I do not think that "cropped sensor" cameras (I think this is any camera with a sensor smaller than FF?) will dye out. To the contrary, several people here state they prefer DX over FX Nikon systems, because of smaller overall rig and WA lens choices.
Last week I was on an exciting UW-photo workshop at the Red Sea. The major goal was, of course, to learn photographic skills (hope it worked a little ), but a minor goal of myself was to compare the different camera systems by looking at the different images, how they are processed and talking to people. Before the workshop I was strongly inclined to acquire an additional FF body plus housing (Canon or eventually Sony with adapter, as I have already three Canon-EF mount lenses that I am using on my Oly EM1II MFT body by using Metabones converters). After the workshop this inclination is reduced, at the moment I see only little reasoning to go for FF and several against. Here are my (subjective) impressions:
#1.: When viewed on labtop screens or the big HD-TV screen that was on the ship, it was not possible for me to judge, whether an image was acquired with a 1" compact or an FX Nikon D850 (at magnification to view the entire image. Of course, these screens are not good enough to see subtile differences).
#2.: The only differences that I could see easily were with cave photos: Here the photographers with FF sensors could switch to high ISO and use short shutter speeds. The MFT photographers could easily compensate the ISO weakness of their cameras by longer exposure times (made possible by the superior image stabilization). At the end, however, the light beams captured with small sensors appear uniformely smeared, while (some) images captured with FF sensors (at high shutter speed) showed razor sharp light rays within the big beam, what is, of course, more beautiful.
#3.: After dinner there was always the "image review", where every participant could submit two images that were critically discussed, also postprocessing was improved. Here I had the impression, that the "reserves" for post-processing (e.g. increasing the shadows, stretching dynamic range, clarity etc.) increase clearly with the sensor size, beeing the smallest with 1" compact and biggest with FF (APS-C, DX and MFT in the middle). With "reserves" I mean the extend a slider in LR can be adjusted, before the image starts to look "artificially" overprocessed. The differences are there, but they are not overwhelming. Alex Mustard, who organized and hold the workshop and held all seminars, said that a FF raw image that is acquired under critical conditions and not exposed to the optimum (e.g. low light in caves) will have less IQ than an image of the same subject, acquired to its optimum, on a camera with smaller sensor.
#4.: Sharpness of the images: Regardless of the electronic poperties of a sensor, bigger sensors have the potential to yield sharper images, just because of the laws of optics (Huygen's principle, i.e. light behaves as a wave and it is intrinsically impossible to resolve an indefinite number of detail in a given area). This means that on a sensor that has 4x larger area (the relation of different sensor areas to MFT (1x) is as follows: Canon APS-C (1.46x), Nixon DX (1.65x) and FF (3.84x)) one can resolve, just by the law of optics, 2x the amount of detail (e.g. lines resolved per height of the image, what is often takes as a measure of sharpness), using the same lens at identical settings. This, of course, can be only seen when the digital resolution of the sensor is not rate limiting. In practice, however, image sharpness is mostly compromized by motion blurr (even when body/lens have image stabilization) and/or unprecise focus. Only with images without any motion blurr that are perfect in focus one might see better sharpness using FF (I doubt this would be possible to see on a normal screen without heavily magnifying the image).
In summary there IS, of course, better IQ with FF cameras, but the differences are not gigantic...
I also could see clear disadvantages of FF systems: Of course, the enormous size, especially when using the WACP, but also the big domeports that had to be used for rectilinear WA lenses. Second, the lens choices for very wide WA (the workshop was entirely on WA photography ): While many APS-C and DX cameras were using the Tokina 10-17mm fisheye as standard lens (I was using the Canon 8-15mm fisheye, adapted with my Oly EM1II MFT camera (=> at least as good, I think)), the FF photographers had to use 15mm fisheyes, sometimes with teleconverters, for comparable angles of views (there is no zoom fisheye available for FF that would cover a comparable range).
Still I feel a little like testing out a FF rig, but this probably will not take place in the near future ("thumbs up" for cropped sensors for me personally). But who knows for sure? ...