Writing
Working notes from building products, tinkering with AI, and thinking about product strategy. New posts appear here as I publish them.
- Why Agents Won't Kill the UI: September 17, 2026 — Agents will change how we use software, but cognition still has costs. Why hybrid interfaces and conceptual integrity will matter in an agentic future.
- Will We Learn to Love the Surveillance Glasses?: September 12, 2026 — Ambient AI wearables may become ordinary without becoming trusted. What would have to be true for constant sensing to become socially permissible?
- Is Europe's AI Sovereignty Debate Missing a Layer?: September 9, 2026 — AI agents will need identity, authorization and trust infrastructure. Europe risks relying on foreign networks to decide which agents can act and transact.
- Ship of Theseus: Can an AI Agent Keep Its Reputation?: September 3, 2026 — If every component of an agent gets replaced — model, prompt, tools, budget — is it still the same agent? And does its reputation carry over?
- Can an AI Agent Have a Reputation?: August 24, 2026 — If agents have no reputation to lose, no court to appeal to, and no colleague who remembers them — what would giving them one look like?
- Designing the Social Temperament of AI Agents: August 17, 2026 — When agents start meeting other agents in the wild, intelligence alone will not cut it. Notes on Anthropic's multiagent research and what kind of colleague an agent should be.