I don't have an article for you, but my basic criticism is its illusory approach to the Byzantine generals problem. Xrp makes this allusion through talking about byzantine fault tolerance, where your node communicates with others it trusts and agrees with whatever a percentage (like 80%) I believe are saying. On the surface this isn't a terrible model, but when you dig into the design basis of a byzantine fault tolerance model you find that its not intended to protect against deliberately malicious data, only faulty data.

If the criteria for trusted nodes is met than you could receive data that is malicious, and you have no objective truth like POW to say otherwise. XRP people say, "but you have trusted nodes", except the nodes that will manipulate you are probably the oldest and most well connected nodes on the network, and those lies will propagate across the entire network.

I am inclined to think there are useful things to build on the byzantine fault tolerant model, but money isnt one of them.