I think that the solution is automated rail transit. Being in a dedicated place with lower likelihood of encountering people removes nearl every issue that self-driving cars have. Being automated means that 24/7 schedules are possible. If there are enough trains and high enough saturation, need for cars and even taxis is removed.
One train transports 100s of people, the driver is a fairly low proportion of the cost. And there's other members of staff that are required even in a fully automated system. (network monitoring, security). Removing the driver is a nice step, but it doesn't fundamentally change the economics of rail transport. If a route is uneconomic, that's going to be the case without a driver too.
Removing the driver mainly removes barriers to running late - meaning things like drunk driving can be significantly reduced since transit in the US is virtually non-existant at drunk'o'clock, effectively pressuring people into bad decisions when their judgement is the poorest.
If a route is uneconomic, that's going to be the case without a driver too.
Infrastructure is vital to economic and other activity. It needs to be treated as an investment or necessary cost, not a business. Doing otherwise inevitably results in collapsing bridges, toxic spills, and other symptoms of neglect as corners are cut to maximize profit.
We're in agreement that night trains are a good thing, but you should push for them whether or not your trains are driverless.
You misunderstand my use of economic. Everything has a cost and a benefit which can theoretically be calculated, with infrastructure like transit that benefit extends beyond fares. Typically governments will do this calculation when deciding whether to pursue a new project, they include all the planning, construction, running costs, and externalities e.g. climate impact, and all the benefits from fares, economic activity, new opportunities for industries and development, ect. This produces a cost benefit ratio. In my research with transport, the best value projects are local safety improvements like cycleways, sometimes the ratio is as good as 10. Large public transport projects are maybe 1-2, and large motorways are usually less than 1. My point was a train driver is a small cost that isn't going to significant affect this. Of course, this analysis often gets ignored and the overpriced motorway gets built anyway.
I can imagine them being cheaper and I only would use people to transit other people when you can have 40 people or so. Where security on big vehicles like bus or train need more caution. A person driving a single person feels like a waste of time or smth. Driverless cars could also be more efficient in routing.