"Why would you eat an orange when you could drink a coffee instead?" Yes there are heavy mathematics behind the structure of both types of network, and it's tr

"Why would you eat an orange when you could drink a coffee instead?" Yes there are heavy mathematics behind the structure of both types of network, and it's true that without centralization the network itself can't reach a certain limit. But this is a double-edged sword, this is only true to the whole system's perspective and it's completely different on the node (user) level graph-theoritically. Also, the article forces certain needs on the users. There are lots of us who DON'T want to magically find anyone or be found magically with a keyword, nor interested in "trends" - this is heavily misleading for we still don't have proper metrics for it and won't have one for a while because the problem implies building fault-tolerant reputation networks (that are decentralized by the way)