This video covers the full arc of the Ambient Library experiment in sequence — the same ground as the eight teaching blocks, but narrated end-to-end so you can see how each piece connects to the next.
Chapters — click to jump
- 0:00 Introduction
- 3:34 From session-bound AI to ambient intelligence
- 8:08 The startup hook problem — why CLAUDE.md wasn't enough
- 9:18 The ambient library needed a skill that designs skills
- 13:08 Model selection was the wrong altitude
- 15:11 The inference router — model and effort as a reusable skill
- 17:08 Isolated discovery for a growing global library
- 18:46 A forked writing team — worked example
- 21:33 Publishing the model-selection guide
The eight written teaching blocks are available in the Ambient Library session if you prefer to read at your own pace or skip to a specific topic.