I guess we're close, but I draw a distinction between "the technology doesn't exist" (it does) and "nobody is doing it yet" (agreed, they're not exactly at large scale). 

When you said "mesh networking doesn't exist" you seem to have meant "doing it correctly is technically possible but there is little will or energy going in that direction". We might agree that the goal is to advance freedom technologies (like mesh but also lots of other stuff) but I oppose blasting out pessimistic "doesn't exist, (but yea I guess you could do it)" takes. Getting people aware of and optimistic about the possibilities is a good way to help make sure they actually come about.  

 I get what you're saying, but I still think that there is a somewhat of a hard line here

if decentralized routing was a solved problem don't you think someone would have implemented it already, even as an experiment? 

 urbit's routing is hierarchical but also decentralized. Planets depend on Stars for their routing, stars depend on Galaxies (Planets can spawn Moons). Anyone can own one of these nodes and put them anywhere. I don't like their approach because of it's overly hierarchical nature, but it is an interesting hybrid. 
This is just an example of an "experiment" that maybe you haven't heard of? (Or maybe you have). The world is big and I don't presume to know of all the things people have tried but haven't taken off because of poor marketing or outright idea suppression by powerful centralized forces.  

 I tried to read about it but couldn't find anything useful, doesn't urbit rely on IP already?