On the standard eye test chart where 20/20 vision is the ability to read the bottom row, I could usually read the copyright fine print below the 1-inch high block letters!
To me it always felt like that is normal vision, and 20/20 is taking an average that includes a bunch of people with poor eyesight. Like... the average person has less than 2 eyes, less than two legs, etc...
It’s been a while since I read up on it but my recollection is that the healthy eye in a young patient (“normal” in the sense of “lacking any deficiencies whatsoever”, not in the sense of “average”) should be around 20/12, with natural variation between 20/10 and 20/16, and 20/20 indicating (extremely minor) deficiencies. The theoretical perfect eye from an optical standpoint, something like if you were to grow an eye from the same materials as a real human eye but you could specify exact dimensions etc., is something like 20/7 (as in, you can’t get better optical resolution than that given the optical properties of our lenses and the distribution and size of our light receptor cells).
Eyes are extremely complicated structures that are either very sensitive to environmental conditions, or that the body is very willing to cut corners on given the slightest provocation, so few people get that “perfect” vision.
There is one very good reason for considering 20/20 to be “normal”, though - since it has been accepted as the standard for over 150 years, most of the human-designed environment has used it as a yardstick for designing visual elements such as roadsigns. The size and shape of roadsign lettering and their placement relative to the road feature they’re describing is all according to a standard that was worked out on the assumptions of (some multiple of) “normal” 20/20 vision, for instance. Likewise for just about every other thing you have to read. So it makes a lot of sense to use that as the threshold. In this sense perhaps it helps to read “normal vision” not as “a normal eye” or “the normal acuity among average healthy humans” but rather “the vision threshold needed to interact normally with the human world”.
Full disclosure: I’m an EndMyopia conspiracy theorist.
Viewing distance also varies a lot - I recall reading somewhere the empirical observed viewing distance for smartphones is between 6 and 15 inches sitting and between 4 and 9 inches lying down. Environmental brightness also varies, obviously.
Just to give a personal example, my near vision was measured at 20/16 with correction, I am currently holding my phone 7 inches from my eyes, and I have my room lit with four 1500-lumen bulbs (to simulate daylight). I don’t know what the correct PPI would be there, but it’s obviously going to be higher than the usual amount.
In addition, it’s not clear that “cannot distinguish individual pixels” is even the same thing as “no more visual quality can be gained with higher pixel density”. At 30fps the human eye can’t distinguish individual frames, but 60fps still has visual benefits over 30fps, and many people are able to consistently detect the difference between 60fps and 144fps. Past 144fps is when you get to “no more visual quality can be obtained by higher frame rates” (as opposed to proxies like “individual frames are not distinguishable”). I suspect a similar thing is true for pixel density.
But as I said originally, I’m excited that microLEDs might make it viable to have screens with a pixel density not just “beyond what’s reasonably good enough in most circumstances”, not even “beyond what’s detectable by most people in all but the best circumstances”, but in fact well into “provably beyond the limits of what the best human eyes can detect in the best possible conditions at impractically close viewing distances”.