> Thats enough to cover the entire ridership of SF's public bus system
I suspect this assertion doesn't have the necessary math behind it. There are a lot of tricky things that would need to be modeled. (Also some of the others, but I'm just going to focus on this one for now.)
Just for starters, let's play with napkin-math for peak demand.
The buses serve an average of >500k boardings each weekday. [0]. Make a simplifying assumption that each person boards twice for a round trip, and that half of those 250k round-trips occur during rush hour periods, and that's 125k people all trying to get to work in the morning rush.
So each taxi needs to satisfy eight people's to-work commutes during that time. That seems unlikely unless all of the commutes are actually very easy, everybody somehow ends up in convenient parties of four, or a lot of people are commuting against the prevailing flow so that the vehicle is always cycling productively, etc.
I might be wrong, but someone who doesn't even start to address those questions can't say they're right.
> each taxi needs to satisfy eight people's to-work commutes during that time
Then the solution might be 7-seater Waymos that operate in shared mode plus standard Waymos that operate as private rides.
Nothing beats the density of grade-separated urban rail. But it's fair to question if buses can be made better by making them autonomous, smaller and point to point.
If the numbers work this would be fantastic and might offer Waymo as well as Tesla a possible future business model. I know this would be a popular choice here in Michigan assuming of course they could figure out how to make them run in ice and snow ;<).
The most frequent gotcha with this kind of arithmetic is treating the streets as if they were already there, free of charge, whereas the subway tunnels, bridges and rails all enter explicitly into the grand total.
Buses are often the cheapest option to implement public transport in places that have only cars for obvious reasons. However, buses are also not very fast and, additionally, get stuck in the daily traffic jam together with everybody else unless you dedicate space to them (bus lanes).
When I lived in Taipei in the 90s they had just implemented a scheme where they would dedicate each vein of the urban grid to traffic in a single direction. Because of the way the very broad streets were built, they had a dedicated bus lane to downtown in a street dedicated to outgoing traffic. They did not see the necessity for a bus lane in the opposite direction, meaning in the morning I had a commute not unlike a Star Trek transporter, while returning home meant agonizing hours of being stuck.
I suspect this assertion doesn't have the necessary math behind it. There are a lot of tricky things that would need to be modeled. (Also some of the others, but I'm just going to focus on this one for now.)
Just for starters, let's play with napkin-math for peak demand.
The buses serve an average of >500k boardings each weekday. [0]. Make a simplifying assumption that each person boards twice for a round trip, and that half of those 250k round-trips occur during rush hour periods, and that's 125k people all trying to get to work in the morning rush.
So each taxi needs to satisfy eight people's to-work commutes during that time. That seems unlikely unless all of the commutes are actually very easy, everybody somehow ends up in convenient parties of four, or a lot of people are commuting against the prevailing flow so that the vehicle is always cycling productively, etc.
I might be wrong, but someone who doesn't even start to address those questions can't say they're right.
[0] https://www.sfmta.com/reports/muni-ridership-average-weekday...
reply