Urban space were divided by several main streets. When the city grows, locations around the main streets would be preferred, and the rest "back streets" became deserted. Because nobody wants to lose the direct access to the commute route, not for the residency or commercial. Or, because we don't have the power yet.

In a few years, the autonomous cars will help us commute in the city. We let the machines to take care of navigating us in this complex urban.

This changes the way we use urban space: with help of autonomous car, we can easily have access to any streets in town. Back street and main street makes no difference. Every street becomes main street now.
It's like shifting from Direct-Access to Random-Memory-Access.
Or, Random-Space-Access.

All we have to do is to design an interface of navigation, to help people forget the complexity of urban space, and focus on our personal locations stored in the cloud. Forget the public map with spider-web-llike streets, and let the algorithm sort out a "My Map", or "Mype".

Detail:
Unlike a plain list of all locations, Mype has a distinctive animation to sort out all locations in one axel. Keeping the basic way of how we understand a city from standing in a street, namely direction down the street or up the street.
Best or all, I discovered that, people remember a location better with the front door photo than a point on the map. So the selection menu is also based on the door photo.

Conclusion:
By re-design the way we understand the urban space in a simple, abstract yet familiar way, I sure we will be able to make use of the urban space in a new level.