Granted a lot of this is just speculation based on following more recent developments and projects on the Oculus forums/subreddit but I think along with the 3d movie idea, telepresence would be a big one.
There are already a few teams working on prototypes that make use of one or more Kinect depth cameras to "record" a room in 3D so you can put yourself back in these recorded 3d scenes later on. Others have been experimenting with encoding and compressing this data for transmission over networks and decoding at the other end to provide a sort of real-time 3d telepresence.
Obviously it's limited by using Kinect depth cameras, current consumer computers, and standard broadband networks but the way I see it, my first Treo smartphone was limited by the SoC and touchscreen tech of the time as well as the wireless data networks of the time. Still, it eventually evolved into the iPhones and Nexuses of today.
I think that at some point, the head-mounted hardware will become more ergonomic and more functional at costs that allow them to be sold to average users. Likewise, the hardware and software to capture and encode 3d video models of a room or "scene" will become more commonplace. Right now it's games and demos...parlor tricks like the first motion picture reels that eventually led to IMAX movies.