similar behaviours which we intuitively know we could live a fair lot better without
The situations presented are indeed such ones that we could live better without, but the whole point of thought experiments is to construct the worst possible world, and find a way to decide that works even under those circumstances. By your logic, we could easily end up saying “it’s useless to argue about how one or two electrons behave, real world objects have much more of them… and anyway, tunneling effects are so weird and unintuitive that we surely have a wrong model”.
So while I agree that faulty models often result in faulty decisions, the question to answer is whether the model correctly applies to the current situation, and not whether the results are intuitive or not. Sometimes they really aren’t...
(edit: that said, you’re right in that before accepting such an unintuitive consequence, it’s worth thinking it thorough another time. We’re humans, after all, running on flawed hardware and having limited time...)
Well, the thought experiments are precisely the experiments in what’s intuitive; that’s why they are thought experiments rather than real experiments. It indeed is fairly useless to argue how one or two electrons behave, you have to take a look in the real world and see.
Likewise with the trolley problems; the implicit presumption is that they are not completely irrelevant to the real world, yet they are because they neglect the false positive rate (assuming it to be 0) while discussing an extremely low probability event (whose probability is well below any plausible false positive rate). This sort of thing is precisely why Bayesian reasoning is so important.
The situations presented are indeed such ones that we could live better without, but the whole point of thought experiments is to construct the worst possible world, and find a way to decide that works even under those circumstances. By your logic, we could easily end up saying “it’s useless to argue about how one or two electrons behave, real world objects have much more of them… and anyway, tunneling effects are so weird and unintuitive that we surely have a wrong model”.
So while I agree that faulty models often result in faulty decisions, the question to answer is whether the model correctly applies to the current situation, and not whether the results are intuitive or not. Sometimes they really aren’t...
(edit: that said, you’re right in that before accepting such an unintuitive consequence, it’s worth thinking it thorough another time. We’re humans, after all, running on flawed hardware and having limited time...)
Well, the thought experiments are precisely the experiments in what’s intuitive; that’s why they are thought experiments rather than real experiments. It indeed is fairly useless to argue how one or two electrons behave, you have to take a look in the real world and see.
Likewise with the trolley problems; the implicit presumption is that they are not completely irrelevant to the real world, yet they are because they neglect the false positive rate (assuming it to be 0) while discussing an extremely low probability event (whose probability is well below any plausible false positive rate). This sort of thing is precisely why Bayesian reasoning is so important.