lukestein’s Twitter Archive—№ 4,314

@itaisher @BasilHalperin @namalhotra Agree with *all* of this. It’s possible that team DAG might express this as something like P(YangGen | YangPri) ≠ P(YangGen | do(YangPri)) ? That said, I think the former conditional probability is interesting (even if it resists interpretation as the latter).