Different people learn in different ways.
https://www.psychologicalscience.org/news/releases/learning-...
Motivation and styles is probably a different matter. There is the problem that self motivated learners can only learn if they are motivated, but also that they can be led astray by materials that primarily motivate and entertain.
In practice, I wonder whether this non-text based learning is more useful in the earlier stages of learning to write code. This might because your learning much more than how to write a valid set of lines that compile to an expected result, there's a lot about style, motivation and norms that you have to soak up too. A lot of that is often implicit in text based tutorials, whereas the video based format makes many minor details explicit (e.g. seeing someone type a line of code wrong, and then explain, 'oh yeah this is a commmon mistake'... this actually encodes a lot of information a text based tutorial would typically miss)
Given that subjects have inherent styles of medium that become more essential as you approach expert and need to interact with experts in an efficient way, if you convince someone they have an innate problem you deprive them of the motivation to become an expert, and you see more and more people going to a junior level and dropping out.
I’m not at all convinced that programming is primarily text oriented in any case. The underlying concepts are mathematical and logical, not textual. People don’t learn through text or video. Those are mediums. They learn through metaphor and relationships.
So the question is which medium gets those metaphors across most efficiently. Consider, for instance, a binary tree. It’s a metaphor. And I’d be willing to bet most people new to the topic would learn that metaphor better (or at least faster) with a person drawing out a tree and pointing to the path a program would take to add/remove/list/etc. the nodes on the tree than they would seeing a block of code and having someone explain it in text.