Comment by bayesnet

2 hours ago

Building useful tests are probably one of the things that LLMs are the worst at tbh. You end up with tests checking `2 * 2 == 4` and similar.

0 comments