It's easy to forget that reading text is in no way "natural" to the human experience, it's just an old, reliable technology. Video, which functions as a proxy to in-person presence and speech, is dramatically more appealing to the average person than the abstract symbol system that is writing and reading.
It would not surprise me at all if a century from now, video is the default format, with text-first things like transcripts redesigned to minimize the downsides of video and replicate the benefits of text.