The higher the resolution the more you hit diminishing returns at the same screen size and distance.
So a game rendered at 1800p will be hard for the huge majority of people to tell the difference between 2160p, yet the raw TF power needed to render it is pretty substantial comparatively.
Not facturing in temporal up scaling, and everything else that's making pixel counting harder.
So if a game looks the same, but one loads in 5 secs and the other 8, then you could argue that it's more easily demonstratable, and easier to see the difference. Even if in the big scheme of things it's not like it's a difference between 1 and 2 mins.
So, how do you think it gets marketed and reported? Unless your DF doing pixel counting (which they are already finding hard in some cases) , the reporting will be both looks the same but one is better because of faster loading. Even DF could end up saying rather the faster speed over perceptually equal IQ, which will cause melt downs on the net.
It's about what can be seen and demonstrated, and with resolution it will be getting hard.
So TF isn't the be all, end all that some people think. As long as in same ballpark, and 9 to 12 is if just lower the resolution of target 4k.