Formalize the Hashiness Model of AGI Uncontainability — LessWrong