The state of AI safety in four fake graphslesswrong.com3 points·allenleee··0 commentsOpen articleSaveView on HN