Blog

What it actually takes to run AI agents in production — infrastructure patterns, security decisions, and lessons from someone who did the same for cloud a decade ago.

1. Passes Everything, Works for Nothing

   An AI agent built me an iOS markdown editor with 100 features. It crashed every time I opened it. I deleted the repo and started over. Then I changed how my tooling works so it can't happen again.

   July 18, 2026·5 min readai-developmentcode-qualityaishorelessons-learned

2. Personality Is Three Paragraphs

   I built a 17-dimension persona schema for AI agents — six psychological frameworks, seven archetypes. Then I scrapped it. For work agents, capabilities are the product. Persona is a settings page.

   July 18, 2026·3 min readpersonaagent-architectureproduct-designcontrarian

3. The cPanel Moment

   AI agent infrastructure isn't in its Terraform era. It's in its cPanel era — the management layer hasn't separated from the runtime yet. Cloud veterans keep pattern-matching to the wrong phase.

   July 18, 2026·3 min readthesiscloud-historyagent-infrastructuremental-model

4. Why My Trading Agent Can't Trade

   I built an AI agent that watches the market all day, briefs me, and proposes trades against a locked plan. It has never placed an order — and it structurally cannot. Here's why that constraint is the whole point.

   July 18, 2026·5 min readai-agentstradingagent-designsovereignty

5. Publish Your Personal Website on GitHub Pages, for Free, Using Claude Code

   A step-by-step walkthrough from a fresh Mac to a live website at yourusername.github.io. One page of plain-English spec; Claude Code handles the build, the repo, the deploy, and the optional custom domain. About 45 minutes including the one-time install.

   May 16, 2026·12 min readtutorialclaude-codegithub-pagesstatic-sitegetting-started