And the drunk driving claim depends on the study. [1] Which suggests it requires further study. Not to decide whether it should be a talking point in favor of Uber but to understand why services like Uber decrease drunk driving if, in fact, they do.

Is it purely a price sensitivity thing? Is it a simplicity/convenience thing? Is it just because summoning an Uber became hip among a certain demographic whereas taxis weren't? It would be useful data because there are possibly useful public policy responses that could address why people drive drunk in some cases.

[1] https://www.nytimes.com/2017/04/07/business/uber-drunk-drivi...