Earlier quoted context omitted.
I've had a similar thought regarding the impact of affordable autonomous vehicles, for two reasons. 1) Commuters who can read / game / whittle while they commute will be willing to commute farther, and 2) if widespread adoption of autonomous cars can mitigate the effects of traffic (i.e. enable us to actually drive the speed limit during rush hour, or at least closer to it), we'll be able to commute farther in the sa…
I have my doubts that autonomous cars will do much for traffic. Traffic doesn't necessarily happen because there are too many cars on the freeway, but too many cars on the much more limited capacity streets which back up onto the freeways. Everyone taking a self driving car into work isn't going to change the fact that traffic is bad because everyone insists on traveling 20'x5' apart from one another through a dense…
These are often because of a bottleneck such as a traffic right (particularly left turn signal) or stop sign. If autonomous vehicles can be tweaked to increase throughput (better reaction times, shorter follow distance, less accidents, higher speed through chokepoints like tolls, not distracted or not accelerating slowly on light changes) you might see significant traffic reductions as a result. Little's law is applicable here.
The problems with bus or commuter lines is it's hard to get people to share a route efficiently. I've had it work in the past but only where I picked my housing according to an effective commuter line for my school or work. But that's hard to arrange at scale.
Where effective commuters line don't scale instead if you combine the above wins with an uber pool like service where 3-4 people are efficiently routed together. You could be looking at 2-3x fewer cars on roads that have better throughput and fewer bottlenecks.