What would be a components of an "ideal" solution? I'd love to hear improvements or corrections to my assumptions.

1) Free to view all approved papers (possibly run by a non-profit and funded by grants around the world)

2) Establish reputations for scientists (giving scientists access to future grants)

3) Peer review handled autonomously (submitted papers are automatically distributed autonomously by the system to other scientists in the related field with a high enough reputation)

4) Easy searching for articles (articles have semantic structure, are categorized by author, title, subject, citations, etc.)

5) Removal of "scientific silos" (all articles published to the same site, categories define content rather than publishing brand)