New social credit formalizations — LessWrong