It seems appropriate to mention Gowers's "Modest Proposal" from fall of 2011. Here is his [original post][1], and then his "[more modest proposal][2]." This came about after the OP asked this question and seems very related, since a big benefit of his recommended mathoverflow-like system of submitting papers was the ability to have discussions of those papers. As others have pointed out, this also allows the community to quickly correct errors in the papers and also would reduce the lag-time between when new results are finished and when others can start to build on those results. I'm not sure what ever happened to Gowers's idea, except that there was a lot of discussion at the time and problems were pointed out. If someone knows, I hope he/she will edit this and fill in the details. This also seems related to Alex Chernov's recent answer; maybe Furmaniak's site was based on Gowers's ideas. I'm not really familiar with Reddit, so I don't know how similar it is to what Gowers envisioned. 


  [1]: https://gowers.wordpress.com/2011/10/31/how-might-we-get-to-a-new-model-of-mathematical-publishing/
  [2]: https://gowers.wordpress.com/2011/11/03/a-more-modest-proposal/