[smt88]

1) I actually think self-driving cars are moving faster than expected. Ten years is still a pretty reasonable guess, considering the pace of governments and the auto industry, but we're closer than I thought we'd be by now.

2) Uber will use self-driving cars, probably exclusively as soon as it's legal.

Uber wants to be the internet of the physical world. With the internet, information could travel anywhere with the tap of a finger. Uber wants the same reduction of friction for objects and people.

The problem for Uber is it's already easy to poach drivers from Uber. Drivers don't care if they drive for Lyft or Uber because the payouts are pretty similar. I've met drivers who switched between both.

When cars drive themselves, Uber will have to become even more competitive to stay on top. The cost of building an Uber-like fleet will be no more than buying the hardware.

Remember how hosting companies proliferated in the 90s and early 00s? Until AWS and similar PaaS options, the hosting space was incredibly fragmented, and the margins were awful. That's what Uber has on the horizon.

[marcus_holmes]

Cool thought, that the Last Mile will soon be commodity-priced and driverless cars will be making deliveries when they're not otherwise occupied. Software eating yet another business.

Providing capital to build a replica of an existing proven business model is easy, you already provided the example: gajillions of hosting companies providing physical hardware for peanuts because a few people had proved the business model worked. The cost of the hardware is amortised over its expected lifetime and any bank will provide the loan.

If this is Uber's future then Travis must be looking to exit before that...