for example, it's common to ask whether someone would divert a train onto another line, knowing it would kill someone, to save some larger number of people. that's interesting because two heuristics we use in common cases are (i) avoid doing something that would harm another and (ii) minimise the amount of suffering.
in that case it's not that anyone thinks it's particularly useful to discuss trains because trains are such fascinating things in themselves. instead, it's an interesting case precisely because it shows how starting from two common rules can give contradicting answers.
The issue is that these problems are never "real world" and so are irrelevant to any actual discussion. Contradictions are, clearly, problematic - but only if we are going to face them. Until then they are just interesting diversions (pretty much like all of capital-E Ethics)
I'm not sure there is a real world problem that is completely morally insoluble.
I don't think moral discussion should focus on unpleasant edge cases. More interesting and useful is how to handle more common situations.