tl;dr: Changing land use is the only way to make transit of any kind more effective than private cars. Technological improvements that made point-to-point tran

tl;dr: Changing land use is the only way to make transit of any kind more effective than private cars. Technological improvements that made point-to-point transit as effective as private cars are completely infeasible due to social and societal constraints. Self driving cars will be a feature for users but will NOT change or accomodate existing cities. Point to point would be great, but it's really, really expensive and doesn't change the economics of land use. Look at what cars cost. Most US cities have ~1 car per adult, and to average expensive cars and paid off cars, let's assume ~$5000 in payments per car per year, plus about that much for gas, maintenance, insurance, etc. So ~$10,000/adult per year. Now assume with automated p2p 1 car can do the work of 10 private cars at the same vehicle cost. But barring some efficiency changes, operational costs would be the same as 10 cars, so each car would cost ~$55,000/yr. This means $5,500/adult, about half the cost of a private car - big savings! For a city of 1M, assume 50K cars, and you need almost $3B/yr just for the cars, the whole organization (gov or private) would probably cost another $1B. $4B/yr would be a big savings over private vehicles and free up all the space used by parking. Big caveats: 1) For reference, Nashville (~650K city, ~1.5M metro) had a mass transit budget of $60M, or about 1.5% of this back of the envelope budget. The entire municipal budget was $1.5B, or about 37% of this budget. This demonstrates that however beneficial such a system would be, it' would require a complete restructuring of government and society to shift that kind of spending into centralized transportation infrastructure. 2) These fairly affordable unit economics only work if everyone participates. But the system also only works if you continue to maintain the existing road network, which makes the switch from private cars that much more difficult to justify. 3) A p2p network is an attempt to get people who live in non-transit areas to use transit. But many (many many many) of those people live there specifically because they want privacy and seclusion from other people, so any kind of transit system is wrong on principle.