"No magical healing, we try to avoid children touching hot stoves once.  I think if dath ilan could get faster learning by - no, that's not valid reasoning on my part, they could attach enough disutility to the students' experience of pain that they wouldn't do it.  I guess I'm still skeptical that you're describing a system that's actually locally optimal and that people aren't messing up?  Because if you get an electrical shock for a wrong math answer - that's a kind of pain we could inflict without lasting injury, if we wanted to go that route - then you don't just learn not to answer math problems wrongly, I'd expect you to also learn not to answer math problems and not to go to classes, in some deep part of you that you can't consciously override.  And it sounded like - from something Carissa mentioned earlier - people end up afraid to point out what looks like an error by the teacher, because if they're wrong about that, they might get pain inflicted on them.  That sounds like - exactly the kind of incredibly obvious failure mode I'd expect to develop, if somebody had the bright idea of trying to use pain to teach things, but you were also so bad at institutional design that students could get better grades by sleeping with the teacher?  I would have been a lot less cooperative with the older kids teaching me if I'd been getting punched by them for errors instead of paid by my parents for successes."

"No magical healing, we try to avoid children touching hot stoves once.  I think if dath ilan could get faster learning by - no, that's not valid reasoning on my part, they could attach enough disutility to the students' experience of pain that they still wouldn't do it, so it's not much evidence that they don't do it already.  I guess I'm still skeptical that you're describing a system that's actually locally optimal and that people aren't messing up?  Because if you get an electrical shock for a wrong math answer - that's a kind of pain we could inflict without lasting injury, if we wanted to go that route - then you don't just learn not to answer math problems wrongly, I'd expect you to also learn not to answer math problems and not to go to classes, in some deep part of you that you can't consciously override.  And it sounded like - from something Carissa mentioned earlier - people end up afraid to point out what looks like an error by the teacher, because if they're wrong about that, they might get pain inflicted on them.  That sounds like - exactly the kind of incredibly obvious failure mode I'd expect to develop, if somebody had the bright idea of trying to use pain to teach things, but you were also so bad at institutional design that students could get better grades by sleeping with the teacher?  I would have been a lot less cooperative with the older kids teaching me if I'd been getting punched by them for errors instead of paid by my parents for successes."