Epistemic status: This post touches on bodywork, NLP, Salsa and Yoga. I went through a 4-year training in Danis Bois style of bodywork, have roughly 70 seminar days worth of NLP training, and started dancing Salsa back in 2008 and have danced it a lot since then. I don't have special expertise in Yoga, but apply general knowledge about body work to it, because I expect that my readers will have an idea of what Yoga is but not know Danis Bois's work. 

In his recent post, Are Woo Non-Responders Defective?,  Scott Alexander writes as one possibility why some people get value from “woo” and others don’t as:

Some people process their emotions in their bodies. Other people process their emotions some other way, maybe in their minds. Both methods have advantages and disadvantages, but neither is better or worse than the other. Neither one is “more natural” - maybe it’s genetic or something.

I have found NLP concepts useful for making this distinction and drawing practical implications.

A decade ago, there were years when I was going Salsa dancing in a club two nights per week or more. Frequently, it happened that I was asking a woman to dance and I'm hearing from her "I don't know how to dance Salsa". Over time I learned, what's required to make it work so that I can dance well with those women. 

If the woman I’m dancing with has a decent amount of body tension I can lead in a way where the natural physical reaction to my leading is to make the correct dance moves if the follower's perception is in the mode where she orients herself via kinesthetic input.
If the woman however looks at my feet and is the in the mode where she orients herself to listen to visual input that hinders our dance a lot.

Part of learning to dance Salsa for a follow that’s used to primarily orient themselves according to visual input is to switch to orientation to kinesthetic input. As long as a beginner follows orients themselves primarily visually they might still manage to learn some dance moves that way, but it gets in the way of the flow of the dance.  

NLP theory proposes that for every action we are doing, we are using one primary channel. A lot of what’s written on primary learning styles found that’s when teachers classified students according to their learning styles and then gave different students different ways to learn based on the learning style in which they classified the students this didn’t help them to improve learning outcomes. You will find some people that this body of research invalidates NLP’s theories about primary learning styles, but I would posit that the research was largely done by people drawing conclusions for the NLP theories that the NLP founders wouldn’t.

If I go back to the Salsa example, the Salsa follower who is used to the visual mode as their primary system has no special need for having videos of the dance moves to review the dance moves that were taught in a class. The videos are useful for everyone to review the moves. The important implication of the theory of primary processing systems is rather that even the followers who are used to primarily consuming visual input have to learn to orient themselves according to kinesthetic input to dance well.

I have been to a lot of Salsa classes but I remember no Salsa class where beginners were actually told that switching from orienting from visual to kinesthetic input is an important part of the learning. The general method seems rather that if a follower takes enough classes they will sooner or later learn to orient primarily toward kinesthetic input during dancing.

On the other hand, if a beginner approaches computer programming and approaches it in a way where you feel into kinesthetic senses, they are going to be bad at learning computer programming. An advanced programmer might still benefit from noticing the kinesthetic feeling of confusion within their body but it’s hard for me to imagine any expert programmer to approach it in the kinesthetic mode.

If the average computer programmer who’s used to approaching most issues with the visual primary processing system starts Yoga, they might approach it as: “Let’s watch the video and see how the position looks. Then let's copy the movement so that it looks exactly like the one in the video.”
If they go to a bad Yoga class they are likely approaching it in the same way: “Let’s look at the teacher and copy what they are doing. If the teacher points out some detail that I’m doing differently, then let’s change so that my movement looks exactly like the one of the teacher.”

While that computer programmer, who practices Yoga in this way, might get a bit fitter, they won’t get the deeper bodywork benefits. The deeper benefits depend on the change to orienting toward the kinesthetic input mode when doing Yoga. For many computer programmers that change would be hard to pick up without a good teacher.

When having conversations with other people both primary modes have benefits. People who are in the kinesthetic mode speak slower as the process of feeling sensations takes time. In the extreme that sounds like Eckhart Tolle. If you want to convey as much information as possible in a short amount of time the visual mode is advantageous. If interpersonal emotions are important to the conversation the kinesthetic mode is advantageous. It helps both participants perceive their own and each other's emotions. It also helps both with processing emotions.