I'm going to take a stab at this.
Uni made a broad blanket statement about perception equaling reality, but it was specifically about experience of a game.
Uni's statement could be applied to other areas of reality, but I think that's an error. I don't think they intended to claim that subjectivity trumps a scale for measuring an object's weight, and trying this reductio ad absurdum has failed to engage them or win the argument because that's not a point they're defending.
They were talking about the experience of playing D&D*, which is one which has a great number of confounding variables which inevitably result in different people at different tables having very different experiences with the game. At one table it may be "obvious" that a given class is more powerful than others, because everyone at the table sees the PC of that class constantly kicking butt and performing heroically. But that could still be a class which is mathematically, measurably, LESS good at their function than some other class. Which either is not represented at the table, or is played by a less competent player. Or one which is nasty enough that the DM overcompensates and hits them with more threats, resulting in that mathematically more powerful character underperforming at the table in the experience of the players observing. Or maybe the mathematically-weaker class is being boosted by a synergy with another class/player at the table. Or by some house rule the DM is using (maybe Flanking giving Advantage? That's a pretty common one). Or maybe the guy with the weaker class happened to roll great ability scores and so his particular character objectively IS more mathematically powerful than an average example of the class, but the other players aren't taking that into account in observing that "Damn, Bob's Fighter is a badass character!"
There are any number of these potential factors which impact the experience players actually have at the table. And what they see and feel at the table IS their reality. Even if it's contradictory to what math and white room analysis would predict.
One repeatedly observed phenomenon that comes up repeatedly over time is that Very Online players often do detailed statistical analyses of classes, DPR, and mechanical features which are rigorous, internally consistent, and mathematically verifiable, but far enough down in the weeds that they're often below the level of interest or visibility in play for casual players. Even for a lot of serious players who are less math-inclined. Even with players capable of following the math (which usually isn't very complicated), experience at the table always subjectively looks different than an ideal white room calculation. I know that as a DM, I've (e.g.) sometimes dealt with a perception from players that their luck is abysmal and they "almost always" roll badly that I did not perceive to be true. But if it's true on enough dramatic occasions, it will absolutely stick in the mind. It will shape their perceptions of and experience with the game.
(*with maybe a brief aside about how perception or personal experience comprise a given person's reality much more directly than abstract descriptions or mathematical models, which has a lot of validity. Remember Hamlet? There is nothing good or bad, but that thinking makes it so?)