Blog
Working notes from the agent stack.
Long-form writing on agent architecture, multi-agent systems, and the practice of building with AI: the systems behind the lessons, explained in plain language.
Field Guide · Stanford CS329A
Build the loop, not the myth
A practical field guide to self-improving AI agents: inference-time search, verification, synthetic learning loops, long-horizon evaluation, and prompts for Codex, Claude, and Gemini.
Read the article →
Field Notes
Don't Outsource the Thinking
A daily practice for using AI to sharpen your critical thinking instead of replacing it. Written from product design, but it applies to any product development discipline.
Read the article →
Build Log
How My Agent System Works (And Where It's Different)
A plain-language explanation of the five-agent system I built for a recent project: how it compares to other AI agent tools, where it is genuinely better, where it is not, and why it matters for product design work.
Read the article →Peer Spotlight
Evals for the AI PM
My peer Cait Batzinger's first article: AI product managers should hold themselves to the same quality bar they hold their products to: a four-tier competency framework plus an AI coaching agent, Meridian, that scores her against real work output.
Read on The Floor Is Lava ↗