[dbmueller]

I'm not sure arXiv is an example of doing without journals. arXiv doesn't involve peer review, so that any result published there are to be taken with a big grain of salt.

On the other hand, there are open access journals that essentially act as a layer above arXiv, so that the main added value is review.

[patrick5415]

I don’t think science would suffer too much if we just abandoned peer review and committed to the arXiv route. Peer review does little to ensure research is actually rigorous. Rather, it mostly ensures the authors know how to play game: use the right language, cite the right people, and don’t stray too far from the current fad while simultaneously selling what they’ve done like it’s the hottest thing since relativity.

Once you’ve read enough papers, you don’t even have to invoke the replication crisis to see that most papers should be taken with a grain of salt.

[riku_iki]

In AI field such layer is recognizable conferences, researchers compete to get their papers accepted, and conference committees are doing peer review.