I'd say neither--it's just after the first couple document.createElement, you're driven to the brink of insanity. No one chooses that kind of torture, smart or dumb.
jQuery has two big benefits for me: cutting down on verbosity a bit, and on making a cross-browser solution possible. For the latter various polyfills solve the problem effectively, but the verbosity problem still remains.
jQuery might as well be a different language, like c++ vs. c. Can you learn c++ without c, or assembler without learning machine code? Sure.
I disagree, jQuery is more of a compatibility layer. Think of it (in a way) like an ORM for various SQL backends. They allow you to change between different SQL implementations without needing to learn the various nuances of each engine. The same can be said for jQuery, allowing you to modify classes even though browsers implement that DOM property differently.