Systems Software in the Large

The "update" project at Oxide Computer Company requires managing hundreds of complex, interdependent system components while maintaining autonomous operation. To navigate this, the team employs a "plan-execute" pattern that separates state planning from execution, enhancing testability and providing critical escape hatches during failures. Managing project scope involves distinguishing between "important blockers" and "important non-blockers," a rubric that prevents analysis paralysis while ensuring long-term technical debt is addressed. Maintaining focus in a remote environment relies on "water cooler" meetings and project planning via incremental demos, which force prioritization and provide proof points. By using a date-driven approach rather than traditional project estimation, the team effectively cuts scope to meet milestones, ensuring that the system remains robust without succumbing to the pitfalls of second-system syndrome or over-engineering.