At that point, though, the defense becomes little more than "why make anything in a dungeon interesting (aside from perhaps its aesthetic flow enough for people to endure its grind with a modicum of sanity) when inevitably, no one's going to give a damn about what they're doing anyways?" (E.g. Least development costs for the most time gating = profit.)
These different groups of people exist either way, and will have an about an equally proportionate effect whether you give them a longer, more complex dungeon with increased rewards or a shorter, braindead one with fewer rewards. The only exception to this are those who actively refuse to learn how to play, or play at all well, and actually facing those issues can only be a step forward for the game, even if it's just by a different path of avoidance (e.g. expert roulette as a rule variant on any of various dungeons that makes them much harder, but the rewards that much better, rather than being limited always to only the 2 most recently released) that ends up creating more options for the playerbase, whatever its composition.
Efficiency is going to be rewards/time. How could you possibly hurt efficiency when the rewards are at least proportionately increased? If anything, the gap would only increase further in the favor of those who play well/efficiently as you add on additional challenges. Are we talking about <reward/time*(level of attention required)> here? That's the only way I can imagine there being some restriction against "efficiency". And even then, there's nothing to say that every dungeon needs to be equal difficulty and reward even within its own tier; you could easily have one that's more "efficient" (in terms of rewards divided by cumulative level of attention required over the run) to an unskilled player and another that's more "efficient" in the same sense to a skilled player.