I think in math, and in many other fields, it is pretty normal to post all papers on arXiv. But arXiv has a lot of incorrect papers on it (tons of P vs NP papers for example), so journals are supposed to act as a filtering mechanism. How well they succeed at it is debated.
It is naive to think that “journal paper” means correct paper. There are many incorrect papers in journals too (remember reproduction crisis).
Imagine, you found a paper on arxiv-like site: there can be metadata that might help determine quality (author credentials, citations by other high-ranked papers, comments) but nothing is certain. There may be cliques that violently disagree with each other (paper clusters with incompatible theories). The medium can help with highlighting quality results (eg by choosing the default ranking algorithm for the search, introducing StackOverflow-like gamification) but it can’t and shouldn’t do science instead of practitioners.
