Code review for Android is broken

on YouTube ↗36m

" summary

Claims that CodeRabbit cuts review time by 30–50% review time reduction by running a sandboxed clone of the repo, executing 40+ static and security tools, checking CVEs and dependency issues, and surfacing real fixes (e.g., captive‑portal transport checks, broken tests, refactoring suggestions) inside GitHub/GitLab/Bitbucket. A field test compares human‑only vs human+AI on the same messy Kotlin PR and reports faster PR turnaround, higher defect catch across correctness, performance, security and release safety, plus a copyable CI gate and ready‑to‑apply fixes.

AI-written summary · about

" chapters (9)

1. 0:00Introduction to broken code reviews
2. 3:52Daily developer life with code reviews
3. 7:29Why reviews feel broken today
4. 11:30AI's role in fixing reviews
5. 16:52How CodeRabbit works
6. 21:43Field test with AI reviews
7. 28:38AI review capabilities today
8. 31:29Why AI reviews matter for efficiency
9. 35:22Summary and call to action

" related talks

• 01h 02mEffective Code ReviewDave Liddament · PHP UK Conference 2023
• 45mCracking the Code ReviewPaco van Beckhoven · Voxxed Days Ticino 2025
• 47mGoldilocks and the Three Code ReviewsVaidehi Joshi · Red Dot Ruby Conference 2017
• 43mCode Review ToolsGautier Hayoun · EuroPython 2015
• 45mCracking the Code ReviewPaco van Beckhoven · Voxxed Days CERN 2025
• 46mGetting the most out of code reviewBrian Plummer · Android Makers 2019
• 47mGoldilocks and the Three Code ReviewsVaidehi Joshi · RedDotRubyConf 2017
• 39mSurprise! Neighboring Code ReviewsOkumura, Sato Taichi · DroidKaigi 2023