Don't know much about the ride share industry, but am interested and curious. Which parts of this type of app do you think would be particularly difficult to scale?

Spatially-oriented, locally-scoped applications pretty naturally shard out by region.

The way its built requires no human operational cost.

I'd be skeptical if ridesharing is possible without some human operational involvement. The most complex part is not the 99% of trips that go right, but the 1% that go wrong. The consequences of failure are more severe than say, a SaaS app not working correctly.

It's not like Uber does a good job here.