Driving straight is not the challenge. The failure mode is much more severe on a highway, and extreme conditions are probably just as difficult to manage for an autopilot as city traffic, if not more. For example, what does the AP do when - the car hits a big pothole and perhaps blows a tire? There's too little time for a human to take over. Sometimes you need to make a choice, whether to stop abruptly or not. - lane markings disappear because of prior construction? Construction detour? - truck blows a tire in front of you? (happened to me twice) - some animal/idiot wanders into the road? Hint: avoid is a better maneuver than hitting and holding the brakes. You don't want to stop dead on a highway.
Give me city traffic with a 50km/h speed limit any day.
Blow a tire? The car has the information from sensors to tell which tire blew instantly, allowing it to apply brakes on the other wheels, hold the steering at the appropriate angles, perhaps even adjusting the suspension. Then it will monitor the other lanes in a 360 degree view to safely move off the road as quickly as possible. In addition, one could use run-flat tires to minimize this chance. It would also monitor tire pressure to minimize the chances. In addition to that, one could have the car monitor tread depth, and require manual override for unsafe tires.
Lane markings disappear? GPS data, in addition to 360 degree visual sensors and data from vehicles in front of it mean that your self-driver is more likely to determine a safe path more readily than a human driver.
Truck blows a tire? The car doesn't have your quarter-second reaction time (at best) to contend with before it begins to analyze the data from its 360 degree cameras which know where it's safe to go, sensor data which tells it how much grip it has, road conditions, brake ability, and is able to determine and begin executing the best case for avoidance before you would have even noticed the blowout.
Idiot or animal wanders in? Car is more likely to notice it faster. Infrared spectrum makes body heat stand out on animals that evolved to blend into the surroundings and night-time walkers that think that black is a good idea. In addition, it's able to track these objects without ceasing to track the location of the vehicle, and surrounding vehicles and objects. The car is able to process visual input from 360 degrees of visual input, in addition to all sensor data to make the best decision on minimizing impact with the animal or idiot. And it does this beginning before you would have noticed them (by about a quarter second) and with a good bit more input than you would have available.
In addition to all of this, it's possible that cars could share this data, so that there are multiple vantages being processed on any of these situations, so that they can react in tandem, and know what other vehicles are intending to do, lowering chances of impact even more.
The car is a better driver.
After 36 months, you have the right, but not the obligation to sell your Model S to Tesla for the same residual value percentage as the iconic Mercedes S Class, one of the finest premium sedans in the world, made by Daimler (also a Tesla partner and investor).
Not only is Tesla guaranteeing that resale value, but Tesla CEO Elon Musk is personally standing behind that guarantee to give customers absolute peace of mind about the value of the asset they are purchasing.
This part makes me want to buy one. Well, among a lot of other reasons. But largely because I want to reward someone for standing behind their product. You don't see a lot of that, these days. Lots of difficult-to-use warranty's, but little real pride in a product.
It shouldn't be innovative, or revolutionary; let's face it: this is something lacking from most companies these days.
The point I was trying to make is that they serve different purposes, and are targeted at different groups. Or maybe overlapping groups, but still for different tasks.
Have you considered adblock and flashblock?
You think the toilets don't need occasional maintenance? Maybe a pressure gauge near the flush? Check the water tanks for leaks? They could have sensors too.
Over head bins provide much the same reasoning as the seats above. Weight average, weight over all, weight distribution, maybe even the ability to check for the weight shifting during turbulence so that stewardesses know which bins are likely to kill someone when they are opened, and can get their smart phones ready for youtube goodness.
Remember, in this world of data analytics, it's all valuable to someone.