Neel Nanda explains how overcoming perfectionism and embracing luck got him into a leadership role at Google DeepMind.

OpenAI and Apollo researchers aren't concerned that current models could carry out the most serious kinds of scheming. As with much AI safety research, testing results primarily indicate risks of ...