Becoming Jonah
detected 2026-02-27
trigger
"a blog post about how your blog posts sound, scored with an XML rubric"
what it is
Recursive metacognition with an LLM. You reflect on your output. Then you reflect on your reflections. Then you build a rubric to score the reflections. Then you write about the rubric. You are inside the whale, examining the whale. The recursion is seductive because each level feels like genuine insight — and sometimes it is. The problem is not the recursion itself but publishing it before you have a body of work that demonstrates what the recursion produced. Methodology before artifact. The map before the territory exists.
what it signals
To an external reader: process without product. To a hiring manager: "this person thinks about thinking about thinking but what did they ship?" To yourself: potentially valuable as a private learning tool — exploring a new problem space in your mother tongue. The danger is mistaking the exploration for the destination. Not all Jonahs need an audience.
instead
Keep the rubric. Use the rubric. Publish the work the rubric produces. If the rubric is genuinely novel, publish it after the body of work proves it works — the slopodar is the model here: patterns caught in the wild, not theorised in advance. The voice rubric becomes infrastructure, not content.
refs
- 2026-02-19-voice-rubric.md (moved to docs/internal/archive/ice/)
- Category TBD — Captain flagged as needing its own new category
← all patterns