If it dies, it will be via dependence on meat space. There's only so much of moving atoms that can be digitally replaced and the core business is a commodity: it could compete with taxis because rides between A and B are mostly fungible and being better than existing taxi service was not a very high bar and being better than standing in the rain waving your arm was not a high bar either.
I'm sure it's software engineering is really good. But over the long term, it can only provide marginal advantages to a business that boils down to personal services.
[1]: https://www.bloomberg.com/news/articles/2016-12-20/uber-s-lo...
https://www.wired.com/2016/10/ubers-self-driving-truck-makes...
This one is not Uber, but you get the idea.
https://arstechnica.com/cars/2017/02/gm-lyft-could-deploy-th...
more like little over 1 year [0]. Originally wanted to say 2 years, didn't know loss climbed to $3B last year.
[0]https://techcrunch.com/2016/12/21/uber-losses-expected-to-hi...
Sometimes profit and social goals happen to align.