Dispatches

A fake citation in a product brief forced the switch: parallel pulls from PubMed, preprints, and trials, then tables with DOIs I can open before I repeat a number.

I learned the hard way that one-shot multi-phase Claude Code returns fast code and expensive cleanup; discuss → plan → execute → verify with atomic commits is what made autonomy survivable for me.

I stopped paying the preamble tax once Claude sat beside the doc or thread—short questions, faster triage; I still reach for Claude Code for files, commands, and APIs.

I filed a vague `/codex-task` on a Friday and spent Monday unwinding five conflicting guesses; the real bottleneck is always the same—context, legible scope, and explicit format—so a new teammate could ship without guessing.

I stopped burning every session opener on the same policy paragraph once `/remember`, `/reflect`, and the knowledge hub gave me one collaborator that already held decisions, guardrails, and corrections.

I keep automations in draft: Claude proposes, I send—parallel drafts for speed, human gate for tail risk after a near-miss taught me the cost of auto-send.

Every skill I write holds the constraint I burned time discovering — the CLI flag that silently fails, the join pattern that actually works, the config the docs forgot to mention.