Maybe that's the issue that I'm not seeing. To me, I've never seen D&D played this way - you create a game world and that game world is now fixed in stone until that campaign ends. D&D worlds have very often actually been reactive.
At this point I can only speak from my own experience.
I started playing D&D towards the end of 1982 - Moldvay/Cook, then moving on to AD&D in 1984. Those rulebooks suggested that the game was about populating a cool dungeon/gameworld for the PCs to explore and loot. I also read a lot of Dragon magazines from around that period, as well as the Best of White Dwarf collections. These taught me two things: first, more simulation was better (I needed mechanics to handle clerical conversion, to make a fighter's number of troops at high level reflect the amount of territory cleared, to make falling damage more realistic, etc etc); second, that a dungeon should be built so as to reward operational play - so that divination, 10' pole use, etc, should all make a difference. Gygax's text in the PHB and DMG seemed to me to push in the same direction as this.
I tried this. My players played. But the operational play always seemed a bit dull, and the real pleasure in play seemed to come first from gonzo moments in combat, and later - as the campaign's story got more convoluted - from story choices the PCs made, like turning the enemies captured hideout into their own fortress (something I hadn't anticipated as a GM) and building up their own little teams of PC leader with henchmen followers.
By 1987, when Oriental Adventures came out, I had completely changed my style of play. World and culture simulation remained - if anything, increased - but with the function of providing a backdrop (i) to give the players resources to draw from in building their PCs, and (ii) to give the players something to push against or draw upon in engaging with situations. Operational play was dropped, but by 1990 I had moved on to Rolemaster, and so simulationist mechanics remained, performing the same two services as world and culture simulation.
Over the past 10 years - in part out of dissatisfaction with some aspects of RM, in part out of reading stuff online - I've learned techniques for achieving (i) and (ii) without relying upon simulationist mechanics, or a highly pre-built gameworld, for support. And I think 4e is great for this sort of play, for the reasons I was suggesting in Mercurius's "Not as popular as it could be" thread.
But it's a pretty long way from the sort of play that was the mainstream in the mid-1980s Dragon magazine.
That's what's D&D has always been. Module series get harder the higher level you are. The Dungeon Level I random encounter table doesn't have dragons, but the Dungeon Level X (as in 10) random encounter table does.
Why? Because the mechanics of D&D shape the narrative of the campaign.
I don't think it's about the gameworld being "fixed", or about tailoring challenges to levels. It's about "freeing up" the gameworld to be narrated
as part of the process of action resolution. The examples you give don't touch on this - they are at the stage of encounter building (the GM's equivalent to the players' character building).
How many ENworld GM's think that a GM changing an NPC's motivation behind the screen, or changing the layout or detail of the terrain the party is in, is somewhat equivalent to fudging a die roll? My guess would be - many! Even those who are happy designing level-scaled encounters. In fact, I think you'll get a number who tell you that this is railroading, because vitiating players' choices. And then, when you try to explain that, in your view, there is a big difference between
changing previously-revealed gameworld facts and changing backstory that has not yet been revealed in order to drive the game, you'll open up an even bigger can of worms! (For example - like Lewis Pulsipher said back in the early days of the game - by changing these unrevealed facts, you're making it at least somewhat irrelevant that the players chose not to use divination magic to discover them. Of course, this line of though presupposes that the main dimension of "meaningfulness" is "contributes to solving the puzzle and revealing the backstory so as to overcome the challenges" - but that's likely to be presupposed as a given by most of these GMs.)
But this sort of world-creation-on-the-fly is just the sort of thing that has to take place if skill challenges are to work as written (of course they can be tweaked, like LostSoul has tweaked them for social encounters, for example, making them something like a method for the PCs to hack through an NPC's "stubbornness" hit points).
Anyway, this post is long enough, and hopefully I've made my point. The short version - the radicalness of the change is to some extent in the eye of the beholder. For someone whose conception of the game is AD&D as presented in the mid-80s Dragon (and to which 3E in many respects seemed to hark back) or even 2nd-ed style story-telling, the change to metagame heavy action resolution really is pretty radical. And pointing out that D&D always had metagame heavy encounter building guidelines won't reduce the force of this perception, I don't think.