Blog
Posts about automation, generative AI, agentic systems, and lessons learned building projects.
A Browser Per Claude Session
Giving Claude Code one browser was the easy part. Running many parallel sessions meant giving each its own authenticated browser I can watch and take over from my phone.
I Hit Claude Code's Limits. So I Gave It Workers.
Running multiple projects in parallel pushed me past Claude's usage limits. So I built a system where Claude plans and reviews while Codex and Gemini implement.
The Config That Makes Claude Code Feel Like a Teammate
Installing Claude Code gives you a capable agent. The MCP servers, skills, auto-approve rules, and hooks are what turn it into something that works like a teammate.
Building From My Phone, Not My Laptop
I used to haul my laptop everywhere to work on side projects in spare moments. Now a cheap home server and my iPhone do it, no laptop required.
How I Automated My Yard's Drip Irrigation
Friends kept asking for my irrigation setup, so here it is: the drip system that waters my whole yard, and the parts I used.
Giving Claude Code Eyes with Playwright MCP
Claude Code in WSL can't see your browser. Here's how I used Playwright MCP to close that gap and build a visual testing workflow around it.
From Excel Macros to Building with Agentic AI
How my relationship with automation evolved over twelve years in paid media, and what each wave taught me about the one that followed.
From $60/Month to Free Website Hosting
From Wix to WordPress to a VPS to Cloudflare Pages: how I ended up hosting three websites for nothing but the cost of domain names.