You should read review papers. There are journals that either only publish such papers (e.g. Review of Modern Physics) or some of the larger journals sometimes

You should read review papers. There are journals that either only publish such papers (e.g. Review of Modern Physics) or some of the larger journals sometimes publish such papers (e.g. Science or Nature). They are usually written by one or more experts in the given field, are less dense than normal papers and focus on the current state of the art and future research directions, so they're pretty much exactly what you want. They will also contain tons of references to relevant work. Arxiv is just an unfiltered stream of mostly garbage (pardon me) so unless you're already familiar with a field and know how to filter out such garbage it's pretty useless I'd say.