May 26, 2026 - 6 minute read

Deterministic Shells Control, Agentic Shells Explore

Every AI agent has a boss. Sometimes the boss is ordinary application code, calling the model for one bounded judgment. Sometimes the boss is the model, deciding what to inspect, what to call, and when to stop. That choice decides the failure mode before the first prompt ever runs.

May 19, 2026 - 8 minute read

You Get the AI You Ask For

Two engineers open the same chat window, with the same model, on the same Tuesday morning. One ships a feature with public database queries, hardcoded keys, no rate limit, and a UI that converts at half the rate it should. The other ships the same feature audited against OWASP, profiled for hot paths, copy-tested against high-converting landing pages, and fully instrumented. Same model, same week, two completely different products, separated entirely by what each engineer knew to ask the model about.

May 12, 2026 - 8 minute read

AI Agents Will Break Any Rule You Don't Test

Most teams give their AI agents an AGENTS.md, a CLAUDE.md, or a Cursor rule full of polite architectural guidance. Layers, boundaries, where secrets live, what may import from what. Three weeks later the codebase is spaghetti reaching across every line they wrote down, and they wonder why the agent ignored them. Documents do not enforce architecture. Tests do. The fix is small, language-agnostic, and unforgiving: take your most important architecture rules, write them as a deterministic test, wire it into a pre-commit hook, and let CI run it again. Now the agent literally cannot finish the work if it breaks the architecture.

May 11, 2026 - 9 minute read

Pick a Lane: Factory or Frontier

A year ago I wrote that AI workers fall into four tiers (Conscript, Cyborg, Centaur, Centurion), stacked from least output per human to most. The tiers are still real. The implied ladder isn’t. The longer I work alongside agents and watch the people who do it well, the clearer it gets that there isn’t one destination at the top of the stack. Two lanes are opening up in front of every serious operator, and they reward completely different instincts. Pick the wrong one and you’ll either over-systemize a problem nobody has solved yet, or you’ll keep hand-crafting work the world has already turned into a commodity. Both mistakes are expensive. Only one is recoverable.