My AI Had Fixed the Code Before I Saw It

Summary

Was this newsletter forwarded to you? Sign up to get it in your inbox.Before I opened my laptop, the code had reviewed itself.I launched GitHub expecting to dive into my usual routine—flag poorly named variables, trim excessive tests, and suggest simpler ways to handle errors. Instead, I found a few strong comments from Claude Code, the AI that writes and edits in my terminal:"Changed variable naming to match pattern from PR [pull request] #234, removed excessive test coverage per feedback on PR #219, added error handling similar to approved approach in PR #241."In other words, Claude had learned from three prior months of code reviews and applied those lessons without being asked. It had picked up my tastes thoroughly, the way a sharp new teammate would—and with receipts.It felt like cheating, but it wasn't—it was compounding. Every time we fix something, the system learns. Every time we review something, the system learns. Every time we fail in an avoidable way, the system learns. That's how we build Cora, Every’s AI-enabled email assistant, now: Create systems that create systems, then get out of the way.I call this compounding engineering: building self-improving development systems where each iteration makes the next one faster, safer, and better.Typical AI engineering is about short-term gains. You prompt, it codes, you ship. Then you start over. Compounding engineering is about building systems with memory, where every pull request teaches the system, every bug becomes a permanent lesson, and every code review updates the defaults. AI engineering makes you faster today. Compounding engineering makes you faster tomorrow, and each day after.Three months of compounding engineering on Cora have completely changed the way I think about code. I can't write a function anymore without thinking about whether I'm teaching the system or just solving today's problem. Every bug fix feels half-done if it doesn't prevent its entire category going forward, and code reviews w...