This is too extreme. If, like Google, you have a flock of Ph.D.s who you can put onto an NLP problem to extract semantics from text, then semantic markup becomes less valuable. Not all of us are in that situation. And I don't think parsing text is the only application of the semantic web. Having hugs databases full of knowledge is interesting in itself.
As for semantic markup being confusing and usually wrong, I don't know where you get that.
Yeah, but i think there is a difference between standardized markup data formats describing e.g. proteins, and generic text with annotations. The latter are redundant
As for semantic markup being confusing and usually wrong, I don't know where you get that.