Drivers are interested in those features because it makes them more efficient. And having a critical mass of drivers is what makes it possible to get a ride in a few minutes. There are other upstarts, but they don't have many drivers, and your potential user market doesn't scale linearly with drivers because nobody wants to wait 30 minutes to get a ride (even with crazy discounts).
Amazon was far from a dominant player in 2003, and AWS wasn't launched publicly until 2006.
From a product standpoint, as others have stated, Uber is a real-time services marketplace vs Amazon which is more about physical goods (again, excluding AWS, which is technically a service). Most of their value is putting all the work into the ground to keep the marketplace balanced, which is a tricky marketing and econometrics problem. One need not look farther than Lyft to see how hard it is to keep the "5 minutes away or less" guarantee.
Also to those who think the app is a non-trivial technical achievement, I would recommend reading some of the blog posts that go into some of the crazy technical challenges they hit [1]. Specifically in some cases, in order to make the app work in all geo's, they ran up against practical limits to binary size at Apple. Not to mention that geo / waypoint data is a genuine "big-data" problem and not easily reproduced by just any company.
[1] https://blog.pragmaticengineer.com/uber-app-rewrite-yolo/
However, I would not want to own a lot of Uber shares with the way FSD is progressing since v12.