I think that is massively trivializing the technical complexity of engineering a system that powers millions of rides, food orders, freight orders, etc globally every single day.

As an example, look at how they've improved upon GPS in urban areas to take into account satellite visibility to determine which side of the street you're on: https://eng.uber.com/rethinking-gps/

They invest tons into just solving small problems like that. Not to mention all the open source work they do, ATG, etc.

Then consider that Uber has to have pretty big operations, sales, marketing, and support teams in each market they operate in. It's not just a handful of engineers working in SF. Their massive costs considering their scale makes total sense to me. This is not to conclude one way or another whether or not they are a viable business.