That's actually a really interesting question. Because it's not necessarily about handling most usage, but also about handling peak usage. Is it worth the cost to keep everything 4-seater if that means they can all enter "carpool mode" whenever required at times of peak demand?

Because once they become ubiquitous, I suspect the vast majority will be operating in carpool mode at rush hour. Most people won't be willing to pay 4x to get a private vehicle if they're by themselves. Especially since the more vehicles there are, the more efficient carpool mode becomes for everyone.

My impression is that the shared ride options on Lyft/Uber only give a small discount and see little usage. Sharing rides is only more difficult and less attractive for Waymo users, who won't have a third party to buffer interactions.

The price depends on a lot of factors. And there are many factors that can turn it into a large discount. And the larger the fleet is, the less difficult it becomes.

Of course it's less attractive at the same price, but if it's cheaper enough it becomes more attractive for the average rider. And we can even imagine cities implementing single rider surcharges at rush hour to keep traffic running smoothly.

Rush hour will be a bottleneck. So something has to be done, and it will involve trading off price and convenience. Whether it's carpool mode or self-driving buses or likely a combination of the two.

That's really interesting because I hadn't actually thought about that in-depth before. I think Tesla's robotaxi prototype was even a 2-seater. My knee-jerk reaction to your comment was "no, 2 seater won't happen because the incremental cost of the additional seats and doors is immaterial to the overall cost of the car."

But then thinking more about it I thought of how great we (all the people who like Waymo) think it performs around bikes and pedestrians. So now I agree with you directionally but you might not be taking it far enough. Once (if?) autonomous vehicles rule the road, and they're known to be safe, the future will likely be the broad spectrum from autonomous buses (on the large side) to super-cheap, bike-like vehicles (on the small side) that cost way less than a car. For a single occupant, if you knew another vehicle wasn't going to kill you, wouldn't you take an e-bike (with a cover and basket on it?) for short trips if the fee was proportionate to the cost of the vehicle? I would. Assumes lidar shrinks I guess and that automated kickstands are a thing, but that seems tractable in the years to come.

Bike share programs already exist and are pretty popular in NYC. Self driving doesn't really seem necessary at that scale.

Can you expand on that? Why is it that a vehicle being bigger vs smaller is the distinction? I’m one person but I care about some combo of money/time/distance/comfort. If I could get a cheaper ride from say Houston up to the park rather than having to drive it myself the value prop to me is the same—I don’t have to drive. Heck if the seat on the bike was comfortable enough like a laid-back sport seat in a car I might choose the bike at the same price point because it would be more fun (again this assumes I “know” it’s safe).

Sounds like you might enjoy a recumbent bicycle.

For a "bike" to balance you need a certain level of physical/mental engagement. How much value are you really getting from the self driving aspect, even if it were feasible, at that point?

At some point smart-car esque self driving vehicles will likely be possible, but self-driving unenclosed two wheelers? Pointless IMO.

Yeah, you're right, uncle. How about some kind of conveyance that just costs a lot less than a conventional US car? Rhetorical. Not expecting a reply to that.

I don't understand the hostility.

A bike already fits your requirements, self driving would just make it more expensive.

If you want something enclosed, NYC has self-driving (from your perspective) trains and busses.

If you want enclosed and not-shared, I guess you just want a smaller car? Losing wheels wouldn't make it cheaper, IMO, because that makes a bunch of stuff more complicated.

I figure the real future will be 2-3x 2-seater separate soundproof bubble domes per car.

This makes lots of sense to me. A 2-seater is often a hard sell for a private owner (even one with no kids), but I'd bet the majority (or at least a plurality) of taxi/ride-share trips are for one or two people.

Or a 4-seater with two rear facing seats. The design space gets bigger quickly once you no longer have to account for a driver and their field of vision, especially if the cars usually travel at city speeds where aerodynamics don't matter as much.