Loop Engineering: Building Web Apps with an Agent in the Loop
Why an AI agent needs a governed loop of observation, objective gates, iteration, and honest stopping to ship reliable web apps.
Read postWriting
Essays and field notes on building AI-powered products from 0 to 1: agentic workflows, product design, and lessons from shipping software fast.
Why an AI agent needs a governed loop of observation, objective gates, iteration, and honest stopping to ship reliable web apps.
Read postAn MCP server that lets you click directly on slide elements to give feedback to an AI agent. Why I built it, and the human-in-the-loop pattern behind it.
Read postGeography, ambient audio, and the pain of Microsoft Store IAP — a year of solo dev notes from building NearbyGame.
Read postA practical pilot of disposable micro-apps that chain from Squad Shuffer to Group Qualifier to Tournament Bracket, with optional paths for different match formats.
Read postWhy I abandoned a venue-discovery idea, reframed the problem with AI, and turned NearbyGame into a calm walking experience built for one user first.
Read postA simple local-first habit helper focused on one action: complete a routine and cross it off.
Read postWhy I pivoted MiniBreaks into a weekly micro-app experiment, and how I structured AI roles to ship one tiny app every week.
Read postWhy I built a local-first chat app with no registration and no cloud dependency.
Read post