How to Optimize Non-Coding Time
summary
Focuses on cutting the 10 hours a day spent on non-coding work by automating chores and reshaping the PR cycle: correlates GitHub to Jira to create tickets automatically, uses an ML review-time estimator so reviewers pick PRs that fit available slots, and runs a bot that tracks where merged code lives in the pipeline. Uses data from ~25k developers showing median PRs are idle half the time and top third idle ~80%, pushes for same-day reviews and small PRs (100–200 LOC) to slash idle and context-switch costs.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 4:47Developer productivity paradox
- 8:01Pull request cycle impact
- 13:38Queuing theory in PR reviews
- 23:23Research on PR idle time
- 32:29Optimizing PR workflow
- 41:16Lunar.dev’s solution
- 46:34Q&A and closing
related talks
38mHow to Optimize Non-Coding Time
14m50 shades of PR dark matter
42m7 Metrics to Accelerate Delivery without Damaging Culture
29mBreaking the PR Review Bottleneck
31mDevOps 2.0 - Bigger, Badder with More Automation
06m50 Haunting Shades of PR Dark Matter
17mThe Perfect Code Review
57mAsync Code Reviews Are Choking Your Company’s Throughput