If people with trained palates for wine cannot tell a chilled red apart from a white, or a room-temperature white apart from a red, it may not prove that the conventional wisdom around wine tasting is bullshit, but it sure suggests it strongly.

Also, wine tasting is absolutely a narrow classifier trained on limited data (in this case, tasters are not trained to classify red wine that is chilled). It's not surprising that wine tasters have overfit their classifier to the types and conditions of wine that they have been trained and tested on.

I believe this, but it's not what I said. I also assume a room temperature white smells and tastes completely different from a chilled one.

If there is nobody in the world who can, with significant accuracy, sort a set of reds and whites held at the same temperature into a set of reds and a set of whites, then it is reasonable to say that a lot of the conventional wisdom around wine tasting is probably wrong.

If wine tasters are too far up their own butts, use regular people instead, or people with more generally refined palates (e.g. chefs). Train somebody specifically for this task, even. At this point it's a hypothetical exercise.

Claiming e.g. "only trained wine tasters are qualified to make this distinction, and serving wines at 'incorrect' temperatures makes it impossible for trained wine tasters to tell the difference because they've 'overfit their classifiers'" is absolute hogwash.