Does it really make sense that Google would build self-driving cars without also thinking about ride sharing technology and userbase? I think the primary use case of Google's cars will be optimal ride sharing, doesn't everyone think this as well?
Sure, their MO is to try things out and ditch the ones that don't work - and this one may not work. But if it doesn't, I'd bet that they'd try again. There's a perfect fit between self-driving technology and ride sharing (hence Uber's interest in self-driving cars).