Are you talking about subpixel rendering?
Leveraging individual R G and B dots needs knowledge of the layout to achieve increased resolution, and the algorithm should be adjusting (best effort) to compensate for human colour perception with knowledge of the subpixel layout + subpixel size.
But usually it might be that the screen has a different layout than the OS picked, e.g RGB vs BGR, vertical vs horizontal, or it has a peculiar layout like RGBG. See down the page for an example when it doesn't match, which usually gives the perceptual effect you described https://www.grc.com/ctwhat.htm
At some point though I found with increasing DPI there's diminishing returns to go subpixel and it's simply easier and just as good to go grayscale AA and be done with it.
I agree that at some point grayscale AA is good enough. But I've found that that point is at a higher DPI than I get on a 4k 32" screen. On my 24" 4k screen grayscale is good enough, so the limit is somewhere in between for me.
Other than that, I'm quite happy with my screen real-estate. So, I would indeed love it to have a higher resolution. I could probably use a somewhat bigger screen, a bit further away, that would do a better job as a TV the rare times I use it as such (don't have an actual TV). But all in all, I just want smaller pixels.
Others are talking about higher refresh rates, but for my use case it wouldn't be as useful since I rarely drag things around (I use a TWM) and don't use smooth scrolling in apps.