I wouldn't at all let the top LCD backlight "problem" influence my decision for the simple reason that the camera, even with the "problem," performs flawlessly within and beyond its specified limits.
Anybody with a 5DIII can trivially confirm this. Put the camera in manual mode and set the exposure to EV 1 (one second at f/1.4 at ISO 100, if I remember right) and find something dark in a dark corner of a dark room where the meter bug is centered. Turn the backlight on and off, and the meter won't budge. EV 1 is the specified minimum light level that the meter is designed to work in, so when you see that the meter doesn't budge, you know your camera is working exactly as designed.
Lengthen the exposure by a stop, find an even darker corner of the room, and the backlight still isn't going to change the reading. You're now working in half the minimum amount of light the meter is designed for, and it still works fine.
Yet another stop darker, at EV -1, which is absurdly dark for photographic meters and not a situation where anybody with a clue would ever even bother with a meter, you might get the backlight to change the indicated exposure by a third of a stop -- but that's irrelevant, because the meter is already way beyond spec and into silly useless range. You need to get all the way down past EV -4 before the backlight changes the measured exposure by a stop, and that's waaaaay into ludicrous territory.
So, buy the camera or not if it's a tool worth the value to you or a toy worth the pleasure you anticipate it will bring. but, for the sake of all that's unholy, don't give even a moment's thought to this tempest in a teacup.
Cheers,
b&