I think a lot of these notions get started the same way as a lot of, shall we say, "debatable" audiophile "remedies." Theoretically possible + infinite equipment configurations = empirically impossible to completely disprove. Therefore, even if it also equals empirically impossible to prove, there will always be hoards of people willing to put forward their personal anecdotes as evidence. These are usually people who just spent hundreds of dollars on exotic cables and whatnot and are thus biased in the direction of justifying their purchase. This is called confirmation bias, BTW. In our society, it seems, scientifically measurable = easily perceived, even if double blind testing doesn't back that up.
For the record, data is data. You could make a CD out of bloody Kryptonite and it wouldn't sound any different. On paper, yes, the gold will have different reflective qualities than a standard CD. But I'd be willing to bet my headphone collection that, in a double blind test, nobody on the planet could perceive a clearly audible difference and reliably identify the gold disc, all other things being equal. 1s and 0s aren't affected by the material they're encoded on, nor the material through which they pass. Either the data makes it through, or it doesn't. The DAC is where the most variability will occur, apart from your amp and headphones/speakers, of course.
CDR's do degrade, but it takes years. Some tips are to keep them out of direct sunlight and in climate-controlled conditions. Believe it or not they can warp and it changes the sound ever so slightly. I know this because I happened to be unfortunate enough to damage an original disc I had such that part of a long track skipped. I had made a copy of that track on another disc, and it, too, got messed up. Luckily, it skipped in a different place. I figured I'd just splice the two good parts together with a DAW. Though I'm sure both versions of the track started out identical, the copy (stored and handled poorly) was actually slightly longer than the original (by a fraction of a second), and thus when played in unison the tracks were phase-shifted.
