While I’m not well versed with Adversarial networks to agree or not with your comments on the shortcomings of the network, I wholeheartedly agree to pretty much everything you have to say on the matter. More so on your comments about Arxiv preprints.
There’s no denying that preprints are of immense importance in this day and age when people refer to paper released a year ago as ancient history, and if someone could put a paper on Arxiv at the same time when you submit to a journal, or even a conference, by the time it’s published, it’s already outdated. I ran into this: https://peercommunityin.org/ some days ago which aims to solve some of these problems. Its a pretty ambitious project and I’m wondering that if manifested into it’s intended form, it might change the way research outputs are handled.