Vladimir_Nesov comments on What would defuse unfriendly AI?