Researchers at Microsoft Live Labs are working on Photosynth based on Photo Tourism from the University of Washington. By structuring the photos based on their relative position to each other, it is possible to navigate in a large photo collection in a 3D style way. The system looks very responsive from the video, but I would be curious to see how it works in a real-world context.
It would be very interesting to create similar navigation tools for audio. I guess someone (probably in the ISMIR community) is working on this, and would be very interested to hear more about this.