I use Bloom’s Taxonomy unapologetically when I write or speak about flipped learning, since the sequencing of learning activities is critically important in flipped environments. Two things I always try to communicate first before pulling out Bloom:

1. It’s called Bloom’s Taxonomy and that’s all it’s really meant to be — a taxonomy, or a naming convention for certain kinds of knowledge. The fact that it’s arranged in a pyramid shape implies a kind of partial ordering on the names but this is just a suggestion, not an empirical fact. In particular, there is no implied order of importance on any of these levels except you could argue that the bottom levels are in some sense more important than the upper ones, as you’ve mentioned, because like in a real pyramid the floor underneath supports the floor above.
2. Progression through Bloom’s Taxonomy is almost always nonlinear, unlike what the “floors of a pyramid” shape might suggest. We don’t always start at the bottom and go up one level at a time. I wouldn’t suggest that it’s possible to fully master knowledge at one level of Bloom without mastering knowledge at all previous levels; but I would definitely say that our mastery can progress at one level without it being fully realized at a previous one. I may not understand everything about the application of a calculus concept, for instance but that doesn’t stop me from making real progress in organizing patterns and seeing trends. Constantly jumping levels of Bloom is just a facet of our ambitions as learners, trying to sound out our zones of proximal development, and it’s kind of fun to do.