ML
Writing

Notes from things I’ve actually built

I write after shipping, not before. Most of these started in production — an agent proposing par levels overnight, a memory system deciding what to forget — and turned into lessons worth keeping.

May 2026

What a year of running an AI agent in a restaurant taught me

The agent proposes, the manager disposes. On confidence decay, approval gates, and why the boring parts of autonomy are the hard parts.

Read article
April 2026

Designing a website for AI agents as first-class visitors

ChatGPT was sending kids to ArcadeLab within 24 hours of launch. What llms.txt, machine-readable briefings, and a publish API did that SEO couldn't.

Read article
March 2026

Forgetting is a feature: building brain-shaped memory for LLMs

Power-law decay, sleep consolidation, and why an AI that remembers everything is less useful than one that knows what to forget.

Read article

RSS feed