> Most new research; sometimes a paper gets posted but most of the audience doesn't have enough background to evaluate the implications of things even if they're highly relevant. I've seen a lot of cutting edge stuff get overlooked because there weren't enough upvoters who understood what they were looking at.
Is there a good place for this? Currently I just regularly sift through all of the garbage myself on arxiv to find the good stuff, but it is somewhat of a pain to do.
I am not aware of any central, public location; I use a combination of arxiv, Google Scholar, Huggingface posts, and private Discords. The problem is that most of the public space has been poisoned with AI hype, and it's nearly impossible to find more than surface-level introductions for a lot of topics because the substandard Medium posts and YouTube hypemongers drowns them out.
Having a background in machine learning helps, because at least I can search for terminology that hasn't been picked up by the hype machine yet.
There's some communities that are more niche; more academically focused Discords or groups where there's better discussion going on, basically. Those are intermittent enough that you can't expect ongoing general discussion and for most I'd have to go back and check if they're still worth reading past the one discussion I found useful.
But for the wider internet, the hype train has forced most of the informed discourse off the road.