitive capitalism asks whose collective knowledge makes production possible. Platform capitalism inspects ownership of infrastructure others need.
The lenses expose different pressures. Treating them as synonyms leaves the word acceleration while losing the structure of ownership and power.
The bills that survive cheap research
AI can reduce research time while fees, distribution, customer acquisition and verification remain. If a provider changes prices or a marketplace limits access, productivity gains may be captured by that layer.
Measure the full cost of producing, checking and delivering an outcome a customer accepts. Cheap tokens and cheap business outcomes are different metrics.
What do you build while you rent?
Renting services is a rational way to test ideas quickly. Use that process to accumulate useful products, expertise, operational knowledge and direct customer relationships. Ask whether recurring spending builds an asset of your own.
Owning every layer is not the goal. Capital allocation means distinguishing the layer you need to control from the layers you can buy competitively.
Ownership can fail too
Ownership brings maintenance and capital expenditure. A system that is not useful, or has no paying customer, becomes a cost. Evaluate usefulness, bargaining power and actual value capture together.
Our test: do lower research costs survive fees, verification and competition? Judge business outcomes rather than demo quality.
Follow productivity beyond its creation. Find out who keeps the gains.
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.