In Seattle, transit ridership is up significantly because of grade-separated light rail: http://seattletransitblog.com/2016/07/16/link-ridership-up-8....
Urban environments with well planned transit corridors make useful service far more likely.
Suburban environments make 'coverage' more likely, but not effective coverage or coverage that is useful for those that would need it the most (anyone unable to afford personal transport).
Rural environments are just shafted; there can't even be the pretense of economically viable coverage.
I can't help but imagine how self-driving cars would change the above; when you can have an automated taxi take you directly from point A to point B, or maybe share one or two stops at most with carpoolers along the way.
Well, he's written about that:
http://humantransit.org/2016/07/elon-musk-doesnt-understand-...
However, it seems very clear that fixed-route transit service isn't optimal, and that technology will help us arrive at something better. People want to get where they're going as fast as possible, and each stop along a transit route makes it less attractive. Ridehailing networks create a real-time market for transportation that will make it easy to optimize each person's travel time and costs. Some people will end up with what looks like fixed-route bus service, just with dynamically generated routes instead. Many will end up with more direct service with few stops, especially if they're traveling between dense neighborhoods we've allowed to thrive by ending restrictions that prevent them.
The key to making that future work is ending the crazy practice of allowing our roads to become congested. We know how to turn off traffic. We just don't do it. Charge fees that increase until traffic flows freely. Rich people get direct travel with more personal space. People with less money share space and travel with more stops. No one waits pointlessly in traffic.
Plus, there are larger apartment/condo projects going in near both Hayward Park and Hillsdale stations.