I'm granting that miles driven or number of cars on the road could go up, maybe way up. I'm just guessing, subject to actually doing the math, that increased efficiency and assumptions of reasonable limits on how much latent demand is actually out there could be sufficient to make time spent in cars, fuel consumption, emissions, or other metrics better regardless. Since we're using networking analogies, basically I'm guessing that current uncoordinated human-driven traffic is copper wire to an automated system's fiber-optic cable.
But again, it's just a guess, and I'd be the first to accept that I was wrong if the numbers didn't actually work out.
But again, it's just a guess, and I'd be the first to accept that I was wrong if the numbers didn't actually work out.