I'm very well aware of these discussions and "rules of 8" for impedance. Over my long years in Headphone space (soon to be 20 years), I personally learned that the important aspect is the impedance curve of the headphones/IEMs. For things like Shure 846, Campfire Andromeda or recently Crinacle Zero:Red, different output impedances result in significantly audible and measurable different responses. And for these I would always recommend to pick the source and even cable very carefully.
For the other aspects, we are more into the subjective section of Hi-Fi.
Like this, what exactly means "sound wrong" in this context and what makes it automatically applicable for all headphones by simple sharing the 48 Ohm specification (not even discussing that it's normally stated just at 1kHz). Even your linked article from Headphonesty puts the limit starting at 2.5:1 already.
We publish output specifications for our models, so the customers can make their purchasing decision based on that, ideally in accordance with recommendations by the headphone manufacturer. Our engineers simple believe that the sound we can achieve with the TPA6120 on some of our models is worth the drawback of increased output impedance.
Last year we released H5, running similar amplifier to M5 Ultra, based on TPA6120 and also running at 4.7/6.6 Ohm impedance. And the feedback on sound is great, even for people using low impedance IEMs. As many things in Hi-Fi, it's often more complicated than just one general rule.
One last thing, different hi-fi communities put different weight on this. For example, on our domestic Chinese market, the output impedance is not really discussed topic and our Chinese website doesn't even include it in specifications:
H2 page:
https://www.shanling.com/product/4691
M5 Ultra page:
https://www.shanling.com/product/4696