Data Before Intelligence
You will want to start with the impressive part: the AI. The order is backwards. The intelligence is only ever as good as the boring, decomposed data underneath it.
Notes
Shorter notes: things I’m learning, figuring out, or just find interesting. Less formal than the writing; published more often.
You will want to start with the impressive part: the AI. The order is backwards. The intelligence is only ever as good as the boring, decomposed data underneath it.
The stack and the reasoning behind this site: static Astro, Markdown-first publishing, and one deliberate constraint that makes writing the only thing that takes effort.
How I work with the agent, day to day. For the framing, the principles, and the longer essay, start at the practice thread.
A skill written for your own agent teaches one assistant. Committed into a public repo, it teaches every agent that opens the folder. Two skills from the sandbox run just made that jump.
An agent gets better not from cleverer prompts but from corrections that stick: turn every fix into configuration (memories, rules, small skills) so the same mistake never comes back.
I build most of this alone, but not really alone. What working with an AI coding agent actually changes, where it carries real weight, and where you still have to be the one who understands.