[taxicabjesus]

It takes at least a 1/2 hour to get back to base. Your cleaning staff can't be waiting around all day doing nothing, they need things to do. So the dirty car sits idle until the staff gets to it. Then it takes another 1/2 hour to get back to where the hoity-toity crowd likes to party.

If someone pukes in the car, it's out of service for the rest of the day.

[dragonwriter]

> Your cleaning staff can't be waiting around all day doing nothing, they need things to do.

The service staff may have more to do then cleaning, and even without major events like "someone puked in the back", a large fleet is going to have a continuous, ongoing level of cleaning and care needs (much of which, while needed on an ongoing basis, can easily be deferred when a puking, etc., incident occurs to focus on that.)

So its quite likely that there will, at least in major markets, be work for a 24/7 service staff that also handles puke-like incidents.

[jonwachob91]

Or, I could just drive the cleaning person to the vehicle in question. I'd still have those 30 minutes of downtime + cleaning time, but no time returning the car to action. So maybe I lose an hour a day per cleaning (and maybe the car should be cleaned once a day regardless, nice way to keep the cleaning team engaged) BUT a self-driving car isn't a person. It can work 24 hours a day, so I'd still get 23 hours out of it.

[Retric]

That reduces the throughput of the cleaning staff and costs you 1+ cars moving them around anyway.

Also, from a capital perspective the utilization from 2-5 AM is going to be terrible. Taking peak demand into consideration you probably average ~6-15 hours per day per car anyway.

[serge2k]

You need to charge them too.

[kamaal]

>>Your cleaning staff can't be waiting around all day doing nothing, they need things to do.

By the time full scale driving AI arrives. I'm sure we will have decently capable robots who can clean this kind of thing.