You know how sometimes you look at a piece of research and think, «I suppose it's an interesting technical problem, but isn't teaching an AI to hunt and kill humans a pursuit fundamentally dangerous to the continued existence of mankind?» (techcrunch.com)