A problem with most everything I read about testing and the relative benefits or drawbacks is that it’s all anecdotal (as you’ve noted).
Furthermore, one style (such as “write mostly integration tests” or “do TDD”) may work well for some projects or teams, but not all. It’s of utmost importance teams understand why they write tests how they write tests!
In other words, if your team decides to scrap most of their work on unit or functional testing based on a tweet — you’re doing it wrong.
Thanks for explaining this viewpoint. Others will undoubtedly agree or disagree with you here, because they have different anecdotes supporting a different strategy. 😖
There is no One True Way.