Our group always pre-publish to arXiv, as a way of saying "we did this first/already" because we usually suspect, and half the times are correct, that our competitor are on a similar research. But as soon as we arXiv, we start the hunt to publish with as high Impact Factor as we can. Why? Almost 100% grants applications are going to evaluate you for the weight of your publications, and some of them doesn't even consider indexed publications below Quartile 1 or 2. The IF of the journal is a direct proxy for your work quality.

In our experience, arXiv doesn't peer review at all. You get some twitter posts, messages of congratulations, at most a couple of questions. But nothing like the average peer seriously reviewing your paper.

I wouldn't pay a dime to publish if my salary wasn't affected (e.g. being Linus). I would submit to arXiv, Github, Twitter or even a "Tell HN", and probably reach a larger audience. In fact, the journals ask you to promote your papers on your social networks, because they know theirs doesn't matter unless you are Science or Nature. But sadly that's not how the scientific world works.

The problem is that the rest of the world does not really like reviewing stuff. A few people hint frauds, which is a great service, but that’s about it.