It is kind of shocking how much the LessWrong scene constantly and articulately describes their own failure modes without ever fixing them.
“Stuck priors” and they can’t update away from the model of AI that made doom predictions make sense.
“Second-order consequences” and their moral analysis stops at “who was directly and non-consensually harmed”
Yud’s beautiful essay on how pursuit of truth is greater than, prior to, and beyond any one methodology while they’ve become method-worshippers.
Truly, do as the rationalists say and not as they do.
yeah, smart rational people would absolutely host 100 person gang-bangs because the second order effects, incentives, ripples and norms around such behavior are just obviously pro-civilizational