Even with my kindle reader font set to its smallest, after dithering the framebuffer, it is still quite readable, and at arm-length you cannot see the difference. It would be a nice option for the PDF reader, and certainly usable for VNC even IF you dither the text. But of course, you really only need to dither the moving parts of the screen.
I have been testing this with a tcc-compiled version of the dither program. I will now cross-compile it to run faster, and then attach it to this post. You can test it from SSH (hint: press up-arrow and enter to repeat), while using the kindle framework to display various menus, images and book pages.
UPDATE: The program mentioned above only worked for K4 and newer because it was reading pixel bytes directly from the framebuffer. That does not work for 2bit/pixel kindles like the K3. I now have a "geekmaster formula 69" universal getpx function that returns a pixel value 0-255 from the framebuffer for any x,y coordinate position on any kindle. I am having problems on K3, where I cannot get both solid black and solid white. I am trying to figure out how to stretch and center the range in that formula. Figuring out these formulas needed for branch-free code is a non-trivial task.
I will post the framebuffer dithering program with it works on all kindle models.