This would be challenging with modern codecs using delta frames. The only way I can see it work is precomputing all frames from the preceeding keyframe. Doable, but decent effort for a fairly obscure feature.
I never saw BeOS do that with video, but I heard it do it with MP3 files. SoundPlay was a kind of crazy bananas MP3 player -- it could basically act as a mixer, letting you not only "queue up" multiple files but play each of them simultaneously at different volume levels and even different speeds. I've still never seen anything like it outside of DJ software.