Reading this thread makes me depressed about the potential for AI alignment thinking to reach mainstream in time :(