One SDC takes up slightly less space when we factor in the necessary buffer around human-driven vehicles to account for human reaction times and safety perception.

A fleet of SDCs snaking down a road can "tuck in" tighter to each other because computer reaction times are better (especially if they're networked and sharing decisionmaking signals).

Of course, it's still not as space-efficient as a bus. But the bus infrastructure as a solution has been challenged by COVID-19.

If autonomous vehicles werebto ever become cheap and reliable, I fear induced demand could have a much larger effect on traffic volumes than the role it plays in traffic now when vehicles rely on human occupants. As people find more creative use cases for autonomous vehicles, the roads would fill up with ghost cars performing any number of progressively more trivial tasks until traffic returns to equilibrium.

Quite possible.

The constraining factor on SDC commonality in the short-to-medium term will be maintenance. These cars require dedicated service (which is why they'll likely start as replacements for taxis, not privately-owned; people will have to be hot-rodders to want to do all the work necessary to keep the sensors clean and tuned).

There may be a threshold in the future where the sensors are hyper-rugged or self-maintaining, but it's a little ways off. The maintenance burden will keep the ghost cars down in the short- to medium-term.

Presumably if all cars were self driving, or say 90% were. Then that would become the next "public transportation". And if they were all electric, the footprint would be less harsh as well.

Personal cars will never be the "next public transportation". The density just doesn't work. Personal vehicles are the least space-efficient (to say nothing of energy efficiency) means of transportation among buses, trains, bicycles, and walking. You can't solve congestion issues by replacing a more-dense option with a less-dense one. The only thing self-driving cars solve for is freeing up a person to read or do something else during their commute.

Congestion itself isn't the issue. Wasted time and fossil fuels is. And most of my friends who take public transportation to work have a commute that is longer than my friends who drive to work.

Right now there is a difference in quality of that time in that your attention is more free when taking public transit but self driving cars fix this issue.

It's not public transportation in the same way that taxis aren't solving the public transportation problem. But also, why is it that there is this assumption that when self driving cars become a reality, they'll also be electric? The two problems are completely orthogonal, and while solving both of them are important, making self driving cars electric isn't a priority for most companies today.

> There's also what's known as the "geometry" problem - 1 self driving car takes up as much space as 1 human driven car.

Human-driver cars consume parking spaces at the destination, self-driving cars don't have this limitation. They park themselves away from high-traffic areas or they could ferry other people around.