This is not the case with fiction or poetry, where what is communicated is often precisely what cannot be communicated socially or even explicitly. The experiences of both writing and reading are often in a different realm altogether than those of speaking and listening face to face -- one is not a watered down version of the other. They have different qualities.

And even in the case of scientific and mathematical exposition, where your statement is more often true, there are many exceptions. For example, I think of professors I've had who could write lucidly but were poor teachers, both in the classroom and in office hours. Either their social skills stood in the way of their communication, or their verbal skills were not as good as their written ones. They needed time and solitude to express their thoughts clearly.

This is what Feynman himself wrote in the preface to the books under discussion.