As other people here have suggested, papers are hard to read. Old one which mark the dawn of a subject are often (imo, but certainly not always) not well written, and poorly type-set, and modern papers assumed a lot of background.
BUT!!!! What you might want to try is just typing key words for subjects you're interested into the search bar on ArXiv. Sift through some abstracts, and find something that sounds really cool to you. Then, you can make a personal goal of trying to understand the paper(s) that you find. This will likely still involve you going back to learn more from textbooks and/or older references cited as you work your way through the paper, but I think it could be a great way where you get to start looking at papers like you want, and serve as motivation for the things you need to go look up.