As someone who does a bit of graphics programming in his day job and as a hobby, it can be said that I'm a fan of graphics technology. Despite that, I think it's long past the time where PC developers should stop worrying about graphical spectacle, because the pursuit of graphics is doing us more harm than good.
In the original Wolfenstein 3D, the "level editor" was a simple little program that let you draw squares on a grid to create gamespace. You could make a playable room in well under a minute. It was laughably simple and primitive by today's standards, but the game was at least forty hours long because the content was so easy to produce. (It would have almost been possible for someone to make levels as fast as you could play them.) A few years later in the Doom and Duke Nukem 3D era, level design had become slightly more elaborate. It took time to get the textures to line up and make the lighting interesting; that same room of gamespace might take five or ten minutes to produce. With Quake, the bar was raised even higher. Level design was basically 3D modeling, and it might take a whole hour to make the same amount of content.
You can see where this is going. The one hour room gave way to two hours, and eventually led to teams of people working for days to make just a few moments of playable content. Now you have someone designing the level, someone else making unique meshes to decorate the space, a specialized texture artist, and a lot of work being done to set up complex lighting systems, moving machinery, special environmental effects, and all of the other steps needed to take advantage of current-gen graphics engines. That's more than a thousand fold increase in the amount of work required to give players a few seconds of entertainment. This inflation of manhours is obviously unsustainable, and even the amount of work we're putting into games now is probably too much. Taking another step forward is folly.
Don't get me wrong, these sexy new polygons look great, and I certainly wouldn't want to go back to the days of pixelated 2-D sprites sliding around a repetitive blocky room, but the problem is that each new graphical step forward has cost us more and given us less in return, and I think at this point we're getting a lot less than what we're giving up.
As these costs rose, we started getting less game for our money. Games began getting shorter. Forty hour games became twenty hour games. Then ten hour games. When developers couldn't make any more cuts to gameplay, they began protecting their investments by simply taking fewer risks. It's one thing to try something outlandish and innovative when a game costs half a million dollars to produce. It's quite another to do so if the game is going to cost twenty million, and anything less than a complete commercial success will spell bankruptcy for your company.