Google’s DeepMind lab has unveiled SIMA — a learning artificial intelligence model that will play games for you. It’s customized to play more like a human than a superhuman artificial intelligence. SIMA, which stands for Scalable Instructable Multiworld Agent, is currently just a research experiment.