This isn’t being realistic. The major benefit of these is peer review. You aren’t going to have enough people to peer review the work of a massively open and public publication system.

On top of that the chance of finding something as you suggest becomes that much more difficult. Smaller findings get published now in a more controlled scenario and get lost in the stream.

Major journals are a net positive for surfacing important science.

Yet "peer review" would absolutely scale if it were actually the review of peers (and not just an editorial board). A large number of publications where submissions are reviewed by previous and prospective authors would be much like how open source peer review works, though not without its own set of issues.

Discovery is a search problem and its pretty clear that we have the technical capacity to solve that problem if there is enough of a signal from wide-spread peer review.

Major journals become those that re-publish and report on the big debates and discoveries of the actually peer-reviewed journals and this would be the work of "journalists".

Peer-review can also occur from non-gatekeepers, from non-experts. You realize you posted this on a massively open and public publication system, right?

Non-experts sometimes bring perspectives that gatekeepers are blind to.