This is the only weak point in your article. You need to address the fact that most people opt for Lyft Line or Uber Pool, which isn’t an individual car per person, it’s a car per 2 or 3 people.
You could certainly argue that a full bus is still more efficient, however what about all those times you ride the bus with only a few other people onboard? In those cases, a single car is more efficient (both in terms of space & emissions). And with self-driving cars, the situation will be similar. People will be incentivized to share rides because it will be much cheaper, and riding with strangers typically ranges from “not bad” to “actually pretty fun.”
You have to weigh to space & emissions savings of collapsing all the single-car drivers out there (ever seen the Bay Bridge Toll Plaza at rush hour?) with a potential expanding of cars from people fleeing public transit. Early estimates were that fully autonomous car fleets would reduce cars on the road by 80–90%, however there’s definitely some debate in this arena now. And it certainly depends on how the rollout goes. If autonomous cars are only available privately and they slowly penetrate the market, then they certainly won’t alleviate traffic. But if areas are fully served by autonomous fleets maintained by private companies or even by local governments, I would expect that to completely alleviate traffic. Then we could debate adoption rates but I think that as long as the fleet were deployed appropriately, adoption would be very very fast.
Last thing is I’d like to take a moment to imagine what public-transit ride-hailing could be like. You can imagine bus stops retrofitted with push-button ride hailing to better serve those who truly depend on public transit. Payment could be arranged with Clipper Cards (or similar) or even with a meter that could accept cash. The size of the vehicle could be tailored to match demand at that point or hour. Ultimately ride-sharing is more flexible than buses or rail systems, which is a tradeoff for reduced efficiency at peak hours. However, there is considerable efficiency to be reclaimed if we 100% autonomous vehicles, which could more than make up for the lost efficiency related to breaking travelers into smaller groups.
Anyway good piece and I fully agree that Lyft & Uber should not be subsidized unless they truly meet transit needs for those who need it the most.