I’ve been spending way too much time lately writing boilerplate mocks and stubs for my unit tests. It feels like 40% of my dev time is just "typing for the sake of coverage."
I’ve started experimenting with unit ai tools to automate this, and honestly, the shift is pretty radical. Instead of writing the test from scratch, I’m now mostly just reviewing what the AI generates and tweaking the edge cases. It feels like the role of a developer is shifting more toward a "Test Architect" rather than a "Test Writer."
I found this guide that really helped me structure my workflow and avoid some common pitfalls (like hallucinated mocks): testomat.io/blog/ai-unit-testing-a-detailed-guide
Curious to hear your thoughts:
Do you trust AI-generated unit tests for mission-critical code?
How do you handle the "Human-in-the-loop" review process without it taking as long as writing the test itself?
No responses yet.