Does auto generating tests even do anything helpful? Don’t they just sort of tautologically say the code does what it does at best or do something completely ridiculous like test and implementation that only exists in the test file at worst?
We have an extensive description of _how_ tests should be written and they're reviewed by a human. All the AI does is fill in the boring middle part.
We have an extensive description of _how_ tests should be written and they're reviewed by a human. All the AI does is fill in the boring middle part.