Strand · RUN

The Runbook

3 essays, August 2026 to September 2026. They test Instruments Over Theory (1). 70 claims across 1 of them are checked against a primary source in the public ledger.

Every essay in this strand

  1. Use AI to Learn Without Getting Worse at It Answer-giving AI left learners worse off once it was gone. Three setups built to keep the thinking with you, and a way to check what stuck. 70 sourced claims
  2. The Skill That Never Fired A Claude skill can fail two ways: its instructions are wrong, or Claude never picks it. Anthropic's own tooling scores whether one skill fires. This tests it against the neighbour that can quietly win its requests, with a small routing eval you run yourself.
  3. Law IV Stop Re-Priming Claude Code by Hand The context you paste at the start of every session, put into one file you invoke with /prime.

All strands All the tools The claim ledger