I once worked at a company with 1000 UI tests. Every PR tests would fail because certain backend failures would cause flaky behavior. Then, we started mocking r

I once worked at a company with 1000 UI tests. Every PR tests would fail because certain backend failures would cause flaky behavior. Then, we started mocking response, and of course they no longer failed. They also no longer caught real bugs anymore. Does that give you some nuance?