Where the term comes from Contextual flow control is Alex Komoroske's name for the alternative architecture he proposes in the interview — inspired by Helen Nissenbaum's contextual integrity, the idea that privacy means data being used in contexts you understand and that align with your interests, beyond consent forms or end-to-end encryption. [moment: 1771] The problem it answers Komoroske frames the current situation with his iron triangle: a safe system can have only two of these three — 1. Untrusted code 2. Sensitive data 3. Network access - The web chose untrusted code + network access, keeping sensitive data out. - The app store model keeps sensitive data + network access, but only by vetting all code centrally. His favorite historical counterexample is the file system, where multiple apps work on the same data and coordinate without any single app owning your life. Breaking the triangle — with tools like information flow control and confidential computing — is what makes contextual flow control possible. [moment: 1771] Why it matters now Under the web's same-origin paradigm, each origin is an isolated island that sees everything you put into it — data goes in and the origin can do whatever it wants with it. That data gravity leads to aggregation. Contextual flow control is his proposed change to the laws of physics: instead of consent forms or one mega-context that mushes everything together (his example: a chatbot surfacing a salary negotiation from months earlier in a totally unrelated conversation), data would flow in contexts people intuitively understand and that align with their interests. [moment: 3592] - Contextual flow control = Komoroske's name for a privacy/security architecture inspired by Nissenbaum's contextual integrity. - It exists to break his iron triangle (untrusted code / sensitive data / network access) the way the file system historically did. - It is grouped with information flow control, confidential computing, and open attested runtimes as the leverage point against data gravity and aggregation.