And does so by being larger, with the only real difference being more area (and gates, and often then critical path length and thus speed) being spent on getting the signaling to the level where it can do word instead of page addressing. The actual flash cells themselves are functionally the same.

There's no fundamental difference in gate technology between the two, so a flash that is addressable to a finder granularity will always be larger than the coarser equivalent. That's the trade off.