OK, so we all had our share of laughs,.. but seriously I think it is a good idea to capture as much of what the lense sees as possible, and leave the final cropping/format decision til later (i.e. in Photoshop rather than in the field) . You could get a bigger square crop or wider (but lower) crop with a circular image chip than our current 24x36 will get us.
There would no longer be a need for vertical grips on the camera, and even the tilting horizons could be saved without cropping the picture.
I appreciate the added cost argument, but this is probably just a matter of time, and original article was on future development. The mirror problem might be a bit more tricky though.
So maybe it is time to start thinking outside the box