... that maybe another nail to the coffin of larger chips /MF.
I doubt it: if and when DMF gets modern active pixel CMOS sensors, it could just as well get this sort of tchnology too. However, one weakness of this idea is reduced highlight headroom: some photosites get the full incoming white light signal _plus_ some diverted red or blue from neighboring locations, while others get only a weaker signal with red or blue light removed, so these "plus" photosites will blow out earlier. All in all, the new approach seems mainly oriented to getting the best low light performance out of the small sensors used on the camera-phones and super-zoom compacts that seeme destined to dominate fixed lens photography, and to sell in vastly greater numbers than sensors for interchangeable lens cameras.
Anyway, I rather think that low light performance is relatively low on the list of reasons that some people will continue to choose DMF over formats 35mm or smaller, behind DR, big, bright VF images, and the inherent optical advantages of larger format lenses that come from the lower degree of enlargement needed to get a given size of final displayed image.