I've used Mendeley (and use papers.app when I'm on a mac) and it has helped me more with organizing papers than in filtering them. What you see as an academic in the morning is a 100 papers staring you in the face and you fall back on pattern recognition at points if you don't have more than a few minutes. Whose names and institutions do you recognize? Which titles seem most similar to your line of research? But you can miss a lot of interesting things and not view things tangentially related, but possibly useful as often. Journal clubs attempt to mediate this by having a group of people each pick an article and bring it to the table but crowdsourcing this and having cross lab "bringing to the table" would be very useful.

Yes, this sounds a lot like Mendely http://www.mendeley.com/

Yes, it sounds like he wants Mendeley