People come in a variety of ways, and with a variety of learning skills and patterns.
Some are highly visual, and learn primarily through watching. The teacher/student model works for them (live or delayed). (There is a vast library of this that already exists on video, on platforms that exceed tiktok. )
Others learn better from source text material. A few good books and they have all they need. The lecturer is "just reading the book" so its faster to skip the middleman.
Understanding this difference is key to creating a good spread of training material. It's not sufficient to "just produce a manual" or "just make videos".
Personally I find some topics work better in visual form (explaining exercises for psyio recently) and some work better as text (api reference.)