That’s a fair question. I probably should have phrased that more carefully.
I’m not aware of any well-controlled listening tests that demonstrate audible differences between standards-compliant Ethernet cables in a properly functioning network. If there are such studies, I’d be genuinely interested in reading them.
My skepticism comes from the engineering. Ethernet is designed with differential signaling, galvanic isolation, clock recovery, error detection, buffering, and (where applicable) retransmission so that, when the link is operating within specification, the same data arrives regardless of the cable.
From a practical standpoint, Ethernet is also one of the easiest cables to evaluate yourself. Swapping Ethernet cables takes seconds, doesn’t require powering equipment down, and can be done repeatedly without anyone knowing which cable is connected. If the reported effects are as large as “removing digital glare” or dramatically changing soundstage, they should be straightforward to demonstrate under blinded conditions.