He's not saying it's difficult to get 100% coverage. He's saying that 100% is a bit suss because you probably wouldn't achieve 100% coverage unless that is your goal (even if Python does make it easy) and that's the wrong goal.
It's a real goal for dynamic languages. You have no other option to be sure your code is not broken. Another option is to use 100%-non-any type hints. It's a way more harder.
I can see it's a non goal then you have access to deployed code and Sentry. But as library author or author of customer apps there is no other way around.