Blog
On engineering orgs, system design, and what AI actually changes — and what it does not.
The Repository Is All You Need
Tech startups don't need twenty platforms. The codebase is already the product — everything else is a file waiting to live beside it.
The Memory Illusion: AI Persistence Isn't Intelligence
AI memory systems solve storage, not memory. Retrieval precision and relevance judgment are both unsolved — and wrong context actively degrades quality.
First Principles Stopped at the Org Chart
Product decisions get first principles. Org structures get copy-pasted. Conway's Law guarantees the gap costs more than most leaders admit.
AI Will Help You Drift With Confidence
The most expensive thing you can do with AI is hand it the wrong problem. It will not tell you. It will help you drift with confidence — and without domain knowledge, you will not notice until it is too late.
You Can't Prompt Your Way Out of Ignorance
Training data is the internet's average, hallucinations baked in. Domain knowledge is the grounding layer that makes AI useful rather than confidently wrong.
Your Meetings Are the New Technical Debt: AI Compounds It
AI doesn't fix broken communication — it accelerates through it. Ignore your decision-making infrastructure while adopting AI and the debt compounds silently.
AI Doesn't Read Code. It Reads Patterns.
AI code generation works by pattern-matching, not reading intent. Why it fails in enterprise — and why pattern-first codebases feel like magic.