> does anyone really believe anymore that it will be cheap, given how expensive ride share services are?

I used to own a car in SF, and I definitely spend way less money every month on Lyft + the occasional weekend Getaround than I did on insurance/gas/repairs/the occasional tow or parking ticket when I had my car (not even counting initial acquisition).

Obviously depends on a lot of factors, but it’s not that clear cut.

If robo-taxis actually manage to reduce the number of people that own cars, it also has other advantages - mainly less need for parking spaces.

A robo-taxi driving 24/7 replaces 24 cars driving 1h/day, 23 of which are parked on average.

This would be a huge improvement in resource efficiency.

Only if there weren't rush hours.

Yes, it was an overreach; we won't get 24x in utilization, but we could pretty easily get a few times. And the cars don't need to then park where people live.

Sure, human movement is unevenly distributed, but a robo-taxi can also be used for deliveries during non rush hours.

There is also probably 1 hour of service per day needed.

You won't get 24x efficiency, but 15x is good enough.

I'd assume so?

The biggest expense has to be the drivers time right?

Maybe--but probably not to the degree that everything else is almost too cheap to meter. I think the IRS rate is something like $0.67 per mile these days which is, while arguably high as a marginal cost, probably a pretty good place to serve as a baseline for self-driving costs (assuming the cost of the vehicle isn't significantly higher).

I mean not really. I would imagine most public transit journeys are less than a few miles in the most part, so at $.67/mi you are going to see public transit become very poor in comparison to a bus or train fare.

>I would imagine most public transit journeys are less than a few miles in the most part

That's because local public transit is mostly not viable unless you have densities that mean most trips are only a few miles and pickups/dropoffs are nearby start and destination points.

If I were to do the 30 minute drive into my nominal office (which is a very typical commute outside of cities), you're looking at a good $25/day commute cost (ignoring any deadheading to get a car to pick me up on either end). People generally ignore the cost of routine driving.

The exciting part to me would be the cost. If about 50% or taxi fares to to the driver, add some more expenses for more capital in the car and other measures you're looking at a totally different price point where it becomes more economical for things like commuting. And there would eventually be other efficiencies when cars are more designed as driverless that could theoretically drive the cost even lower.

Actually, this is the biggest problem and what government needs to be focussing more on.

If the price drops 50% (or more) then you will have a huge surge in demand for cars. This will cause much more congestion than we currently have which will become completely intolerable.

Really we need to see some tax applied to miles driven by all vehicles which pushes this up and suppress demand - maybe it could only apply to robotaxis at first (would probably be much more palatable politically as well). This could then be used to spend on public transit, so you don't see a huge dropoff in ridership of public transit to robotaxis.