There are a lot of comments debating whether 8K is worthwhile vs 4K, and whether we can see a difference. I decided to do some math.

Let's assume someone has a 65in (165cm) TV and sits 6.5ft (~2m) away. This is a larger TV than most people have, and a closer viewing distance than most people sit. Let's also assume that the viewer has 20/12.5 (6/3.75) vision. 20/20 (6/6) is the low end of normal, and it isn't uncommon to have slightly better vision. My optometrist (my fiancée) can correct me to about 20/15, but 20/12.5 is unusually good.

The screen width is 56.65in. At this distance, that subtends an angle of about 40 degrees:

    >>> math.degrees(2 * math.atan(56.65475464133182 / (2 * 6.5 * 12)))
    39.9191582380095

20/12.5 vision corresponds to 48 cycles per degrees, or 96 pixels per degree, distinguishable (20/20 is generally accepted to be 30 cycles per degree). Therefore, the viewer can distinguish 96 * 40 = 3840 pixels horizontally. Which is UHD 4K resolution.

I chose the inputs to get to that result, but my point is that viewers would need to have unusually good vision, and unusually large TV, and be sitting unusually close to hit the limits of 4k. And that's only for very high contrast images, like black text on a white background. Our eyes are less sensitive to low contrast transitions, as found in most videos. Try reading an eye chart with the letters printed in gray on a lighter gray background. It's more difficult!

I think there are much better things we can use our limited bandwidth for than increasing resolution: we could use better compression codecs at higher bitrates, reduce or eliminate chroma subsampling, increase color depth and/or gamut for HDR content, or use higher framerates for content where that is appropriate.

EDIT: Here's a source for 20/20 = 30 cycles per degree, and 20/12.5 vision being unusually good: https://www.opt.uh.edu/onlinecoursematerials/stevenson-5320/...