I keep wondering what happens when all of today's cars are 10 or 15 or never mind 20 years old. If so many of these cars have infotainment issues out of the gate, do we really think automakers will have fixes years down the road? Is there going to become a huge aftermarket to just replace all these units? That won't be easy with fully integrated units. Take the Volvo XC-90. Large screen controls just about everything and has had problems since launch. If that doesn't work, the whole car is useless. Imagine having today's smartphone fixed 15 years from now? How about connectivity? Do we really think all future phones will connect to the Bluetooth in the cars we're buying now? Hell, people have trouble connecting current phones with current cars. The industry is really setting up a bunch of problems.
They are already hacking the Mazda CMU. AIO-Tweaks is an example of that. One step further would be something similar to what you suggest - an all out after market CMU replacement.
Bluetooth, will probably be a round for quite sometime, however. Its now entering ubiquity and cutting across all sectors which means multiple product types will be using it, if not already. There is still lots of room for Bluetooth to grow. An example would be the fact that right now, as I type, Pilatus is going through official certification of their newest aircraft concept the PC-24. It is a Very Light Jet which will seat 6-8 comfortably and be single-pilot certified. That aircraft will be Bluetooth enabled, not just in the Cabin but also in the Cockpit. In fact, right now, there are Bluetooth Pilot Headsets (just bought my first pair) beginning to pick-up favor with pilots both private and commercial. So, I think Bluetooth is going to be with us for quite a while given the simplicity of the technology behind it.
Regarding the "Fully Integrated" Head Units. This too was for years a hot topic of conversation in general aviation. That conversation has pretty much been put to sleep as the move towards full cockpit integration of avionics that places everything into a singular system (engine monitoring, aircraft systems & performance information, navigation & flight planning, flight instrumentation information, etc.) all under essentially one roof or what are called PDF (Primary Flight Displays) and MFD (Multi-Function Displays) has been underway for a while now. And, as you say, we pilots wanted to know what happens when the lights go out! Or, more to the point, what happens when the PFD goes out?
We now know the answer. The system is Fault Tolerant, Redundant and Failover Protected - so what was the the MFD now becomes the default screen during an inflight outage of PFD. All required information for SOF (Safety Of Flight) is cast over to the MFD and the aircraft is flown in routine fashion. After landing, the mechanic repairs or replacing the faulted unit and puts the aircraft back in service after signing off on their work.
Getting that level of fault tolerance, redundancy and failover into a mass produced vehicle is not impossible - but I can imagine it would push up the costs somewhat if not carefully designed. And, it would have to be a Single Screen solution as you would not want to drive around with multiple Screens in your vehicle. How you get there would take some creative engineering design work and I suspect it would also require a lot of back-end architecture that isolated individual functions that come through the screen such that when/if they fail, they don't take the entire CMU system with it down the tubes (so to speak).
Like this morning, after doing a TPMS and Maintenance Schedule (Oil Interval) initialization, the CMU would not boot after starting the engine. I had to shut down the engine then re-start a second time to get the unit to boot itself. If the system was fault tolerant, the failure in the TPMS and Maintenance Schedule initialization would not have been enough to hang the entire system causing an engine shutdown/restart sequence. That kind of design can be done on the coding side of the system without requiring multiple screens to cast remaining systems information that survives the fault.