The cost of saying yes has changed

Summary

Generating initial code patches has become significantly cheaper with AI agents, shifting the primary expense of small feature requests from writing code to debating scope in meetings. Teams can use agent-generated patches as diagnostic probes rather than final deliverables, turning abstract scope debates into concrete artifacts that reveal true system touchpoints and risks. However, low generation costs do not translate to low ownership costs, as changes touching areas like authorization, compliance, or public contracts still demand significant human review. Constrained attempts allow engineers to price uncertainty quickly and shift scope discipline from pre-implementation speculation to evidence-based code review.

Takeaways

1. 1

   An AI-generated patch acts as a price check to inspect diff sprawl and testing resistance rather than serving as an automatic deliverable.

2. 2

   A software change is only genuinely cheap if a human engineer can confidently review, validate, and own the resulting behavior long-term.

3. 3

   Changes affecting authorization, data-retention semantics, billing, privacy, or public contracts require rigorous evaluation regardless of how clean the generated diff appears.