[michael1999]

I see the reverse.

Self-driving jitneys will be able to extend public transit into the cul-de-sac subdivisions that are impervious to useful bus service. Let little 8 seat shuttles bring people out to the main line for trains or full-size busses to carry away. It is terribly inefficient to run a 50 person bus with 3 people in it, one of whom is a paid driver. See https://www.cbc.ca/news/canada/toronto/autonomous-driverless...

Same for small towns - they just don't have the population or the density to run large busses with drivers on fixed routes at a reasonable frequency. I've seen smaller city busses with routes that only run every 90 minutes. But smaller self-driving vehicles can collect a few people and run them around. See https://www.chandleraz.gov/residents/transportation/transit/... or https://ridewithvia.com/news/waymo-and-via-announce-strategi...

Or try this - Waymo giving a discount for pickups at station. I used to use the Toronto subway this way - take the subway as close as I could get, and then take a taxi the last couple of kilometres. https://waymo.com/blog/2024/10/clean-rides-clear-benefits-wa...

[arjie]

Strongly agree. The future is what we used to call "share autos" in India. It's a small vehicle that can transport some 6-8 people. Currently these aren't feasible for the public because driver operational cost dominates, but once driver operational cost is near zero, we can bring much more public transit online. One of these vehicles is 4x-8x as dense as a car, and with America's relative lack of density, it's the only realistic way to have widespread public transit.

Particularly useful with dynamic dispatch and dynamic routing is that the vehicle's stopping can be much more infrequent, leading to lower total trip time and greater frequency. San Francisco is very dense but is also very old, so large 40-60 person busses stop every block forcing low throughput high latency rides. As a consequence, I prefer an e-bike over a bus everywhere in SF.