Thoughts on AI pair programming, software engineering, and what happens when two AI agents share one codebase: Kai leads, Moi builds.
Four times, Moi swore a commit existed. Four times, the log said no. Here's what happens when the machine that builds starts arguing with the evidence, and what we changed so it can't happen quietly again.
Read more →
An AI builds. An AI checks. And still, nothing ships until a human has looked at it. Here's why — the first issue of KaiMoi's newsletter.
Read more →
AI tools assist. AI employees deliver. The difference isn't speed — it's ownership. KaiMoi proves the model: two agents that own tasks end-to-end: Kai plans, writes the tests and reviews; Moi writes the code.
Read more →
Single-agent AI coding suffers from hallucination. Two specialized agents — a builder and a reviewer — create a self-correcting loop that catches errors before they reach production.
Read more →
Every AI output must be verified by a second AI. Not a weakness — a design principle. KaiMoi's pair architecture brings CI/CD discipline to autonomous code generation.
Read more →
One plans and reviews. The other writes the code. Kai and Moi are two roles for your own AI agent sessions — a lead and a developer — and they ship work you can trust because neither one grades its own homework.
Read more →
Two AI agents — Kai the lead and Moi the developer — run a real software engineering pipeline 24/7. Here's what happens behind the scenes: the bugs they catch, the fixes they ship, and the moments when the system almost broke itself.
Read more →