Can a culture of more open, bottom-up communication within and between organisations reduce the fear of being wrong, and produce better collective intelligence? Yes, of course, social software advocates might say. But could that also apply within intelligence agencies?