I think one of the really neat things about games as a medium is that "the experience" is inherently a super malleable concept. Gaming blows my mind when I think about how adaptive you need to be to run a tabletop roleplaying game, like Dungeons and Dragons — no matter how elaborate your plans are, players will always find a way to throw a spanner in the works. Video games have the same unpredictability of how players engage with the world you've made, but a much smaller ability to respond and adapt to ensure that they're getting the correct "intended experience".

In some respects, I agree with you, because when I play games, I care a lot about the intended experience. However, the reality is that I bring too much of myself to any game that I play to be able to think of my experience in that way, and I think that's probably one of my favourite aspects of games as a medium — a dialogue between gamer and game developers. Especially because sometimes, the intended experience of a game isn't well executed; there are plenty of times I have gotten lost or confused in games because the game didn't sufficiently communicate to me (or other players with similar experiences) what it expected us to do. Part of the role of the game designer/developers role is to be guide the players so they get something resembling the intended experience.

Honestly, part of why I am on the pro-accessibility side of this issue is because I'm a bit of a snob — I think that being able to adapt a message or experience to a diverse audience shows a singularity of vision that's more powerful than experiences that target a much smaller audience.

For example, let's say that the subjective difficulty level of a game (the "experience") equals the "objective difficulty level" of a game (the difficulty setting) minus the player's skill level. For the sake of this example, let's imagine that 10 arbitrary units is the correct level of the subjective difficulty level, and above/below that, the experience is degraded; also, let's say that player skill ranges from 1-10, with most people clustering in the 4-6 range. In that world, if a game could only have one difficulty mode, 15 ish would probably be best, because 15 (objective difficulty) - 5 (average player skill level) = 10 (intended subjective difficulty level). I don't begrudge game Devs for targeting limited audiences if that's what they feel capable of, but I do massively respect the craftsmanship of being able to build a game that can serve a subjective 10 to a wide range of people, by having a range of difficulty settings.