Miriam Yevick on why both symbols and networks are necessary for artificial minds — LessWrong