Drawing on his experience at Imprint during a period of rapid growth, and on the way AI tooling has accelerated engineering work, Will Larson revisits his rules of engineering leadership for this new era.
His central argument is that AI changes the cost of many things, but not the fundamental bottlenecks. Migrations that once required teams can now be done by one or two engineers, yet quality judgment matters more than ever, since “even small sharp edges will break your colleagues’ mental models.” First-pass code is cheap, but truly working code still depends on the development harness of tests, CI/CD, and validation environments. And he urges leaders to optimize processes for agents in the base case, reserving human attention for higher-altitude planning and edge cases.
Throughout, Will emphasizes that durable, domain-context-rich teams matter even more in an AI-accelerated world, and that fast, durable decision-making is a prerequisite for capturing AI’s benefits at all. The constraints that limit progress, he notes, remain “organizational misalignment, lack of clarity, and poor technical architecture,” AI just makes everything else faster.
You can read the full post on his blog: Revised Rules of Engineering Leadership.