I could see it going either way. I actually have become more likely to use transit since ride-hailing is so much easier—if I'm ever stranded anywhere (lack of good bus route, after hours, etc.), I just have to open the Lyft app.

Furthermore, when car ownership is no longer considered a necessity (as it is today in most of the US), the economics will change. Right now, my personal vehicle is a sunk cost, so I view car trips as costing close to zero—unlike the up-front, immediate transactions inherent to taking the bus or Lyft. Once I no longer own that vehicle (I'm not planning to replace it), all of my options will be on more equal footing, since they'll all be priced per-trip. Even self-driving vehicles will not be able to compete with mass transit when they're both priced similarly.

> Even self-driving vehicles will not be able to compete with mass transit when they're both priced similarly.

I think there's likely to end up being a third option: Public transport with self-driving cars for last mile at each end. That way you have the convenience of door-to-door, combined with the efficiency of mass transit for the bulk of traffic.

There are also other advantages, like it expands the effective catchment area of stations, making them more viable even at lower densities. And the cars themselves can be much smaller and more efficient as well, since they don't need the range, and they don't need to be as comfortable if people are only sitting in them for ~5mins.

> I think there's likely to end up being a third option: Public transport with self-driving cars for last mile at each end.

And maybe a fourth one: self-driving road based mass-transit (autonomous buses basically).

Exactly, owning a car in a city can easily cost 25+$/day. That buys a lot of public transportation.

Self driving cars will eventually be much, much less expensive to own and operate than cards of today. That's assuming there isn't a massive per-mile tax imposed.

I would also guess public transportation will evolve as well. I could see small and medium sized cities getting rid of buses, fixed schedules, and fixed routes and instead deploy a fleet of van-sized vehicles to pick up people on demand. I could see this as being something that gets outsourced to Uber and Lyft. Maybe it will be less expensive and better to just subsidize rides from private companies.

Mechanically, self driving cars are not going to be cheaper to build or maintain, and they will need just as much fuel if not significantly more to go park somewhere distant. You also need 2 parking spaces @ 100+$/month (one at home, one at the office) and pay to park other places. Insurance might get slightly cheaper, but I don't see the economics changing much and the added electronics are not going to be free.

Electronics will do what they always do - get cheaper and better.

The cars (eventually) will not be like current cars. If they don't crash, they don't need as many safety features. They can be made much lighter and from cheaper materials. When the car is way lighter, it can use a smaller, simpler drivetrain. They will likely be electric. Maintenance on electric cars is far less than gasoline or diesel vehicles. Your robot car will likely be serviced by another robot.

Don't think of a Lexus with an instrument package bolted to the roof, think about adding a few hundred dollars of sensors and actuators on a fancy golf cart. I wouldn't be surprised if self driving cars are eventually available for less than $10,000. I think the low cost of the vehicles combined with the low cost of operating them will result in an explosion of ownership, not a reduction. I think when the cost of the vehicle drops and the cost of a trip drops, more vehicles will be sold and more trips will be made.

A self driving golf cart is not a self driving car. Until we outlaw humans from driving anything self driving cars will need all their current safety tech. That might happen in 100+ years, but it may also never happen because hitting something due to mechanical / software failure is always going to be possible.

PS: Adding a camera may get cheaper over time, that does not mean it's ever cheaper than not having one in the first place.

100 years? I was thinking more like 50.

Mechanical and software failures will always be possible. However, we already accept a really high rate of death and injury when both could be eliminated. I don't think that will change with self-driving cars.

Adding a camera is cheaper when it results in being able to eliminate hundreds of pounds of metal and airbags. We don't make bus riders wear a seatbelt because it just doesn't solve a huge problem. When the injury rate of car passengers gets to a similar level, we probably don't need seatbelts there either.

We have zero production self driving cars today or in the next 3 years and will not be on 90% of new cars for at least a decade after they show up. Hyundai for example is targeting 2030 for urban driving as an option.

Cars last for ~25 years in normal use (average car in US is 11.5 years old and accidents take many out before reaching their lifespan). So, at a minimum we are talking 50 years before essentially every car is even capable of being self driven. Laws are unlikely to change before self driving goes from a novelty to assumed behavior call it another 25 years at a minimum. Now, add a few more years for phasing in the law, and then a few more for statistics to validate reducing safty equipment.

So, 100+ seems fairly optimistic as that's the happy path assuming zero major issues show up.

You're assuming that the ownership model stays the same. Many people (myself included) are predicting that the convenience and low cost of autonomous ride-hailing services will make car ownership a frivolity, much like owning an additional non-daily-driver is today.

While self driving cars are likely to make ride hailing cheaper, that does not translate to cheaper car ownership.

Of course. But it does translate into cheaper transportation options. Many people may opt out of car ownership once ride hailing reaches a certain level of affordability.

But I get value beyond just point A-B travel out of that.

That's one side of the coin, but on the other hand I can also envision self-driving cars enabling more urban density. Presumably self-driving cars don't really need to be parked on-site at the destination and could instead drop off their rider at the destination and park in some centralized parking garage. The car could even make money for its owner by being an automated uber driver while the owner is working or shopping or whatever. American cities give up a lot of space to parking lots, it'd be great to be able to reclaim some of it

I heard an estimate that something like 25% of traffic in some areas at a given time are just people driving around looking for a parking spot. That's just nuts!

It really remains to be seen what self driving cars would do to traffic.

In the future, the car will just circle the block while you shop.

Sounds like an immense waste of energy, energy that can be better utilized.

The more I hear about self driving cars the more I wish the US government would take investment into public transportation seriously.

I did not mean to imply I don't like the suburbs. I was implying that too many we're built to quickly with the expectation that the economy was going to be different or that it could sustain these new build outs.

Maybe self driving cars will help. But I see that being a thing for those who can already afford to live in the city. These people are going to likely prefer new stock over run down hoods.

Totally! Get in the car, kick back, take a nap, finish some work, do some video calls, watch some TV... long commutes will be so much more tolerable.

sadly I want to think you're wrong but in reality every last remote place will be destroyed by them. Why live in walnut creek when you could commute from yosemite?

Yup, and self driving cars full of one person each who is taking a nap on their longer commute.