I voted for no change because I feel the system has unrealized potential.
The XTrans array is the first generation of an innovative approach to digital imaging. The demosaicing algorithms (even Fuji's in-camera rendering) are also first generation. As on-board and off-camera computing power becomes more practical, more sophisticated algorithms will take full advantage of the Xtrans array.
Given how well the XTrans performs right now, I believe it deserves further development.
I do not agree that crops or large prints are negatively impacted by the XTrans system. I do know that I must use entirely different parameters as well as a different LR 4/5 work-flow for parameter optimization with XTrans raw images. From my point of view, the main disadvantage of the XTrans system is different types of subjects require more diverse raw-rendering parameter settings compared to Bayer raw images. It is difficult to sort out fundamental problems with XTrans raw rendering from sub-optimal poost-processing techniques. It didn't help that the first ACR tools for XTrans raw were completely unsuitable for some images.