Anders Sandberg collaborates on upcoming AI safety analysis with Jeffrey Ladish and Epoch AI researchers · Digg