Aningan: And besides there is no reason to drop graphics all together. I mean just take a look and what moders did with the UT engine in these past years. Some very good mods or stand alone games. If a developer studio can't even match what some independent moders are doing ( and some can't) are we really supposed to give them the benefit of the doubt and buy their games?
I'm doing game development right now, having finally fleshed out the full complement of my team (5 individual human beings, woohoo!) and we're using the UDK with PhysX and various other middleware I'm paying for primarily out of pocket. (PhysX is thankfully free!)
I've been messing with shaders and physics code off and on while my other team members make assets and concepts, and I find it to be extremely simple to make a game hit or surpass AAA graphics quality. Models that are blocky can have that easily hidden by masterful texturing, look at the Quake 1 modding community and their monster skin packs.
It's not hard to make what would be considered top-of-the-line graphics today, especially because 90% of the time you actually don't need to worry about things like polycounts anymore, filling the screen with tris will barely knock the frame rate.
Animations are harder to do, but considering you don't have to worry about optimizing much of anything else (unless you're working with retardedly detailed stuff to begin with) you have all the time you could ever need to make your animations at worst passable and at best lovely.
Know what I think bogs down commercial AAA game development? Massive teams and feature creep. The more people you pack into a room to work on something, the more time it will take you all to shut up and start doing the work -- and this is even more obvious in polarizing things such as game development, where your idea of "good game" will 100% certainly deviate in some measurable way from that of any other person you will ever meet, and nobody is going to back down on their opinions about games.
Feature creep, well.. I don't see it occur often in AAA studios because they tend to have deadlines and such. But I'm sure it's a serious consideration, especially because of how many independent games fail due to it. "let's add this, and this, and this, and this" -- even I'm guilty of that. Currently I'm culling stuff I'm unlikely to implement in my work, and that's one hell of a lot harder to do than add more bullshit you do not have the time or experience or manpower to see through to completion.