article Writing

Notes from the Build Process

Practical notes on local-first AI, RAG systems, assistants, software engineering, and what breaks when prototypes meet real workflows.

article

calendar_today July 07, 2026

AI Agents Need Control Systems, Not Just Better Models

The current AI conversation is shifting from model capability to operational control: agents need permissions, audit logs, kill switches, and domain-specific governance before they can safely act at scale.

Read article arrow_forward

calendar_today July 07, 2026

OpenClaw and the Real Test for Personal AI Agents

OpenClaw is interesting because it exposes the real product problem behind personal AI: agents need memory, tools, permissions, and safety boundaries before they can become useful daily software.

Read article arrow_forward

Follow the Work

Follow my GitHub and LinkedIn for new projects, technical notes, and local-first AI experiments.

GitHub LinkedIn