I'm in philosophy, though personally working in formal philosophy, so this is quite different from CS. In my area practically all top journals reject more than 95% of all submissions, and if you get accepted, it takes about a year until the article is published. Many journals get so many submissions that they do not provide any reviews or feedback - not even a paragraph or line - if they reject a paper.
Although I have been able to stay a fully paid postdoc for more than 10 years without problems and get decent amount of publications out, I'd agree that the peer review system is broken and seems to be strongly biased in favor of mediocrity over originality. It's particularly bad in fields like philosophy in the humanities, where you could argue that there are no clear-cut criteria for evaluating the content other than merely formal issues. (Philosophy is not a science.) There are lots of people who game the system, I'd say many of my colleagues are doing it, though not (yet) the majority. By gaming the system I mean writing articles they barely endorse themselves, tailored to get past reviews. I've literally had a very successful colleague say "Who cares? It's one more publication!" when I raised doubts about a joint paper, because it involved too much hand-waving and didn't seem to work.
Long story short, I have the impression that in the humanities whether you're accepted or not mainly depends on the writing style and on whether you win the "reviewer lottery." One and the same paper can get accepted in one journal and rejected in another, for journals with similar reputation. I also believe that there are certain niche areas in the humanities that are (almost) pseudo-sciences and based on a small group of researchers colluding with each other and passing opportunities for publication to each other. I won't mention the disciplines, though.
I'm a bit dismayed that you describe CS in a similar way, I've always thought its much more rigid there and easier to evaluate CS papers.
Having said all that, I don't think there is any alternative to the peer review system. Any alternative I've seen so far or come up with myself seems to be even more prone to manipulation and gaming the system. I believe that it's not the peer review system itself, but over-reliance on objective indicators by university administrators that has led to most of the problems. But the humanities have certainly changed to the negative, at least in philosophy I'm certain that many seminal and famous articles from the 70s of last century and earlier would not make it past the peer review today.
sadly, CS is also about the reviewer lottery. however because it is conferences rather than journals, turnaround is much faster at least. and there is a culture of preprints, so having a paper rejected once doesn’t delay getting it out into the world.
Although I have been able to stay a fully paid postdoc for more than 10 years without problems and get decent amount of publications out, I'd agree that the peer review system is broken and seems to be strongly biased in favor of mediocrity over originality. It's particularly bad in fields like philosophy in the humanities, where you could argue that there are no clear-cut criteria for evaluating the content other than merely formal issues. (Philosophy is not a science.) There are lots of people who game the system, I'd say many of my colleagues are doing it, though not (yet) the majority. By gaming the system I mean writing articles they barely endorse themselves, tailored to get past reviews. I've literally had a very successful colleague say "Who cares? It's one more publication!" when I raised doubts about a joint paper, because it involved too much hand-waving and didn't seem to work.
Long story short, I have the impression that in the humanities whether you're accepted or not mainly depends on the writing style and on whether you win the "reviewer lottery." One and the same paper can get accepted in one journal and rejected in another, for journals with similar reputation. I also believe that there are certain niche areas in the humanities that are (almost) pseudo-sciences and based on a small group of researchers colluding with each other and passing opportunities for publication to each other. I won't mention the disciplines, though.
I'm a bit dismayed that you describe CS in a similar way, I've always thought its much more rigid there and easier to evaluate CS papers.
Having said all that, I don't think there is any alternative to the peer review system. Any alternative I've seen so far or come up with myself seems to be even more prone to manipulation and gaming the system. I believe that it's not the peer review system itself, but over-reliance on objective indicators by university administrators that has led to most of the problems. But the humanities have certainly changed to the negative, at least in philosophy I'm certain that many seminal and famous articles from the 70s of last century and earlier would not make it past the peer review today.