There's an economic imperative to keep the process secretive in paid-access journals, rather than improving it - they can claim this vetting process is their 'secret sauce' and it's one way many attempt to justify their fees.
As we've learned from open-source, if you can get enough eyes (identity isn't actually that important) then all bugs become shallow.
This is an often repeated claim, but I don't think it holds up.
What is more accurate is that, given enough eyes most bugs are shallow, but the deep ones are not.
Some things are just subtle, and some things are technically deep. Throwing a thousand naive[1] reviewers at a problem like that wont get you anywhere near the result of just one person who knows the domain.
In the domain of peer review, this means that you can't replace (domain) expert reviewers with a random sample of interested people. You can certainly enhance the expert reviewers with additional people, and this is a good idea.
[1] Naive with respect to that particular problem domain.