Introspective update: I seem to have such unusual rules for how I act and respond to things that I feel hopeless about my intentions ever actually being understood, so I kinda just give up optimising at all and accept the inevitable misunderstandings. This is bad, because I can probably optimise it more on the margin even if I can’t be fully understood. I should probably tone it down and behave slightly closer to status quo.
Not expecting this to clear anything up, but the cheap-to-compute rules are downstream of three hard-to-compute principles:
A) Act as if you believed social norms were marginally more optimal than they already are.
Bad social norms are stuck in inadequate equilibria, and even if everyone sorta-knew the norms were suboptimal, it would take courage or coordination to break out of it. But if we are more people who are courageously willing to pretend that we already have the norms we want to have, then we’re pushing the frontier and preparing for a phase shift.
For example, the idea that you shouldn’t brag or do things that make it seem like you think you’re better than anyone is really pernicious, so I try to act as if no one agreed with this. This also means that I try to avoid constantly making self-protective disclaimers for when I act in ways that can be perceived as hubris, because the disclaimers usually just reinforce the norm I’m trying to defeat.
B) Act according to how you want other people to expect other people to expect you to act.
This goes to recursion level 2 or 3 depending on situation. Norms are held in place by common knowledge and recursive expectation. If I’m considering pushing for some norm, but I can’t even *imagine* a world in which people stably expect other people to expect me to behave that way, then that norm will never reach fixation in the community and I should look for something else.
C) Imagine an optimal and coherent model/aesthetic of your own character, and act according to how you expect that model would act.
I call it the “SVT-SPT loop”. We infer our intrinsic character based on how we behave (self-perception theory), and we behave according to what we believe our intrinsic character is (self-verification theory).
This is about optimising how I act in social situations in order to try to shape my own personality into something that I can stably inhabit and approve of. It’s kinda self-centered because it often trades off against what would be best for others that I do, which means I’m at least partially optimising for my own interests instead of theirs.
Introspective update: I seem to have such unusual rules for how I act and respond to things that I feel hopeless about my intentions ever actually being understood, so I kinda just give up optimising at all and accept the inevitable misunderstandings. This is bad, because I can probably optimise it more on the margin even if I can’t be fully understood. I should probably tone it down and behave slightly closer to status quo.
Not expecting this to clear anything up, but the cheap-to-compute rules are downstream of three hard-to-compute principles:
Bad social norms are stuck in inadequate equilibria, and even if everyone sorta-knew the norms were suboptimal, it would take courage or coordination to break out of it. But if we are more people who are courageously willing to pretend that we already have the norms we want to have, then we’re pushing the frontier and preparing for a phase shift.
For example, the idea that you shouldn’t brag or do things that make it seem like you think you’re better than anyone is really pernicious, so I try to act as if no one agreed with this. This also means that I try to avoid constantly making self-protective disclaimers for when I act in ways that can be perceived as hubris, because the disclaimers usually just reinforce the norm I’m trying to defeat.
This goes to recursion level 2 or 3 depending on situation. Norms are held in place by common knowledge and recursive expectation. If I’m considering pushing for some norm, but I can’t even *imagine* a world in which people stably expect other people to expect me to behave that way, then that norm will never reach fixation in the community and I should look for something else.
I call it the “SVT-SPT loop”. We infer our intrinsic character based on how we behave (self-perception theory), and we behave according to what we believe our intrinsic character is (self-verification theory).
This is about optimising how I act in social situations in order to try to shape my own personality into something that I can stably inhabit and approve of. It’s kinda self-centered because it often trades off against what would be best for others that I do, which means I’m at least partially optimising for my own interests instead of theirs.