No. A CD audio recording can contain literally all the information in the audio spectrum with enough dynamic range to capture human breathing and a jet flyover simultaneously.
You are not understanding his point. Linear PCM systems encode quiet parts of music with a lower signal to noise level than louder parts. So by over-engineering the bit depth, you keep the quantisation noise further from perceptibility than without it.
I think everyone gets sampling theory by now, your response was to an audio engineer.
But with typical production flow for a CD, you don't really have any classic quantization noise because you dither (probably with noise shaping) during downsampling to 16 bit; the effect is a perceptually-weighted 120dB dynamic range.
It doesn't matter that the poster is an audio engineer: there is simply no good science or objective evidence that people can hear the difference between well-recorded music played back with 24 vs. 16 bit media.
It does matter in the sense you are referring to basic sampling theory that anyone with 30 minutes to spare could learn, but quoting that to someone who's job literally depends on understanding it. You can assume he gets it and is making a point with that in mind.
https://en.wikipedia.org/wiki/Nyquist–Shannon_sampling_theor...