Earlier quoted context omitted.
Of course it's a sarcastic comment, and I know it when I was comment. It seems that it was you, who are unfamiliar with its background and implication. "Reproducing the ultrasonics" here is a sarcastic reference to any sampling rate higher than 24 kHz (it actually backs back well before digital music). Audiophiles often prefer it because it's claimed that it has a wider frequency response, a lower quantization noise…
When recording audio, it is necessary to filter our frequencies above Nyquist frequency to prevent aliasing. It is easier to make a good filter when you have Nyquist freq. at 48 kHz rather than try to implement a filter that can silence everything above 24 kHz and doesn't touch anything below 20 kHz. Or am I wrong here?
His point is that once oversampling and filtering is done in hardware, the final output only needs to be encoded in a low sampling rate for its benefits, and all high-quality DAC and ADC already oversample the signal under the hood. Using 192 kHz for its output or music distribution is unnecessary, and only introduces drawbacks like additional IMD and wasted space.