No harness, no goal: spending LLM tokens just to be surprised
I set up a daily automation that reads my own notes and surprises me each morning with one small, finished artifact I did not see coming. This post explains how it works, shows a few it has made, and shares the exact prompt so you can run your own.