Re: Tesla has plenty of problems...
Apple has even less chance of developing a top notch autonomous vehicle than Tesla
That was the point of the article - Apple are not going to develop a car, nor are they doing to make the AI to drive a car. They went there, saw the problems, and realised it was never going to offer a payback.
What Apple want is to control the human interface of future vehicles - entertainment, information, navigation, and the telemetry side of things. They know they need to control the hardware modules that the system runs on, and the car makers then need to be able to link that to what they've built. In theory its a good bet, because hardware makers (including car makers) are universally shit at software.
The difficulty is that development timescale and support lives in automotive are an order of magnitude different to what Apple is used to in its existing products and services. There's the whole certification and safety needs that are "red line" requirements. And there's the fact that Apple don't like having to work with other parties that have significant control over implementation. Agreeing to work with (say) BMW on car systems is all very well - execs can agree this over a well padded lunch - the reality will be a stark difference of corporate cultures that makes progress slow and difficult. In this respect, Field's experience or product design at Tesla is of mixed value - he knows the systems issues, but getting things done at Tesla is worlds apart from working with the rest of the car making industry. And Apple need to work with multiple players - just getting their system into one brand might give the volume they want (say VW Group) but that would be a single buyer - a high risk strategy, and doesn't play to the premium segmentation Apple have always wanted.
As I write that, it seems that in-car systems isn't a good fit at all for Apple - it might defend the walled garden a bit, but the whole concept is B2B sales (that Apple are not good at), varied hardware environments (that Apple don't like), risk averse, cost conscious partners who move at glacial speeds (that Apple will struggle to work with), and to get the volume to make it economic, there can be little segmentation of the end users (which challenges Apple's high cost, high markup, high value customer model). My expectation is they'll get the software platform well advanced, maybe even have demonstration versions custom built into driveable vehicles that can be show to the press. Then they'll find all those challenges remain, and they'll conclude that the benefits to Apple are outweighed by the costs and risks, and they'll either shut it all down, or reposition as a much more modest entertainment and satnav platform with modest integration for vehicle settings, so that what is then nothing more than software and a better touchscreen can be sold as a $2,000 option. That would fit the segmentation model Apple have, fit with their expertise and margin requirements, and minimise the problems of a more encompassing control system.