The author argues that team speed is a property of the system rather than the individual performance of the strongest developers. In large projects, delays often occur between the initial idea for a change and its actual deployment, rather than in the writing of the code itself.
To reduce the costs of making changes, the author suggests implementing a modular architecture. This approach allows developers to launch large sections of a project separately, focusing only on the required functionality without waiting for a full product build.
Running modules in isolation enables faster verification of hypotheses and allows new developers to contribute to specific sections without needing to understand the entire project architecture first.
