From what I understand, this isn't about replacing faster centralized servers with a slower centralized approach for all files. It's more about creating a situation where if a file were to be censored by any server, you could still access the file