Social behavior curves, equilibria, and radicalism — LessWrong