Do you let LLMs write both your tests and your code?
Summary
The article debates using LLMs to generate both tests and code, highlighting risks of tautology and overengineering. It considers governance approaches and suggests rules to prevent LLMs from generating both aspects in a single MR.