Low quality papers will get identified quickly when we passed a certain userbase threshold. The scientific crowd can evaluate a paper a lot better and more dynamic than simple peer reviews.
The whole platform will work very similar to steemit, but with a seperate curation system. Every account will be connected to a real life person, and we let them do whatever they think is best. It's their work, and their reputation. We can't let someone else decide for them what's good or not.
Filter functions will allow everyone to create their own bubble when necessary.
I also think bad or even outright hilarious papers and their metadata like votes, voters and comments may provide interesting data in some ways.
In general I do not know for certain that the general scientific community possesses the specialization necessary to accurately judge the quality of all published works outside of their field of interest. While I am interested in a decentralized paper publication platform, I do not know that it would be able to effectively curate articles of actual quality, and would focus more on those with more flashy and easy to digest conclusions. In all it is a worth while experiment to see how the results would turnout, however I remain skeptical (as any scientist worth their weight should be, a healthy dose of skepticism keeps work judged fairly).