Interesting that you bring up VR workflow immediately after acknowledging everyone abandoning 3D implementation. Making a 3D mode for games is probably the best pre-VR way to help establish best practices for VR projects, considering it also involves 3D dimensionality and two seperately rendered images. Games with 3D modes could also be fairly easily retrofitted with VR functionality once the peripherals arrive, to help strengthen the portfolio of compatible software out the gate. If everyone's so damned interested in VR development, why does no one seem at all interested in something that helps cut teeth on pipeline skills they'll need for VR?
3D is at the bottom of the list of what makes VR a challenge to develop for. All you have to do to make a 3D mode in a game with 3D space is add a second camera. Games with 3D modes would not be fairly easily retrofitted for VR and there are no VR practices learned from simply making a 3D mode in a traditional game.