I think there's plenty of room for setting design that is focused on building elements for players to interact with that may not see play if players do not pursue it or to provide dynamism in the setting that has knock on effects players may discover if they play skillfully. A setting element does not have to be known to be motivated by the needs of play.
There seem to be a few different things being canvassed in these posts:there seems to be an underlying assumption that the players will never experience the majority of what is designed unless it is specifically prepared at the moment, in collaboration, or for the next 2-3 sessions. Which I disagree with strongly. PCs can interact with situations and NPCs that are informed by circumstances that players may never discover.
* Setting elements that are prepared by the GM in advance may end up not being known to the players, simply because the players declare actions for their PCs that leads to different setting elements becoming the focus of play. These setting elements thus end up not mattering to play, but they might have - it wasn't and perhaps couldn't have been known in advance that the players would not declare actions that made the ignored setting elements salient. The paradigm example in D&D play would be a dungeon room that the players end up simply never having their PCs explore.
* Setting elements may be prepared by the GM in advance with the expectation that it is likely they will end up not being known to the players, because it would take skilled play on the part of the players for them to declare actions for their PCs that oblige the GM to reveal the setting element in question. The paradigm in D&D play would be a secret door concealing a valuable treasure. These setting elements matter to play even if the players never learn about them, because they create the scope for skilled play.
* Setting elements that are prepared by the GM in advance, and that are not known to the players, may be used by the GM to make decisions about setting elements that are known to the players. (DW has a version of this, in the form of "fronts") - in that way they "mattered" to play.
* Setting elements may be prepared by the GM in advance with the expectation that it is likely they will end up not being known to the players, because it would take skilled play on the part of the players for them to declare actions for their PCs that oblige the GM to reveal the setting element in question. The paradigm in D&D play would be a secret door concealing a valuable treasure. These setting elements matter to play even if the players never learn about them, because they create the scope for skilled play.
* Setting elements that are prepared by the GM in advance, and that are not known to the players, may be used by the GM to make decisions about setting elements that are known to the players. (DW has a version of this, in the form of "fronts") - in that way they "mattered" to play.
The second and the third of these can combine (as I think @Campbell is suggesting): an element of the third type might also be an element of the second type, and hence something the players can learn about through skilled play. A secret treasure room probably isn't a good example of this; but a secret villain might be.
This is quite relevant to the OP, and to similar sorts of issues that can arise in a type of fairly traditional, fairly mainstream RPGing. If a setting element is doing a lot of work in my third category (eg the GM is imagining very active secret villains) and is very opaque as an instance of my second category (eg the action declarations the players would need to make to learn about it are less "skilled" and more somewhat lucky or arbitrary - this can happen, for instance, if the GM is changing the fictional parameters around the setting element a lot) then the players can lose their grip on the fiction, and essentially become subject to GM dictation of what happens next.
This is something on which I think it should be possible to give quite detailed and effective GM advice, but I don't recall ever having come across such advice myself. (Maybe it's somewhere on the web, and I'm mostly thinking of rulebooks.)