Teach Your Agents to Manage Up
The post shows how to train AI agents—like the OpenClaw/Hermes “Claw”—to manage up by treating them as outcome‑focused employees rather than generic tools. The author shares a six‑step prompt that forces the agent to repeat back the task, outline steps, break work into tiny actions, present recommendations with alternatives, track feedback, and adapt over time. Applying this framework reduces wasted effort, improves decision quality, and creates a feedback loop that teaches the agent a manager’s preferences. The approach mirrors best‑practice human‑to‑manager communication but can be explicitly programmed for AI.
Motherhood & Ambition, 6 Months In
The author reflects on six months of motherhood after a difficult IVF journey, describing how the birth of her son removed a protective layer and intensified her sense of purpose. Contrary to expectations that parenting would shrink her professional world,...