tl;dr: Brooks's Law ("adding manpower to a late project makes it later") still applies in the age of AI agents, just in mutated form: sequential tasks remain sequential, and coordination overhead has shifted from managing people to managing agents, defining "done," and catching hallucinations. Unleashing multiple autonomous agents on a codebase without clear orchestration leads them to overwrite each other's work. The author's proposed 2026 update: "Adding agents to a project with unclear requirements gets you the wrong answer faster."