Do not try to do it because you know you can do it.
Gameplay should be first, all things (even graphics) are secondary. If the game is fun and enjoyable but has poor (or not so next gen) graphics, it will still be fun and enjoyable and people will play it, and also your metacritic will be good. Otherwise if the game has awesome graphics and features (realistic fluids, procedural mesh destruction, etc) but is not entertaining or difficult to play (bad controls, etc), people will not play it. No players = no money and also bad metacritic.
So first plan the game you want to do and think about its gameplay features, its playable situations. Do not try to push trying to make that feature X fit into the game just because it looks cool. If it doesn't really make sense, or if it's not going to represent an important part of the gameplay, simply drop it.
Also, avoid trying to build gameplay around one of those awesome features if it doesn't make sense or you think it's not going to be fun. For example: you might think "I have procedural mesh destruction, so let's force the player to destroy everything before he can continue advancing in the game (so he can see meshes being procedurally destroyed)".
To sum up: first think about your game and about its needs. From that base, plan your development phase, and then you will see if there is enough room to fit one or more of these awesome features (and if it makes sense to add them to your specific kind of game).