Sometime between the release of the N64 and the release of the previous generation consoles we hit a point where everything that exists in the real world or we can imagine could be represented using real-time computer graphics and be recognised by most people. At this point in time we hit a threshold where the practicality of improving real-time computer graphics for communication purposes was very limited.
Now, the OP seems to be heavily centered on when people won’t see much of the benefit of further processing power put towards graphics. If you look at how graphics improved over the past 10 years (from Half-Life to Crysis) and then project a similar improvement over the next 10 years we will have easily passed a point where the typical person will have difficulty seeing any further improvement in real-time computer graphics.
