On the empirical question, I would just say this: When a subject has been discussed for decades and the vast majority of relevant experts near-unanimously report a particular conclusion, and I see that same conclusion reported in all my personal experiences with non-experts as well, then in the absence of any reason to believe otherwise - which so far hasn’t been provided - then it seems appropriate to believe that the conclusion is widespread.

But, with that said, I think you might be interpreting my claim in a way that’s far stronger than I intend. I’m not claiming that opposition to the repugnant conclusion is somehow deeply hardwired in any essential way into every single human being, regardless of culture or education or anything else. After all, it’s very likely that, in order to understand or engage with the question in the first place, you probably need to have at least some shared foundation in basic concepts related to the modern study of population ethics, which many people on planet Earth don’t have. So I certainly don’t mean to suggest that, like, some random villager in the Amazon has some pretheoretic commitment one way or another. Rather, I’m just claiming that, for those who attempt the question - that is, for those who are engaging with it as intended in the context of analytic philosophy more broadly - a conclusion in favor of one world is much more common than a conclusion in favor of the other. You can think the “rules of the game” involved in that engagement are fundamentally flawed or misleading, but that’s beside the actual point.

Otherwise, I don’t think it matters at all whether we’ve agreed on what constitutes a life barely worth living, or if we’re imaging the same thing, or anything like that. And frankly, I’m not sure why you think it would matter. The question is about these particular states themselves, and not what actualizes or instantiates them. So even if we disagree on what a life worth living looks like, we should both be able to render our own judgments about the implications of such a life existing or not existing. As an analogy, let’s say I ask a hundred people, “Would you prefer a president who has good economic policies that increase the GDP by 5%, but bad foreign policy that starts an unjustified war that kills a million people, or a president who doesn’t start any unjust wars but lowers GDP by 5% through bad economic policies?” Two people who answer “Good foreign policy, bad economic policy” are rendering a judgment that reports something meaningful about their priorities, even if they radically disagree on what good economic policy would be or what constitutes an unjust war (Although maybe you would also dismiss that question for the same reason?) Similarly, if two people say they prefer a hundred lives that are extremely good over a billion that are barely worth living, that’s an important shared judgment regardless of what they take those claims to be describing in any detailed sense - again, because the question is regarding those states themselves and not what instantiates them.

The comparison with asking whether you’d like to live with elves or dwarves doesn’t make any sense to me, because in that case the issue would be that the overall goodness or badness of those worlds is obviously underdetermined by the description. You’d ask for more details - what is the society like, what level of technology do they have, whatever - because you’d need those answers to determine the overall quality of the life you’d be living. But in the case of the repugnant conclusion, it’s the exact opposite: What you’re given is only the overall quality of the life, and if you already have that, no details should matter. That’s why I still can’t wrap my head around what more information you could possibly want! All the information anyone could ever give would, by stipulation, necessarily align with original description of the overall quality. It’s sorta like if I asked you to think about a shiny red car you wanted to buy, and you asked what kind of paint the car had and what model it was - the answer would just be “Whatever paint would make it shiny and red, and whatever kind of car would make you want to buy it.” It’s just not clear to me why anyone would need access to the first-order facts of the situation, whose only role would be to help you determine the qualitative appraisal that’s already been provided.

Similarly, the situation you’re describing with people entering into the simulation doesn’t make sense to me either. How could observing the situation change their judgments, if the nature of the stipulated situation itself is pegged to ensuring that a particular judgment is given? It just doesn’t seem to me like anyone could be wrong about judgments like this. It would be like if I asked someone, “What would you do if someone stole your shoes, but they were a pair you didn’t really like?” and them responding “Hard to say, I might end up really liking them.” We just said you didn’t! It would be one thing if I was describing a specific society and assuming that life inside that society would be barely worth living; then, I might legitimately decide otherwise if I experienced that society first-hand. But if I just assert, axiomatically, that I’m discussing a society in which a life would be barely worth living, then the idea of my opinion changing once I experienced that world is literally incoherent to me; I can’t understand how it would be possible, even in theory.

I suspect this is just another example of a fundamental gulf between us in terms of basic methodological considerations, but at some point, it’s hard to not notice that your approach generally forecloses on the possibility of meaningfully exploring a large number of questions that I am extremely confident other people have made real progress on through the use of other competing frameworks. And that to me is a major point in favor of those other frameworks, especially since the alternative explanation - that the entire field of population ethics is literally meaningless and/or confused - is in direct contradiction with my own first-person experience of reasoning through these questions in productive ways myself. I guess it’s just hard for me to understand what motivates this skepticism of abstraction, given that I can’t see where it’s actually going wrong, as opposed to just generating conclusions that don’t meet an (imo arbitrary) level of specificity. Anyway, I won’t drag this on any longer, but I would at least encourage you to consider what interesting, self-consistent conclusions are possible if you did accept the possibility of abstracting out this way, and whether they might be worth “purchasing” through your own methodology.