Also, the リニア will cut travel time between Tokyo and Nagoya from about 100 minutes to 40 minutes, which is a big difference for people who do business between those two regions.
The good news is that it's "just" an information problem (well, so is landing a rocket, in a way), the bad news is that the number of possible endpoint pairs in a network is bloody high.
In the end you get back to the old vision of on demand "pods". I suppose that those could still be made "rail-efficient" if you make them robotically jack into a paceline behind a tender that provides drafting and power grabbed from an overhead line (you would not want this per pod), but run on battery power only to get to/from the tender at endpoints (which the tender and its followers passes non-stop) and when the paceline gets broken by someone exiting. If you can start over on a new right of way, this is what you should do (in a pinch, you could also take over the leftmost lane of existing roads, but that's not a good place for doing construction and maintenance).
There tend to be at least local and express services on almost all routes, and often multiple tiers of express. Another common pattern I've seen is having several commuter services which each run stopping for a different section of the route, and then run express to the central stations. This gives passengers the benefits of swapping between local and express services, without them loosing time changing trains.
Of course, all of this is a lot easier when you already have the passenger numbers to support it.