The problem with physics for me is that the performance hit for what they've shown off so far just isn't worth it for me. A lot of this stuff can be done by "scripting" it. Sure, it won't be as realistic, but you can get most of what it adds for immersion without the massive performance hit. At least that's my opinion based on the cloth/particle/small object physics that has been made a fuss about. Likewise, with water, in most games it just wouldn't be worth it to implement it, and so you'll basically have to base the gameplay around the interaction with it. The bad thing is, while that would be awesome (imagine how much better the gravity gun would be), the performance hit would be so high that it'd be like Crysis.
I'd say that's true of a lot of these eye candy stuff recently, its not integral so it seems tacked on and the performance hit for it is just too high to justify.
It seemed like games were moving towards a return to software focus for processing (rendering, etc), which I was looking forward to as it seemed like that would allow them to implement a lot of stuff quicker/easier, where you could control the performance aspect more, where if you want better performance you just pay for more computing power (which granted is the case now, but you weren't limited by hardware support for certain features). It seemed like that was the best way to get audio processing out of the old Pre-Vista days, add physics, and even stuff like ray-tracing. This way, also, those things become integral to the game.
I was never big on Ageia and the PhysX card (mostly because of the aforementioned move to software focus), but it kinda sucks that dedicated physics hardware didn't take off, mostly because CPU based seems to have stagnated (CPU peformance improvements in general just has been mildly disappointing for several years, at least for desktop, although a lot of that is the low support for more cores and doing anything interesting with them), and GPU physics is pretty much a joke and it seems like its already being tossed aside in favor of tessellation.
Sony or Toshiba should have made that Cell add-in card they talked about way back when. It could have been used for physics (and other things) in games and video processing, and plenty of other things (3D conversion, audio processing).
Facebook
Twitter