Model, component, workflow, trajectory, and system
The right boundary contains the mechanism you want to attribute: a trajectory is evidence of one execution, while the outcome confirms whether the change matters to the system.
Links containing ?t= open the video at a specific second.
Video summary
The ideas to retain
01
Start by defining the object, not the metric
A useful abstraction for an AI system is:
02
Level 1 — Evaluate the model
A model eval tries to measure model behavior under a controlled harness.