How building changes when agents take part. Coding agents can plan, write, test, localize, and debug. The shift that matters is not that the machine writes more code — it is that the developer becomes responsible for the constraints, the review, the tests, and what counts as done.
Coding agents can plan a feature, write the code, run the tests, localize the strings, and triage a crash. Developing Intelligence is the discipline of building with them without surrendering authorship. The keystroke production drops; the responsibility does not move — it concentrates.
What the developer keeps is the part that was always the work: the architecture, the constraints, the review, the tests, and the decision about what counts as done. The agent proposes. The developer still decides — the same boundary the rest of this section is built on, now applied to the act of building itself.
The agent writes the code. The developer still owns the architecture, the constraints, and the verdict on done.
The Developer page is the workshop — the craft of building excellent Apple-platform software. Developing Intelligence is the theory of that workshop: what changes when agents participate, and how authority survives it.
The essay, the build, and the field notes arrive in dedicated sessions. This page states what is coming rather than pretending it is already here.
Building with agents without surrendering authorship. The review, the tests, and the verdict on done stay with the developer.
Agentic coding is governance applied to the engineering loop. ClawLaw is the safety model: tool permissions, gates, and an audit trail for what the agent did.
Governed Intelligence →The practical Apple-platform toolchain — Swift, SwiftUI, Xcode, the agentic workflow as a daily practice — lives in the workshop.
Developer →If agents write the code, the human's edge becomes judgment: how to read, question, review, and decide what to trust. The formation behind the authority.
Learning Intelligence →