#223 – Neel Nanda on leading a Google DeepMind team at 26 – and advice if you want to work at an AI company (part 2)

Neel Nanda, who leads the mechanistic interpretability team at Google DeepMind, discusses his career, offering advice on how to succeed in AI safety research. He emphasizes the importance of luck, being in the right place at the right time, and seizing opportunities. Neel shares insights on crafting effective cold emails, leveraging LLMs for skill enhancement, and the nuances of AI safety work, advocating for efforts that advance both safety and capabilities. He also addresses the overconfidence in AI safety predictions and the need for intellectual humility. The conversation explores strategies for impacting AI governance within large companies, the value of technical expertise, and the significance of mentorship in research, highlighting the balance between autonomy and support in academic versus industry settings.