Self-driving cars

I suspect that in a while we’ll just call self-driving cars by their name today; cars.

There’s an idea out there that The American Family will buy a self-driving car to replace their old boring human-driven car. I don’t think it will happen that way.

First, we already have self-driving cars available to most people. They’re called taxis. Government restrictions aside, a taxi costs roughly what a self-driving car does. You need to buy an old Crown Victoria and pay someone $50k a year. Let’s say you buy one every three years, that’s $160k ($10k for the old cop car). That’s roughly what a self-driving car will cost soon and it too will last about three years.

We’re not all excited about running out and buying Crown Vics. Instead we rent them for short periods. And the rent is going to come down one or two orders of magnitude with car sharing services and their new ride-along sharing services.

Second there’s a nice reduction argument for how many self-driving cars there will be. Since everyone will have one, I won’t need one. If I want to visit my friend then they can just send their self-driving car to pick me up. Maybe I buy beers along the way as a thank-you. Therefore you won’t need one either, and neither will most people. The number of cars will drop off a cliff.

You’re not going to buy a self-driving car.

Instead every service that you use will offer them. Want to go to Pizza Hut? They’ll send a pizza car to pick you up, so long as you buy a couple of all-you-can-eat tickets. Want to go drink at Hooters? The free car will pick you up and take you home. Need some groceries? Safeway will have a car at your door in 10 minutes.

If the number of cars on the road drops off a cliff then the utilization has to rise. And that’s a good thing. Your car today spends 90% of its life doing nothing sitting in a garage or rusting in the rain. Self-driving cars will be the other way around and spend most of their time being used. And that means it’s ok if they’re dramatically more expensive. Like, $200,000 or $2M a pop.

When you push the sunk cost in to that range then suddenly electric cars with huge batteries are not an issue. And that’s useful since electric cars have fewer moving parts to break down, and all the pollution happens far away instead of being emitted by the vehicle.

The conundrum is the commute. The daily mass migration from home to office. Or rather it’s the inverse. It’s the down time in between commutes. Today we just park our cars for that down time but it would be nice if we didn’t commute, staggered the commute across the day or some other nice solution. Trains are horrifically inefficient, especially when the government runs them. But perhaps privatized trains with last-mile self-driving car solutions would work out.

One Response to Self-driving cars

  1. Frankie Roberto August 19, 2014 at 2:31 pm #

    I was with you up until the point about publicly run railway services. What’s your justification behind them being ‘inefficient’, and by what measure are you considering efficiency?

    Most evidence seems to suggest that railways, like roads, are a natural monopoly, and can only be provided by some level of governmental control. There are few very fully privatised railways! In the UK, the privatised railways haven’t been a great success.

Powered by WordPress. Designed by WooThemes