A community of scientists sharing their papers would be a good thing already now.
I personally know active scientists who don't even try anymore to look up the paper, but rather go directly to sci-hub for any doi they need. I can understand why, but I also think that this doesn't lead to a sustainable publishing culture.
But honestly, how often does one have to skim that many papers in a day, to a level where the freely available abstract is not sufficient?
Perhaps every once in a while when one compiles a survey of a new field they enter. Once the project is set on the rails, one rarely has to read that much.
More often than you might think.
To take an example from my own work, I was doing assay design a while back, and needed to collect all existing primer sets in the literature. I probably went through a hundred papers over a several day period.
Are you sure that the usual suspects don't make authors assign copyright or at least distribution rights? I wouldn't put it past them...
Sites like ResearchGate make this very easy. And often a simple email does the job, too.
Advantages:
* It is legal
* The author gets feedback that someone out there reads their research
* Making direct contact to your peers is a good thing.