Well, this is a nice article but as fellow commentators mentioned before we can’t help but take into account there is no way for autonomous vehicles, should they be shared to replace NYC’s subway downtown. Not ever. Even with shared vehicles. This is, physically-speaking, impossible. Beyond this, for high-traffic public transport like NYC costs typically fall close to 0,05$/passenger/mile, which you wouldn’t achieve easily even in best case of autonomous vehicles.
Your argument is correct however that many public transport subway/light rail/bus lines will be jeopardized by the competition with AVs, because they offer poor-quality of service, low frequency and high costs per passenger. This is especially in the US where you have weak public transport compared to Europe. In that case your scenario might be relevant.
You have actually two conflicting forces fighting each other:
- Cannibalization between AVs and public transport, where the former could be cheaper, more convenient and more efficient
- A nice combination of AVs and PT when the former are used to reach metro/train station and increase traffic of the latter. This makes sense because you can’t all reach city centre by car, and because it makes sense economically. If you add Mohring effect to this you get a nice virtuous circle: more traffic in PT leads leads to higher frequency (to accommodate demand) and better P&L, which in turn increases economic utility of the service (for the users it is more appealing to be offered trains more often), then traffic…
You’re right to say PTAs have to carefully review their long-term investments to make sure they design a relevant mobility system and take advantage of AVs to increase mobility in cities and not to stall these cities with countless AVs!