I like the PRT idea but it requires a HUGE investment in infrastructure.
While your car pooling idea requires less (additional) infrastructure, all rail-based transit has much higher infrastructure cost than PRT, and PRT should be cheaper than highway lanes of similar capacity. Of course PRT can't replace the flexibility of roads (which carry things like moving trucks, construction equipment, bikes, ambulances, etc), so it's not really right to compare it to basic road infrastructure (which is inevitably required); but it compares quite favorably to the cost of increasing road capacities. And as an immature technology there's more avenues open to improve it.
While responsive car pooling or taxi service doesn't require new infrastructure (instead increasing the capacity of current infrastructure) the energy intensity improvements seem limited. It would feature smaller vehicles, and by being more responsive and adaptive it could carry a higher number of average passengers. But it still requires vehicles larger than a typical car, and I imagine cars would frequently be running at less than full capacity. I wonder what the energy intensity of a normal taxi is? Probably atrocious -- they frequently drive around with no passengers at all, and tend to use fuel-inefficient vehicles. Ride share systems would certainly be a big improvement over that, and probably buses, but not a huge win. Though by using less vehicles more intensely you can afford to invest more heavily in fuel-saving vehicles or in vehicle upgrades. Anyway, an interesting idea.