The point of having LW posts around is not to take their titles as axioms and work from there. My hardware, corrupted as it is, has no intrinsic interest in traumatizing children, so I don’t suspect my brain of doing something wrong when it tells me “if it were reliably determined that traumatizing children led to awesome outcome X, then we should traumatize children, especially considering we are in some sense already doing this.”
In other words, I think an argument against traumatizing children to make superheroes, if it were determined that this would actually work, is either also an argument against mandatory education or else has to explain why it isn’t suffering from status quo bias (why are we currently traumatizing children exactly the right amount?).
Edit: I’m not sure I said quite what I meant to say above. Let me say something different: the post you linked to is about how, when humans say things like “doing superficially bad thing X has awesome consequence Y, therefore we should do X” you should be skeptical because humans run on corrupted hardware which incentivizes them to justify certain kinds of superficially bad things. But what you’re being skeptical of is the premise “doing superficially bad thing X has awesome consequence Y,” or at least the implicit premise that it doesn’t have counterbalancing bad consequences. In this discussion nyan_sandwich and I are both taking this premise for granted.
Even if net awesomeness increases though, do awesome ends justify non-awesome means?
The point of having LW posts around is not to take their titles as axioms and work from there. My hardware, corrupted as it is, has no intrinsic interest in traumatizing children, so I don’t suspect my brain of doing something wrong when it tells me “if it were reliably determined that traumatizing children led to awesome outcome X, then we should traumatize children, especially considering we are in some sense already doing this.”
In other words, I think an argument against traumatizing children to make superheroes, if it were determined that this would actually work, is either also an argument against mandatory education or else has to explain why it isn’t suffering from status quo bias (why are we currently traumatizing children exactly the right amount?).
Edit: I’m not sure I said quite what I meant to say above. Let me say something different: the post you linked to is about how, when humans say things like “doing superficially bad thing X has awesome consequence Y, therefore we should do X” you should be skeptical because humans run on corrupted hardware which incentivizes them to justify certain kinds of superficially bad things. But what you’re being skeptical of is the premise “doing superficially bad thing X has awesome consequence Y,” or at least the implicit premise that it doesn’t have counterbalancing bad consequences. In this discussion nyan_sandwich and I are both taking this premise for granted.