Yeah, it is a general criticism that can apply to a lot of thought experiments. And don't get me wrong, I enjoy the problem, it's just that I also enjoy critiquing it.

I believe most people's initial response to the original problem is to pull, but far fewer people will push the fat man, and this is framed as highlighting a contradiction in people's beliefs. In reality, it shows that if you create an unrealistic scenario, it trips up people's intuitions. One reason people's intuitions tell them not to push the fat man is because their moral intuition is outweighed by their physical intuition. We all know that pushing a guy off a bridge won't actually stop a trolley (and even if it would, we can't know that), so if we start to consider doing something like that, our brains say, "No stop it idiot don't do that." But it's not telling us anything about morality, it's just telling us "that's not how physics works, dummy."

Our intuitions are grounded in a world where physical and scientific laws apply and where we can never have perfect knowledge of future events, and the further you break from that, the less useful they are. But if the idea with the trolley problem is to help us identify what (if any) consistent logical precepts we can apply that match our moral intuitions, then we need to have simple, straight-forward questions. The original trolley problem is a little contrived, but doesn't break from reality nearly as hard as variations like the fat man. That means that the intuitive response to the original problem is more trustworthy and reliable, compared to whatever we feel about the more contrived ones.

Imo pulling the lever is the correct answer, and whenever I see a variation that tries to contradict that, I look for ways that that variation breaks from reality in ways that would trip up my intuition. So in this case my intuition tells me not to convict, in "contradiction" of saying I'd pull the lever, but that's because my intuition hasn't internalized all the assumptions about magical psychic foreknowledge and stuff. When I consider the problem with all those assumptions, then I say, "Oh well in that case it's just like the trolley problem so pull," and in some cases that answer might make me come across as a psycho, but that's only because the hypothetical doesn't allow me to consider the full effects, risks, and ramifications that the action would have in the real world.

So that's my full solution to the problem. I studied physics in uni so sometimes I might be a bit too inclined to find a final objective answer to a problem that's supposed to be open-ended lol