I can see it now: "Our self driving cars no longer travel to Toledo because not many users go there."

Actually, your statement is not too far off.

Google's self driving is far less dynamic as they would have you believe. Areas are extremely heavily analysed with a massive array of sensors, analysis and even human input/sanity checking before cars are allowed to drive in a specific area.

No doubt this preparation is expensive and the resources to do it are finite. I could easily see some very accessible normal place be unavailable to a self driving car.

Source: 2 of my coworkers worked at the start-up that originally developed the core self-driving platform that was acquired by Google(No, Google did not develop the origin technology). The both worked for some time at Google working on the project.

Huh. Funny - you can probably come up with a few scenarios where that would happen, like if their large-scale deployment software for self-driving cars is heavily dependent on the wisdom of crowds, for instance.

The reason it was removed was because it was extremely hard to set up and use properly. Glass is seen as a similar device as Chromecast in that it is supposed to be completely plug and play where even the most basic users will be able to figure it out with minor instruction from a Guide. Since Hangouts was unable to work with this experience, it was not seen as adding value to the device.