If we can pull off *genuinely decentralized* curation of simple lists, I think everything else falls into place.
Consider that two lists is all it takes to specify a graph: one list for nodes, one list for edges. If we can crowdsource graphs, we can crowdsource data models.
Want to design a verified credential for application XYZ but don’t want to depend on W3C to do it for you? DCoSL it!