Since the LLM needs more context around mocks for the unit test, it is better in generating test cases above existing test cases. Not good for generating from scratch for complex applications.
how do y'all feel about AI-generated tests? I gave it a shot with ChatGPT on a side project and it saved me a ton of hassle. Like, I wouldn’t have written those tests myself. Trust-wise, I was still a bit iffy on refactoring, but hey, better than no tests at all. Overall, seems like a win for low-risk stuff.
Also nice to see more Go based LLM apps