Thinking about this a bit, the idea of taking two readings from the sensor at different sensitivities and then merging them into an image that has both shadow and highlight detail is a bit primitive. Why not take multiple readings? Why not keep sensor data across the board from the lowest ISO to the highest and use a sophisticated raw converter to meld them? Or in-camera, look for detail in the lightest areas at all ISOs?
It something that film just cannot do, and it's exciting to see the film mindset start to be discarded.