I think the MB framework is decent. The issue is that the test is noisy, is context dependent, unstable, and a middle value isn't well defined. But the tail end

I think the MB framework is decent. The issue is that the test is noisy, is context dependent, unstable, and a middle value isn't well defined. But the tail ends are very interesting, relevant ways to categorize personality. If you forget the test and just consider the ideas, it's useful