4 ms·

Perhaps you're not over-valuing pain, but under-valuing numbers. There's no reason to think that morality can't or shouldn't center around minimizing a pain met

Perhaps you're not over-valuing pain, but under-valuing numbers. There's no reason to think that morality can't or shouldn't center around minimizing a pain metric. That would put it in line with any other trajectory-planning-under-constraints problem. Once you've done a bit of work in trajectory optimization, you start to feel the numbers come alive, and develop a visceral feeling for them. I wince when I see a car scrape to the edge of its traction limit as it rounds a corner at high speed. Nature itself plans trajectories to minimize action, the integral of a Lagrangian: it's literally just a number, but it decides the motion of molecules and galaxies. What's wrong with a moral system based around minimizing a pain integral, subject to constraints? Still just a number. But one that can easily take an important position in our decision-making, and find a deep-seated place in our minds as well. Of course, we probably ended up with morals in the first place as a distillation of game-theoretic heuristics, which helped us trust and cooperate with each other, which would have been an evolutionary benefit. But as with anything else, we're accountable to the drives that evolution left us with, not evolution's reasons for putting them there. The morals are in our minds, not in nature, but either way it's numbers all the way down.

N0b8ez 2y ago
Consider whether we should define electrons to be "positive" or "negative": there isn't any fact of the matter, it's merely a convention. But it doesn't seem like experiencing pain works this way: pain feels a particular way. If pain is just a mathematical entity, there shouldn't be any particular way that it feels, since it's just a symbol. Maybe we're deeply wrong about our intuitions, and actually no conscious experience is happening in the brain at all, despite the brain having an incorrect model about itself which predicts this (and causes it to make utterances claiming that it has experiences). That's the school of thought known as eliminitivism. Under this view, all suffering is a myth, in the same way that luckiness and angels are.
- mitthrowaway2 2y ago
  Sign conventions are not arbitrary in trajectory planning, though. You'll get very different results if you solve for a robot trajectory that moves from A to B while minimizing motor energy consumption, than if you solve for maximum energy consumption. You'll notice the difference - it's a mistake you'll only make once! If we have some internal decision-making process that pushes us to make decisions that minimize a pain metric, the sign we attach to the "pain" number will make a very big difference, both in internal state and resultant observed behaviour.

N0b8ez 2y ago

Consider whether we should define electrons to be "positive" or "negative": there isn't any fact of the matter, it's merely a convention. But it doesn't seem like experiencing pain works this way: pain feels a particular way. If pain is just a mathematical entity, there shouldn't be any particular way that it feels, since it's just a symbol. Maybe we're deeply wrong about our intuitions, and actually no conscious experience is happening in the brain at all, despite the brain having an incorrect model about itself which predicts this (and causes it to make utterances claiming that it has experiences). That's the school of thought known as eliminitivism. Under this view, all suffering is a myth, in the same way that luckiness and angels are.

mitthrowaway2 2y ago
Sign conventions are not arbitrary in trajectory planning, though. You'll get very different results if you solve for a robot trajectory that moves from A to B while minimizing motor energy consumption, than if you solve for maximum energy consumption. You'll notice the difference - it's a mistake you'll only make once! If we have some internal decision-making process that pushes us to make decisions that minimize a pain metric, the sign we attach to the "pain" number will make a very big difference, both in internal state and resultant observed behaviour.