Several different problems. First, we need an architectural design that folks can actually agree on. Then we need open protocols. Then we need implementations.

Several different problems. First, we need an architectural design that folks can actually agree on. Then we need open protocols. Then we need implementations. Then we need infrastructure. Then we need to migrate existing content. Finally we need network effects. If we ever get there, I suspect it'll be through a combo between a decentralized registry core and federated and commercial gateway system (similar to VPN or usenet providers). The biggest challenges are related to identity, moderation and abuse, while appealing to a mainstream audience. "It's not the format, it's the business model."