Not quite mate, as you know storage is cheap but compute is not. Without proper data modelling you’re leaving it up to the user to execute a query that could have been solved once and upfront in the model itself. I would of read the rest of your blog but due to this failed reasoning the rest of the blog falls under a false assumption that simply misses the point of dimensional modelling altogether.

Popcorn analytics may be ok with OBT, but scalable analytics depends on well thought out data modelling, Kimball got right decades ago because we’re still using relational tables to solve business questions.

Learn from the master before you dismiss his lessons as obsolete.