Hacker News .hnnew | past | comments | ask | show | jobs | submitlogin

It's not the vibe coding, it's the lack of tests


And I think human written tests at that. If the LLM is blind to the failure mode X, does it know to reliably write a test to evaluate the behavior of X?


wow. there are actually no tests here.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: