TDD with AI?
summary
Shows using Gemini in Android Studio to generate tests, fixtures and a repository implementation, claiming obvious wins for simple tasks and documentation but frequent failures once tests grow or refactorings span collaborators. Live demos produce IDE crashes and flaky generations, incremental TDD “baby steps” don’t work reliably, and generated refactors sometimes introduce subtle, nonobvious bugs. Recommends generating fixtures and docs for speed, warns that it fails on incremental baby steps and that refactorings produce subtle bugs, so avoid for complex tests and junior-only guidance.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 3:00TDD and its importance
- 9:20TDD cycle benefits
- 10:03Why choose Gemini
- 11:50Live coding with TDD and AI
- 27:23AI limitations in TDD
- 31:54AI-assisted coding research
- 33:17Takeaways and recommendations
- 35:04Q&A
related talks
45mTDD with AI?
47mTDD & Generative AI: A Perfect Pairing
46mTDD & generative AI - a perfect pairing?
43mTDD: Test-Driven Development vs TAB-Driven Development
39mPractical TDD for Android
26mAI and the Future of Testing
48mTDD & generative AI - a perfect pairing?
01h 06mDoes GenAI Mean the Death of Test-Driven Development (TDD)