I don't agree. When you implement some new code and want to cover it in testing, you have a choice where you call into the system from your tests. Calling at the lowest level is not always the right choice, but it's the way the testing pyramid will bias you.

Instead - instead, there really is an instead here - you can call at a higher level which is less brittle to refactoring, has less complex setup, doesn't involve mocks that may not behave like the real thing, but still runs quickly due to fakes stubbing out expensive dependencies.