LN doesn't 'need' a decentralized routing solution to work. Someone is going to figure it out eventually, that's not the challange. LN is still a work in project. This is not what you'd call vaporware (if you weren't biased):
Lightning Network
Elements Project - Lightning

Two independent implementations (there are probably more, but these are the most advanced ones yet) of the same idea, but they're working together (plan to agree on protocol level). I've already stated this in the past: Anyone could code up LN if they had the skills to do so.


Their simplistic thinking and bias is what is causing problems these days (toxicity towards people trying to improve the infrastructure). They don't even know that the most simple concepts are quite hard in CS (I could give examples of cache/paging calculations). I can't find a good reason for which people would think that the development of previously undeveloped software/features is easy?