With Scoop, for example, all the people who Driver are just everyday folks. They're commuters like you, me, and the Riders they get matched up with. The economic incentives are like a bonus - where the real value is the improved time in the car and the time savings from the carpool lane (where applicable). And of course, other benefits around sustainability and parking that really impact and motivate some commuters.
Hell, I bet users will even sit in the front seat together.
It would subsidize $6.50 of my parking but I'd end up with an extra 30-40m of driving each day. Perhaps it's better than nothing?
You're right that if you need to go out of your way at pickup or dropoff, or if you weren't going to drive into the city anyway because parking is $$$, this isn't going to work for you. But then that wasn't the intended target audience. If the goal is to get cars off the road, and you weren't going to drive, but take light rail as the saner way to commute into the city --- then by all means, continue using rail or light rail for your commute!
So it is not designed to be something where the driver will make a living, but really just as a way for a driver to get some company, maybe take advantage of the HOV lanes, and it gets some cars off of the road to reduce congestion.
This has impacts to the sort of insurance drivers will need to get, since it's really no different from traditional car pooling. I would not be surprised if, over time, the system tries to pair people who have previously shared rides together for subsequent trips (unless either party rates the other as, "would rather not share a ride with this person"). Knowing that you are sharing the ride with the same person that you normally share your car with would make for a much better user experience, since people are, after all, social creatures.