If you thought the speed of writing code was your problem, you have bigger problems
summary
Claim: coding speed isn't the real bottleneck; IDC finds engineers spend 16% time in editor so writing is rarely the limiter. AI agents like Claude Codex and Pi speed code but GitHub/GitClear/DORA data show code up 55% while delivery and stability fall and code churn doubles. Weekend projects (automated chicken coop, egg counters) get built fast with agents, and a team used AI to add Cantonese across 100 repos in 2 weeks — the org bottlenecks and review/CI issues still stop value.
AI-written summary · about
chapters (13)
- 0:00Welcome and introduction
- 3:18Personal coding journey and AI impact
- 10:20Theory of constraints in software
- 16:28Why coding isn't the real bottleneck
- 21:29AI coding stats and unintended consequences
- 24:42Bottlenecks beyond coding: discovery
- 28:31Review and CI/CD as bottlenecks
- 33:01Lean principles and waste in development
- 36:33Breaking silos with AI coding
- 39:47Measuring productivity beyond code output
- 41:56Actionable steps to find and fix bottlenecks
- 47:13Using AI to improve review and CI/CD
- 48:58Summary and closing remarks
