It's the DMs job to entertain the players by presenting them with challenges to overcome. If the players enjoy the challenges they are a Good Guy (and they keep insisting that he continues to DM, like forever), if the players don't enjoy them, they are the Bad Guy.
Thing is, that doesn't depend on the rules, or the challenges, it depends on the players, what sort of thing they like.
Only a minority? It seems to me that a great many people like to play videogames on Hardcore Unfair difficulty settings. My players keep talking about their Honour Mode playthroughs of BG3, and how they are getting on with Elden Ring. That's not for me, but it seems plenty popular.
A minority
of D&D players can still be "a great many." A minority
of video gamers can still be an enormous number of people.
You are conflating absolute number, which is in the hundreds of thousands (a large
amount), with relative percentage (which may be quite low).
Only about 10% of attempted Honor Mode runs of BG3 succeed, per Larian's own collected data. About 1.2 million failed attempts, and about 140,000 successful attempts. This compared to the total sales of about 15 million mean that only slightly-less-than 10% of players (assuming one attempt per player, because otherwise it would be a
lower percentage) even try to attempt Honor Mode, and only slightly-more-than 10% of
those people succeed. So, yes, I think it's quite reasonable to say that this is a minority preference. That
absolutely does not mean that it should be deleted, not by any means! But it does mean that if BG3 had been
set at Honor Mode from the get go, and only allowed players to attempt a lower difficulty after jumping through a bunch of hoops and being told repeatedly how much they were making the game super crazy easy etc. etc., there is absolutely, positively no way that it would have succeeded the way it actually did.
And the Elden Ring analogy is also fraught, because while it is hard, it is hard
BUT FAIR, and it includes what are functionally "easy mode" approaches to play. That is, there are some playstyles or approaches that are much easier to use than others, to the point that some players disdain them for being
too easy. (I have not played the game myself, but I know several people who have.) Such has been true of every "Souls-like" game, or whatever we wish to refer to it as.
Looping back to the "but fair" side of things, this is something a lot of old school DMs don't really seem to grok very well. They love what I would call bovine feces challenges, where mere crap luck or coincidence not only easily kill, they
frequently kill. That is precisely the antithesis of the "Souls-like" subgenre's offering. The whole point of this kind of "tough but fair" difficulty is that the game is absolutely consistent and (almost always) NOT random; if you die, it's
very specifically because you yourself actually did make a bad tactical decision (or, more commonly, a string of them), not because a boss got an unlucky critical hit or whatever. Not because there was a dumb "gotcha!" monster (ear seeker, cloaker, piercer, lurker above, black pudding, etc., etc.) and the DM gleefully put another character sheet in the bonfire, but because the player actually could learn, theoretically even in the very first encounter, how to fight and thus how to beat absolutely any monster or boss.
That's the "hard
but fair" gameplay that does, in fact, attract a reasonable number of customers (though still not the majority of gamers, because most gamers want a reasonably-paced leisure time activity, not a gauntlet they must overcome with great effort.) It's also almost nothing like my experience of old-school D&D, and nothing at all like what I keep
hearing about old-school D&D, which very much prides itself on puzzles and problems and monsters that
cannot be solved until you have already lost to them at least once before.
