How does one auto-caption a white/black-board in a world before insanely cheap computer vision?
Now if you are talking about the blind then sure thats different but the blind don't need captions.
What if they are hard of sight and can't read the whiteboard or blackboard, even if they might be able to read a captioned video?
Lord knows I remember barely being able to read some professors handwriting, even in person, or being able to understand what the hell a professor was saying in person.
How many people are there that can efficiently caption complex mathematical equations?