That post is missing the network effects. Put another way switching costs are high not because it is easy to bootstrap these but the fact that to switch, your entire social circle would need to switch around the same time.

It is no use half switching, if you have an app which has 50% of your friends and another with 100% which will you use? * that for every person and your retention drops and you end up with a feedback loop which is negative(churn) instead of positive (growth).

Now with Uber the switching costs are low (currently), you can just as easily request a lyft and drivers can use both apps.

The long term play of uber though will make it difficult for others to compete since network effects will kick in. If you have more passengers and drivers, it makes it that much easier for you to do real ridesharing I. E 2-3 people per a car driving down costs. Once you add additional revenue streams like last mile delivery + passengers + ondemand x and I can see why they are investing this much in owning the market.

I thought the whole point (of the post) was the network effects can effectively be bootstrapped from the mobile phone book.

if all drivers (who aren't uber employees) and all users can be part of multiple networks (bootstrapped from mobile phonebook and almost zero cost to both) how can uber maintain a dominant position?