CONCEPT
The Integration Leak
The most consequential and hardest-to-diagnose class of failure in AI-generated systems — not a bug in any component but a mismatch between assumptions that components make about each other, embedded in generated code, implicit and unrecoverable, discoverable only through the failure it produces.
An integration leak is a failure that does not live in any single component's code but in the space between components — in the contracts each makes about the others, in the assumptions that were never negotiated because no human was there to negotiate them. In conventionally built systems, developers who build different components talk to each other, agree on interfaces, document assumptions (imperfectly but somewhere). In AI-generated systems, components may be generated in separate conversations with separate contexts under separate assumptions. The authentication service generated by Claude in one prompt may assume sessions are stored in memory; the load balancer generated in another prompt may distribute requests across multiple instances. Both components are individually correct. Together, under load, they produce a failure that neither component's code explains. The leak lives in the integration layer, where no human has looked because no human had reason to.
