Why do you believe the network effect would prevail when self-driving transport is available?
When I want to get from point A to point B, it doesn't really matter what my friends or other people are using, I just personally want good service at a good price - and I don't think I'd be particularly loyal to any one provider.
I expect that smaller regional providers of self-driving taxis would be able to compete on even footing with larger companies, once self-driving vehicles are purchasable commodities without heavy R&D costs.
What most people want is a car to arrive very quickly (perhaps in seconds) of predictable quality for the best fee. The largest operators with the best software and economy of scale will win this game. You won't be able to automate a single car and compete, not the least of the reason being users aren't going to jump through 15 apps to request a car, and big players will not hand you their customers.
If Uber can operate 100,000 robo-cars in San Fran you can walk up to the sidewalk and press Hail and presto you have your ride. They can dynamically balance load between nearby cities on a day-by-day even hourly basis dispatching streams of robo-cars from their garages when the surges hit. Large events which drive demand are pre-scheduled and automatically reflected in both car distribution and the number of gasoline futures they trade on the options board for the upcoming week, etc. etc.
You can bet these robo-cars will be getting robo-tire-rotations and any other effeciencies they can squeeze out. (Imagine a line of them going through Uber's robo-washers)
5? 10? 15 years out? Doesn't really matter but I'm convinced it will eventually happen. It will be awesome and terrible all at once.
When I want to get from point A to point B, it doesn't really matter what my friends or other people are using, I just personally want good service at a good price - and I don't think I'd be particularly loyal to any one provider.
I expect that smaller regional providers of self-driving taxis would be able to compete on even footing with larger companies, once self-driving vehicles are purchasable commodities without heavy R&D costs.