AI-assisted development journal
Remember Dick and Jane? “See Dick run. See Jane run. Run, Dick, run!” Those simple sentences taught millions of children to read. We’ve borrowed the formula for our orchestration layer, except our protagonist is an AI agent named Ralph, and instead of running through yards he’s running through codebases.
The ralph harness has been running effectively for weeks now. It tracks progress, summarizes history, and can grind through complex goals for 30 hours when needed (though most finish in a single iteration). We have a solid foundation for autonomous development.
Sub-agents are working. This has been on the roadmap since the beginning, and last night it finally came together. Agents can now spawn other agents.
I’ve been actively working with AI coding agents for months. I thought I was ahead of the curve. Then I went back and re-read something Steve Yegge wrote in March 2025, almost a year ago, and realized I’m still catching up.
This post is emerging thinking, not a finished design. Something clicked recently and I want to capture it while it’s fresh.
Running multiple AI agents in parallel requires a workflow that keeps them from stepping on each other. Here’s what we’ve settled on for Ikigai development.
Ikigai is getting close to usable for real tasks. Tonight I decided to test that by copying my prompts and skills from Claude Code to Ikigai, then using it to review plan files that Opus was writing.
Ikigai rel-10 is out, and we recorded a video walkthrough to go with it.
Dead code removal is one of those problems that looks simple until you try to automate it.
When you’re running multiple agents in parallel, the way you plan work changes. You stop thinking about what to do next and start thinking about what can be isolated.