[jprete]

It's a challenging problem. The whole of Web 2.0 is built on lowering scaling costs as far as possible. The trust problem is itself very big - if someone makes the modern equivalent of the Yahoo directory, how do they establish trust that they're only incentivized by site quality? If it's a graph of trust, how does one handle the subtle changes in baseline expectations and norms that qualify as "trustworthy", when moving from one part of that graph to a distant part? If such a system goes massive, how does one prevent the member nodes from being overwhelmed by requests, or even from offers to contribute by people more interested in Internet karma than giving quality contributions?

It's a tough nut to crack.