The Mythical Man-Month in the Age of AI

Last weekend, I witnessed four individuals struggling to connect a laptop to a screen. Their efforts, however, piqued my interest. Upon being asked to assist, I responded with a chuckle, stating that adding more people would not expedite the task. This observation inadvertently alluded to Brooks' Law, a concept introduced by computer scientist Frederick P. Brooks Jr. in his 1975 book, The Mythical Man-Month. The law posited that adding manpower to a late software project only serves to prolong its completion.

As I contemplated the relevance of Brooks' Law in the age of AI agents and Large Language Models (LLMs), I found myself questioning whether it remained applicable. At first glance, it seemed that the traditional bottlenecks – onboarding time, endless stand-ups, and office politics – had vanished. An AI agent, after all, can grasp a concept in milliseconds and does not require a coffee break. Yet, upon closer examination, the law had not vanished entirely; rather, it had evolved.

The infamous "Baby Problem" still held true: nine agents could not deliver a baby within a month. Sequential tasks remained sequential, and overhead had not disappeared; it had merely shifted. Instead of managing people, one now found themselves managing agents. In the absence of a clear definition of "done," someone had to review the code, identify hallucinations, and ensure that the pieces fit together.

Assigning nine autonomous agents to a single codebase without precise orchestration resulted in each agent inadvertently rewriting others' work in a rapid loop.

To address this modern iteration, I posited that the 2026 update to Brooks's Law should read: "Adding agents to a project with unclear requirements gets you the wrong answer faster." In essence, the law had transformed, but its underlying principle – that the addition of more components, whether human or artificial, without proper coordination and clear objectives – could exacerbate problems rather than resolve them.