I mean, is it really wrong though? When these settings were created, they were created with the rules of the game in mind, and they wanted the setting to reflect those rules. Hence why we get things like cataclysms, times of troubles, spellplagues, sunderings, and whatever else, radically changing the settings around because of a rules update.
I mean, let's look at Dark Sun with the 2e rules.
*While survival spells existed, they competed with other spells people wanted to use so they weren't used as often (or so I've been told in threads on this forum where people gripe about goodberry et. al, and I point out, "but we had those spells in AD&D!"). But TSR staff, unlike WotC staff, had no problems whatsoever printing several pages of "spell changes" to their settings in boxed sets (see also Spelljammer, Ravenloft, and Planescape).
*Magic items were assumed; increased rarity was offset by letting players have superhuman to godlike ability scores (20's for humans, up to 24 for a half giant). This was further augmented by:
*A Psionics system had been produced, and was directly integrated with the setting. Every character had at least one (and possibly more!) Psionic powers in addition to whatever else you wanted to do.
*TSR was also perfectly willing to create new classes at the drop of a hat to meet setting requirements. Thus the Defiler, who goes up levels faster than the Preserver, the Templar, non-spellcasting Bards, and so on. 5e's approach would make all these subclasses, which would mean they work much more like their parent class, and they would have to make Defilers stronger than other spellcasters (in an edition where most would argue spellcasters don't need serious buffs) to make it worth not just being a Preserver. Further, when DS was made, there were TWO arcane spellcasting classes that needed adjustment. One simply lost their magic. Now upthread, someone pointed out how many spellcasting subclasses exist in 5e; you'd have to ban the lot of them!
(I will admit that you could probably make Warlocks into Templars, which 4e kind of did, though the only allowed Patron is the Sorcerer-King).
If you create a setting to reflect rules elements, when those rules elements are drastically altered, the setting is no longer the same. Dark Sun runs counter to enough assumptions that 5e uses that after all the work needed to adapt it, it won't feel like 5e anymore. Since we know "how a game feels" is important, consider that. Maybe you are fine with that because you're ready for 5e to change drastically anyways. But does the majority think that way?
If you say "hey here's this great new 5e setting but...well, it's not going to feel like the game you're playing now" to people who actually like 5e the way it is...what then?
I use Eberron as an example in that Discord post because we know for a fact it was the first setting made with 3.5 in mind. Which is why subsequent iterations have felt hollow by comparison.
I see these as significant hurdles to the development of a Dark Sun that feels anything like the setting that was created for 2e. So the question you have to ask is, if the update is that different from either what it was or whatever new system you've adapted it from, who is the product for?
Old gamers who loved the old Dark Sun? Obviously not, they have their old books!
New gamers who love 5e? Why would they change?
Let's not forget the largest assumption of 5e: make products that the largest percentage of your fanbase will buy for maximum profit.
