Notes on software engineering, shipping products, and building fast, maintainable web experiences.
Chrome's new Agentic Browsing category audits how well a site works when the visitor is an AI agent, not a human — WebMCP tool registration, agent-centric accessibility, and layout stability. Here's what it checks, and how it maps o…
AI-driven multi-agent systems (MAS) are architectures where multiple autonomous agents collaborate, compete, or coexist within shared environments to solve complex tasks beyond the capacity of a single agent
OpenClaw can run tasks across chat apps, browser automation, and your filesystem. Here’s a security-first guide to using local-first AI agents without turning your laptop (or VPS) into an attacker’s playground.
AI raises the baseline for speed—but also for correctness, security, and ownership. Here’s a practical playbook for staying valuable and shipping high-impact work with AI in the loop.
Tokens cost money and time. This guide shows practical ways to shrink prompts, tighten outputs, and use TOON-style structure to cut back-and-forth.
Agent skills are reusable, scoped behaviors that make assistants consistent and safe. This guide covers what they are, how to design them, and how to ship them in teams.