Evaluate useful outcomes before changing behavior. Justify delegation, rehearse restoration and inspect the evidence from Lucy’s unattended shop day, then budget what the model sees.
Read and practice
- Chapter 16: Evaluation as measurement: error bars and paired comparisons (draft chapter, two practice units)
- Chapter 17: Optimizing against an evaluation: the winner's curse and preferences (draft chapter, two practice units)
- Chapter 18: When a second agent pays: parallelism, errors and bounded delegation (draft chapter, two practice units)
- Chapter 19: What a model call costs, and a deployment that survives (draft chapter, two practice units)
- Chapter 20: A whole day: reliability, honest reports and readiness (draft chapter, two practice units)
- Chapter 21: Context engineering: a budget for what the model sees (draft chapter, two practice units)
