Some thoughts:

Dave Moulton(he of Golden Ears listening course) may be able to hear differences that we lesser mortals cannot so i cant agree with the concept of discounting the inability of most to be able to discern an audible difference between audio samples, no matter what the context, making these tests completely meaningless. Though i would agree with Miro that who is listening/comparing (eg engineer or lay person) is an important distinction.

Using any test as a good/bad type of judgement device is, IMHO, not the best way to go. IOW making value judgement's from these kinds of test can result in mistakes. That does not make the testing itself useless. Instead, it means that, like most things, it can be abused.

Changes in timbre, tonality, small changes in the envelope of sound itself, whether they can be obviously discerned by lay person or engineer, may be objective or subjective, but ultimately can be used in the decision making process of sound recording, design and mixing. IOW the tests can generate useful information if we avoid the good/bad paradigm and instead focus on the idea of learning what a particular piece "sounds like" to us, since we are the ones that have to use it.

As Miro points out, we are all going to hear stuff differently and have differing tastes, so we naturally audition equipment. I see these tests as a starting point for getting a preview of the general timbre/tonality/ etc and hearing it along side a known standard, even if imperfectly matched. Just a head start on the auditioning process , as it were.