華文

Chapter 4: Responsiveness in Adaptation

Having a ‘competent’ system is not enough, systems need to be able to respond to lessons and to adapt. That’s what makes a good “gardener” adjust to “the speed of a garden”. Responsiveness stands for ‘care-receiving’ - a system hears back, corrects fast and learns in public. And people can easily participate in the process, it is their lived experience that matters.

Quick version

Results we want

Why Responsiveness?

Competent action creates new information. Refusing to hear it is the fastest path to failure.

A simple picture: A clinic posts hours; patients show up to a locked door. A responsive clinic apologizes, posts why it happened, updates hours, and texts people next time. The fix becomes part of the system.

Simple ideas behind this chapter

What good responsiveness looks like

From ideas to everyday practice (step by step)

  1. Expose the “appeal” button. Everywhere a decision is shown, a one‑click appeal exists with a clock.
  2. Accept “harm drafts.” Let people submit proposed evals in plain words; convert to tests with partners.
  3. Triaging & severity. POs classify S0–S3; S0/S1 trigger immediate pause or reversible defaults.
  4. Fix or explain. On the clock, publish the remedy or the reason with next steps.
  5. Memorialize. Turn the incident into a test; add to the eval registry; link from the contract changelog.
  6. Check back. Close the loop with those who appealed; measure trust‑under‑loss.

Plain tools (buildable today)

Flood‑bot story (Part IV: repair in motion)

What could go wrong (and quick fixes)

How we keep ourselves honest (what we measure)

Interfaces with other packs

A closing image

Imagine a workshop with a wall full of retired “broken parts,” each tagged with the story of how the parts came to break, how to avoid and fix future damage.

Previous Next