But this goes w/ the underlying assumption that graphical advancement is the only thing that can be made, which makes no sense. It's all about how the graphics are supposed to draw you in, while we're still stuck with the conventional controller scheme limiting how we affect the world that's being drawn with ever more realism. You'd have a far more profound effect on advancing games if a little more money's put into digesting Kinect/Wii/Move/etc control schemes, advancing that to another level, and providing experiences that can really change the way games are played out. Money just never goes in that direction because people have been taught to be solely fixated on visuals, and throw out everything else.
Extra graphical fidelity is great, but what really made this gen was the true rise of social integration into our consoles. I don't mind more graphical bling myself, but I do not think it'll convince a lot of people to switch over by itself.
I can't say I agree with your opinions on what constitutes good progress for games as a medium. I was happier before I had to flail my arms to do what amounts to a button press or tell games that, no, I don't want them to tweet about what I just did. The social media integration is mostly just a tool to keep us stuck on a particular format or game by using our friends as an anchor rather than something that has meaningfully contributed to games as a medium. Halo: Reach death match isn't a terribly different experience from Halo 2.
Beyond that, more power doesn't just allow things to look nicer. It allows for a more complex game world with more details, more interactivity, and greater persistence of details.
Consider the following. All objects in the environment potentially interactive and dynamically destructible, with properly calculated weight and density. Battles with the scale and complexity of the real thing rather than 60 or so mannequins on screen shuffling around like a Dynasty Warriors game. Logical, persistent social interaction with NPCs in games like Skyrim rather than the entire town colluding to beat you to death for stealing a cheap sword and meanwhile standing on top of a pile of their slain family members having a chat about the weather. Complex hit detection, procedural animation and model deformation so that everybody's clothing and hair moves realistically, armor and swords don't clip through the player character's body, and you can hit objects and people with flexible objects like whips and damage and animation will be determined and displayed dynamically rather than just relying on static hitboxes and canned animations.
Many of these elements have been used individually in the past as a central gimmick for a game, but they could all work in tandem on more powerful hardware. And that would create new experiences that are more robust and engaging than people even considered possible a few years ago. People who think we've reached the point of diminishing returns just aren't thinking ambitiously enough.
