Interestingly enough, the making-of videos run in Flash, but use an MP4 for the actual video itself, which means it's possible to replace that Flash object with a <video> tag. A bummer this was never implemented officially, even though the capability was there for about two years.
The original Flash objects had timed displays of the tool used during each steps. However, I do not see that being carried over to this in near future for two reasons:
The steps are represented in placeholder strings, and the file that converts those strings into actual words has been lost
More importantly, the timing of the step displays are based on the current frame number, while the <video> element's current time is displayed in seconds. Converting one to the other would involve making assumptions about the video's framerate, which I would not be comfortable with doing.