For nearly 50 years, software engineering operated under a simple, painful truth: throwing more people at a late project just makes it later. Fred Brooks coined that idea in 1975, and it became gospel. Now a new generation of AI tools is making the whole framework look quaint.
Companies like OpenAI, Anthropic, and the AI-powered code editor Cursor are demonstrating that smaller teams, armed with the right models, can produce output that once required headcounts several times larger. The traditional scaling bottleneck wasn’t computing power. It was communication overhead between humans. AI doesn’t have that problem.
Brooks’s Law meets its match
Here’s the quick version of Brooks’s Law: every new person you add to a software team creates new communication channels. Three developers means three pairwise connections. Ten developers means 45. The math gets ugly fast, and so does the coordination tax.
AI coding assistants sidestep this entirely. They function as non-human contributors that don’t need to attend standup meetings, don’t misunderstand Slack messages, and don’t take PTO. In English: the coordination overhead that made large teams inefficient simply doesn’t apply to an AI pair programmer.
