Sure, I'll buy that: Aluminum and oxygen are great friends that love eachothers' company, and on a long-enough timeline here in Earth's atmosphere they'll always be reunited.
But how does that timeline compare to that of a human? Or even of the compact disc itself (a bit over 42 years old now)?
I mean: At least anecdotally, I have never discovered rot on any of my CDs that did not also have other contributing condition issues. I haven't even experienced the once-reported issues of air-dried, solvent-based (instead of UV-cured) inks. (And although my sample set is not infinite, it is also not particularly small.)
cds are both younger and built better, so they will not only last to a later absolute date, they will last longer relative to their manufacturing date, but cds have a few other things that mask rot even when it starts, which is both that they are digital and the player has buffering and interpolation, and also that the data format includes redundant data for error correction. (ld is analog and has neither, later better players do add some digital processing but it can't do the kind of good job with a 6mhz analog fm ntsc video signal that a cd player can with a simple audio bitstream)
A cd with the same rot that is visible on a ld (visible in the output not visible to the eye on the physical disc) will appear to play perfectly even in the cheapest junk player.
So, it will take longer, but I see no reason to treat "longer" as "indefinite".
There is no specific time, but it is inevitable and I don't think it's in the 100's of years but in the 10's of years, and the 10's of years, especially when many are already 30 years old, is not very many more 10's of years left.
And if that turns out to be pessimistic and they last another 50 or more? That's just a bonus. Lucky future rippers who get a chance to rip with even better tech later.
Tangentially I do also assume that some day long before the polycarbonate disintigrates, there will be a practical way to read even fully oxidized discs with a different frequency laser or even a camera or microscope-based head, or even a bulk scanner that just rasterizes the whole surface without even bothering to read the track in a spiral until after the fact purely in software.
As time goes on, tools get both better and more accessible, so in 1995 it would not be possible for a person to make their own laser head, but today it probably is, and in only a few more years will just get easier and easier, and probably at a rate that outruns the rate at which the discs fully degrade.
