To clarify, the point is comparative. If there is a “standard” experience of difficulty, or even an “average” experience of difficulty, there will be people who have an experience that is much easier, and people who have an experience that is much harder.

Getting better is also a part of this, since the rate at which you get better also depends on previous experience. If you get better at a particularly slow rate due to the aforementioned factors, then you’ll have a more difficult time.
Theoretically, it should be “possible” for anyone to beat the game simply by getting better given accessibility controls, but this is meaningless unless we examine the differential experiences of beating the game. It took me 20 hours to beat this game; for my parents, who have only played video games from the 80s, I would expect a few hundred hours of “getting better” (if they could be bothered). That’s the point: there’s a wide divergence here, and it applies generally.

I had a look at your article. The problem with the diversity point is that you also need to clarify why specific avenues of diversity are meaningful. For example, it would be somewhat absurd to say, “On Sekiro’s main screen, the average red weight of the prime-numbered pixels read in row-major order is 45, which is differentiating and therefore meaningful.” There is also a political point to be made that what underlies people’s obsession with games without easy modes (if such a sentiment is widespread) is often elitism — something which has harmed the gaming medium and its spread for a long time.

Temptation is a valid point, but games without difficulty modes are not free of temptation. Anyone playing more than a week after release on PC has access to a slew of modding resources that are at least as powerful as difficulty controls. There are also the game-internal temptations like grinding (again less common in Sekiro). The question therefore must turn to the proximity of temptation, which can be controlled even for games with difficulty modes.