It's all Flash with some FLV video files. In the Billy Harvey site, it would appear that they used Navigation cue points but event cue points could work as well.
The on(press) or onRelease() would trigger either a seek to whatever navigation or, it would just load a new FLV. The short time between would need either a still shot of his face to mask the second change.
The rest is just all actionscript for animation and movement.