Presumably we would get better at building more tunnels after the first one, and presumably once an existence proof is in place the economics will justify accel

Presumably we would get better at building more tunnels after the first one, and presumably once an existence proof is in place the economics will justify accelerated R&D, resulting in a technological feedback loop. You can only layer so many roads above ground. But as said in the post, tunnels have no such limit. 100 years from now one could imagine a very thick subterranean transportation network. The interchange and routing problems I'd guess are solvable iteratively over the course of decades, since you have a lot of free space to work in with minimal constraints once you are deep enough to allow "forking" the network as you build more advanced technology. Slapping two freeways together is not strategy that lands you ultimately in that future, but starting on the problem building cheaper tunnels is.