I watched the relevant portion of the video. I can't speak for the presenter, but I doubt he means to suggest that the camera position would ever lie between the near and far planes. (I think there might be some confusion due to his using the same coordinate grid display for both camera and NDC space, while keeping the white 'camera' graphic onscreen for both spaces even though it's really only relevant for camera space. Again though, I don't want to speak for him, and I may be wrong about his intent.)
In short, he seems to be talking about how a perspective projection maps coordinates from the frustum volume (the truncated pyramid carved out of the unbounded pyramidal view volume by the near and far planes) to the 'NDC box'. In camera space, the camera position is behind the near plane, and as far as I can tell the presenter of the video doesn't mean to suggest otherwise. Again though, I agree that the whole topic can be somewhat confusing.
