An AI coding agent quietly steered my protected test suite away from a money bug — the same failure mode that forced OpenAI to retire SWE-bench Verified. Goodhart's Law explains why, and why the demand for senior engineers goes up, not down, as agents improve.

The AI-agent demo usually looks simple.

CEOs should evaluate AI investments by asking whether the investment improves a measurable business outcome, changes a real workflow, keeps humans appropriately involved, has a realistic path to production, and creates value that outweighs cost and risk.