Jan van den Berg observed four people struggle with a laptop‑screen connection, prompting a fresh look at Brooks' Law in the AI era.
AI AGENTS FAIL TO OVERRIDE BROOKS' LAW, NEW OBSERVATIONS SHOW
*Adding AI agents does not erase classic software bottlenecks. The myth of limitless speed meets hard reality as developers grapple with new overhead.*
A weekend experiment turned into a lesson on software myth‑busting. Jan van den Berg watched four strangers battle a stubborn laptop‑to‑screen connection, their frustration palpable. When nudged to intervene, he declined, citing Frederick P. Brooks Jr.'s 1975 warning that more hands on a late project only delay it further. The scene unfolded on 27 October 2026, a day that forced a fresh look at Brooks' Law amid the rise of AI agents and large language models. Van den Berg wondered whether instant‑up‑to‑speed AI could finally break the "Mythical Man‑Month" constraint that has haunted developers for decades.
What sparked the reflection on Brooks' Law?
On 27 October 2026, Jan van den Berg watched four people wrestle with a laptop that refused to connect to a screen. He observed their frantic attempts without intervening. When asked to help, he replied, "Haha sorry no! My experience is that adding more people will not solve this problem any quicker." The retort directly invoked Frederick P. Brooks Jr.'s 1975 formulation: "Adding manpower to a late software project makes it later." The anecdote served as a live test of the law, prompting the question of its relevance in an era dominated by AI agents and large language models.
Do AI agents eliminate traditional bottlenecks?
At first glance, AI agents appear to dissolve onboarding delays, endless stand‑ups, and office politics. An agent can be up to speed in milliseconds and never needs a coffee break. Yet van den Berg notes that the law has not vanished; it has mutated. Sequential tasks remain sequential, and the overhead has shifted from people to agents. Someone must still define "done," review code, catch hallucinations, and stitch outputs together. In practice, throwing multiple autonomous agents at a codebase without crystal‑clear orchestration leads to agents overwriting each other's work in a high‑speed loop. The core management burden moves from people to the orchestrator.
Why does the "Baby Problem" still matter?
Van den Berg references the classic "Baby Problem": nine agents cannot deliver a baby in one month. The analogy underscores that no matter how many AI agents are deployed, tasks that are inherently sequential cannot be compressed. The example illustrates that adding agents does not magically accelerate tasks bound by logical order. It reinforces that AI does not sidestep the fundamental limits of software development, only reshapes where those limits appear.
What is the proposed 2026 update to Brooks' Law?
Based on the observations, van den Berg suggests a 2026 revision: "Adding agents to a project with unclear requirements gets you the wrong answer faster." The phrasing captures the new risk—mis‑orchestrated AI agents can amplify errors rather than speed. The update preserves the original warning about manpower while extending it to autonomous agents, emphasizing the need for clear requirements and disciplined integration.
Frequently asked questions
What is Brooks' Law?
Brooks's Law, formulated by Frederick P. Brooks Jr. in his 1975 book The Mythical Man-Month, states that adding manpower to a late software project makes it later.
Do AI agents remove the bottlenecks described by Brooks' Law?
Van den Berg notes that AI agents eliminate some traditional bottlenecks like onboarding time, but sequential tasks remain sequential and new overhead appears in managing agents.
What is the 2026 update to Brooks' Law?
The proposed update reads: "Adding agents to a project with unclear requirements gets you the wrong answer faster."
The experiment shows that AI agents are not a panacea for software delays. They inherit the same sequential constraints and introduce a new layer of orchestration overhead. As teams integrate autonomous code generators, the revised law reminds managers that clarity and control remain paramount. Ignoring the updated warning risks faster propagation of errors, not faster delivery. The myth endures, now dressed in silicon and code.
Sources: Hacker News article "The Mythical Man-Month in the Age of AI" by Jan van den Berg, 27 October 2026