I think there’s some confusion about the typical zone definitions. I haven’t read the article, but in most five-zone frameworks, “under AeT” means zones 1 and 2 while “over AnT” means zones 4 and 5. In that context, it would be zone 3, not zone 2, that is to be avoided minimized.
The reality is that zone 3 is useful, but too often over-emphasized and over-used (because it’s gratifying and most people confuse effort and gratification with progress.)
An effective approach would be to use a largely polarized model early in a macrocycle and then include the appropriate amount of zone 3 close to a goal event. As suggested, “appropriate” is often over-estimated.