I know I'm a geek - but there's a certain degree of satisfaction in having a test made to fail after a refactoring actually fail as expected and then implementing the fix as expected.
And to imagine some people want to offload that satisfaction to a LLM tsk tsk tsk.



