This document discusses how to address frequent build failures in continuous integration (CI). It notes that CI requires continuous integration of code, integration testing, and a stimulus-response to build failures. However, many organizations blame individuals for failures rather than fixing issues quickly. This can negatively impact commit frequency, risk-taking, and motivation. Build failures are not always in a developer's full control, similar to merging traffic. The document recommends focusing on fixing builds quickly rather than blame, tracking red build time and unavoidable failures, and creating an atmosphere of trust rather than blame.