Insights on AI, software development, and building better digital products
AI-assisted projects turn to slop when there's no foundation. Lock two files โ a plain-English object definition and a derived data model โ and AI has to build on rock instead of shifting sand.

We broke AI agents into domain-specific subagents with explicit ownership of routes, types, and business rules. The codebase got better overnight.

Under deadline pressure, API routes accumulate bad habits. Here's how a dedicated agent enforces REST standards, Zod validation, and permission checks on every route, every time.

The frontend is always last and always rushed. Here's how a dedicated frontend agent with a fixed execution order eliminated merge conflicts and built UI against stable contracts.

Schema changes are where features go wrong. Here's how isolating database work behind a dedicated agent eliminated a whole class of deployment failures.
Tests written after a feature ships are just documentation. Here's how we made testing parallel with implementation using a dedicated test agent.
You have reached the end of the posts.