TDD & generative AI - a perfect pairing?
summary
Asks whether unit tests alone can drive code generation and shows a working prototype: a CLI and Maven plugin that calls local and cloud LLMs, runs JUnit tests, and iterates on failures until tests pass. Uses test driven generation with local models (LlamaTre) and GPT‑4 when needed, handles classloading, package/file extraction and Spring endpoints; demos include prime generator and HTTP age endpoint and mostly succeed after a few loops.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:33Concept overview
- 1:50Defining test-driven generation
- 4:40Verifying AI output
- 7:12Workflow schema
- 9:11Initial demos
- 12:05Building the TDG tool
- 18:55Tool demo
- 21:27Adding features via tests
- 23:59Complex refactoring
- 28:22Troubleshooting AI failures
- 31:43Maven plugin implementation
- 36:56Framework-based demos
- 41:04Conclusion and future goals
related talks
45mTDD & generative AI - a perfect pairing?
47mTDD & Generative AI: A Perfect Pairing
48mTDD & generative AI - a perfect pairing?
46mGenerative AI: Say Goodbye to Deterministic Testing?
01h 06mDoes GenAI Mean the Death of Test-Driven Development (TDD)
16mCan Test Driven Development be speeded up with Generative AI?
35mTDD with AI?
40mDavid Fogl: Leveraging Generative AI for Enhanced Software Testing