The best part about self-driving cars

Here’s that Google self-driving cars promo video that’s been making rounds lately:

This is obviously completely world-changing.

The thing is, though, this video and its surrounding coverage in the press don’t even begin to explore the best, most exciting externalities of self-driving cars. They get to, maybe, level 2 out of 7.

So I thought I’d lay out some ideas.

Level 1: People who couldn’t use cars before can now use cars. This is wonderful for the disabled and elderly. And maybe even for kids. This is most of what you see in the video — and these groups of people have every right to be excited. But (and I’m definitely wearing my executive business hat here) the real economic impact of this change is relatively not that big a deal.

Level 2: Instead of driving, you can do other things behind the wheel. (Well, “wheel.”) This is the part that the angry internet masses presume Google is probably most excited about. The worry is that when Google says this affords an opportunity to “catch up on email”[ref]For the record, by the way: I absolutely despise this phrase.[/ref] they really mostly mean it’s an opportunity to be served ads or create data for Google to serve ads.

Personally, I absolutely love the fact that being behind the wheel is one of the last few places where you’re forced away from media and left to your own thoughts. The thought of losing that saddens me. I’m not as worried about malicious ads being served. All the modern soccer mom vehicles have TVs in the backseats so that the kids can stay abreast of the latest happenings with Spongebob, and that seems to bother nobody.

I’m a little curious to think about what the most productive & interesting replacement activity would be instead of navigating a steering wheel and pedals, but again, in a large-scale economic sense, this is small potatoes.

Level 3: Everybody switches to self-driving cars. This seems inevitable to me. Once the technology establishes a base level of safety and reliability — it struggles with snow and rain at present, but I can’t imagine those issues will last forever — then the value proposition of a self-driving car is just massively better than a traditional vehicle.

Put it this way: A full-time chauffeur typically costs $50,000-$80,000/year. So you could argue that — at least for the wealthy elite — the self-driving car is worth something like $200k in savings over a 4-6 year use of the vehicle.

That value won’t be the same for lower classes who don’t value the chauffeur (or, the opportunity cost of their free time) as highly, but it’ll still be substantial.

Economically: We lose the traditional chauffeur industry, but that’s not so bad. The car industry’s shaken up, but not gone. It might even boon.

Level 4: Suburbs are more accessible. This could just happen anyway as the internet and digital collaboration get better. (See: Oculus Rift. Read: Snow Crash, Ready Player One, and so on.) Many people avoid the suburbs because of the immense drag of hours of commuting to work in the city or elsewhere. But what if you can work (or even just sleep) on your way to the office?

What happens when the suburbs aren’t just for married couples who want the space to raise a family and can trade off a few hours’ commute time? Do young people move out to take advantage of the substantially cheaper real estate? Does the urban sprawl balloon?

Level 5: The roads are insanely safe. I anticipate that we’ll eventually see the tides turn completely on the perceived safety of driving. It’ll take several years.

Today, self-driving cars are unknown so they seem dangerous. We feel like we can trust human drivers to behave a certain way.

In the future, computer drivers are a sure bet. The faults and lapses of the human mind and senses will still be unknown, so driving a vehicle manually will likely be the outlier and the threat.

An interesting comparable: You can power up the latest Grand Theft Auto game and just start driving around the city. What you’ll find after a while is that the only car that ever gets in an accidents…is the one controlled you, the human.

Once perfectly coherent computers are responsible for manning the heavy majority of vehicles, I’d expect accidents would drop precipitously. I also can’t fathom self-driving cars racking up many speeding tickets.

Here we really start to face some stranger side-effects. What happens to hospitals who no longer have the patient inflow from auto mishaps? What happens to local law enforcement who no longer have the monetary inflow from traffic tickets? Do we need fewer doctors and policemen? Or do we just get better and more aggressive at treating and policing other things?

Level 6: There’s no need to buy a car. A few more years down.

Here’s a big, universal secret that pretty much nobody talks about: Owning a car actually sucks. Even putting aside the fact that it’s already by far the most dangerous thing anyone does on a regular basis (we solved that at Level 4). Your car is incredibly expensive to buy, it’s incredibly expensive to insure, it’s incredibly expensive to use (gas), it’s incredibly expensive to maintain (repairs). It loses value precipitously the second you drive it off the lot. There’s an infinity of teensy tiny nuanced issues (“Has that scratch always been there?”) that quietly add stress to daily life.

Still — today in 2014 (and in years past), that’s a small price to pay for the freedom of being able to say “I need to be 10 miles away from here, right now.”

In the future? Look at what Uber has already done with its on-demand car service. I’m confident their company gets better and better every day at understanding supply and demand — knowing where their cars need to be and when. With self-driving cars, a future company like Uber cuts its most expensive operating cost out of the equation (see: Chauffeurs) and could realistically have an army of cars always standing by. You might realistically be able to schedule regular pickups and routes for yourself. Because you don’t have to pay for a driver, and because you don’t have to pay for the car’s idle time(see: sidebar[ref]Important, relevant side-note: For most of the 24-hour day, Your car’s pretty much just sitting there. Which is the idea driving this whole sharing economy in the first place. You’re paying lots of money to own a car 24 hours a day, and you’re really only actively using it for maybe 60-120 minutes.[/ref]), I’d imagine this could drive individual trip prices down to meet other transportation services where you don’t pay for individual drivers or idle time — you know, like the bus and the subway.

So why own a car when for all intents and purposes it’s just as convenient, and probably cheaper, to rely on hailing cabs which perform all the same functions but cheaper and with less hassle and maintenance?

Also at this point in the future, the symbolic statement of buying and owning a car is all but gone. Getting a car when I turned 17 was a huge deal — it was my freedom and my coming of age. With self-driving cars… who cares? Youths in this future year 20XX have had the freedom to go wherever they wanted since the day they got their first smartphone.

So now the car industry is toasted. It’ll be interesting for the advertising industry too. Car commercials apparently account for 25% of TV revenue. Maybe some day, not everyone will have “15 minutes could save you 15% or more on car insurance” beaten mercilessly into their skulls.

Level 7: Real Estate. This part’s deep, deep future. Perhaps accordingly, it’s definitely my favorite.

Today, we use an ABSOLUTELY, POSITIVELY RIDICULOUS amount of real estate to just let cars sit idly and do nothing.

New York City has 1,100 parking garages and around 4 million metered and unmetered curbside parking spaces. That’s unfathomable. Think about how incredibly selfish it is to hoard all that space just for the sake of having a car waiting nearby.

Our self-driving taxis, however, should never have to sit idle. Like present-day cabs, subway trains, and the like, they’re either always driving, or if they need to be stationary, they’re parked in far-away lots where it’s cheap and not an inconvenience.

All that means we eventually really don’t need parking spaces. Which means for NYC alone, at roughly 150 sq. ft (19 long by 8 wide) per space, the army of self-driving cars has the potential to unlock over 600,000,000 sq ft of land, or 21 square miles. For perspective: Central Park only (“only”) takes up 1.3 square miles.

What’s fascinating about this challenge, though, is that eliminating all the parking spaces in NY isn’t the same as adding 16 new Central Parks, because these spaces are all distributed on the sides of streets throughout the city. So the problem for the future: What’s the best public or private use of a parking space? Of a parking garage? Maybe it’s housing. Maybe it’s some kind of energy production like solar. Probably it’s something just something else we haven’t even begun to dream of.


  1. Regarding Level 7, my guess is: a lot more downtown pedestrian-only streets, parking garages replaced by housing/commercial space, and for individual parking spaces, parklets. (

    I’m really curious to see what happens with suburbs / sprawl. Any predictions for how that will change?


Leave a Comment.