Curbed is ... not the best experience I've had in ride-hailing. To be honest, it's easier and more convenient for me to have the office call a cab for those relatively few times when I'm in Austin. They may have ironed out some of the quirks of their platform since then (scheduled cabs being called out immediately, an attempt to get a ride to the airport being denied because the app decided I was trying to be dropped off on the tarmac itself), but there's no killer reason yet for me to try it out again.

Having said that, it's definitely true that there are basically no meaningful barriers to entry beyond mindshare (after all, every Uber driver is also running Lyft, and likewise for the drivers -- and Lyft's better driver rates are also why I have to use Lyft when I leave for the airport at 3 AM).

Things like UberEATS may provide some stickiness, but I'm far from sold that their basic economics are going to work in the long run, let alone keep them in the market pole position.

I'm pretty sure UberEATS must be a horrible customer experience based on the random shit I see happening at registers in places that I get food from. Stuff like the driver having no idea what the customer has ordered and just taking any old order the people behind the cash register are willing to give them. Since both ends of this transaction don't qualify as the sharpest tools in the shed, it's basically the blind leading the blind.

I get that it makes sense from Uber's point of view -- increasing driver utilization and smoothing out the supply side of the relationship is probably great for their unit economics. But they actually have to execute it well if they expect people like me to use it.

I use UberEATS multiple times a week and every time experience is very good. They deliver fast, no tipping and so far all orders were correct.

> At some point, they have to raise prices.

They don't have to raise prices, they have to cut costs. This will be quite straightforward once there's no driver to pay.

Will it? Setting aside the horserace between spending VC money to subsidize around 60% of every trip vs. the legal framework and a viable self-driving car that's for sale in large numbers, there are some real costs for Uber without drivers. For one, they need to buy the cars, which the driver does now. And maintain them - what's the TCO of an autonomous car? Well, nobody knows. Then of course there's fleet management, which Uber has zero experience in. And don't forget, Tesla and Ford have publicly indicated that they're baking sharing features into their own plans for self-driving cars, which means the platform owners will have the upper hand, while Uber plays with their little app. If that's straightforward, I'd hate to see complex :)

I think you've hit on the underlying issue with the common belief that Uber will be saved by self-driving cars.

If I can produce as many self-driving cars as I want (i.e. I'm Ford), what is my incentive to sell them to you?

I can make more money by owning them myself. Remember also, my cost to make is less than your cost to buy. Then I just tell them to drive off the assembly line and start picking up fares.

If self-driving cars are really a panacea, then the car manufacturers will be the ones who profit. It will also lead to extreme consolidation amongst low end manufacturers, since those with the best self-driving systems will make a lot more money per car from self-ownership than regular cars sold to customers can earn over their lifetimes.

You have to park them somewhere, you have to fuel or charge them, you have to clean them, and you have to maintain them. That means infrastructure. Expensive infrastructure.

Self-driving cars as a service have all the problems of personal rapid transit except track cost. Personal rapid transit has never been very successful. There are at least four startups selling driverless PRT shuttles right now, and nobody has more than one demo installation. Still, shuttle buses ought to be driverless very soon.

(Olli, from Local Motors, is very cute.[1] It talks to its riders, and using IBM Watson, can advise on restaurants and points of interest.)

[1] https://www.youtube.com/watch?v=9joEsWiYFEI

Well, you don't actually ever have to park them. They could run near constantly, or any schedule that makes the most return on investment. Self-driving cars, if taken to their logical extreme, result in total gridlock.

I do think that between Ford and Uber, Ford is much better equipped to handle maintenance of a giant fleet of cars. Uber has zero experience in this area, whereas Ford has a giant, distributed network of dealers. If cars are predominantly owned by the company rather than being sold to consumers, these dealers might be in a position to pivot towards fleet management rather than sales and service.

However, I don't think any of this will happen for the simple fact that self-driving cars are not as close as people currently think. Maybe trucks and buses, but even then I remain skeptical.