#870: Sebastian Mallaby, Biographer of Demis Hassabis — Lessons from 100+ AI Insiders on The Race to Superintelligence, The Religion of AI, and Spotting Breakthroughs Early

Overview

This episode features Sebastian Mallaby, author of The Infinity Machine, in a deep dive into the race for artificial general intelligence (AGI), the personalities driving it, and the societal implications of AI. The conversation spans the history of DeepMind, the challenges of AI safety, the geopolitics of AI development, and the philosophical and practical questions surrounding superintelligence.

Notable Quotes

- Religion is the lexicon for dealing with something that we find too mysterious to really understand. – Sebastian Mallaby, on why AI often evokes religious metaphors.

- You could do the best book in the world, an A+ book on a C- topic, and it would get you nowhere. – Sebastian Mallaby, on the importance of choosing impactful subjects.

- Prepare your mind. – Sebastian Mallaby, emphasizing the need for intellectual readiness in the age of AI.

🧠 The Genesis of DeepMind and Demis Hassabis

- Sebastian Mallaby recounts how Demis Hassabis, a polymath and former chess prodigy, founded DeepMind with a mission to build artificial general intelligence (AGI).

- Hassabis’s obsession with AGI began at 17, inspired by books like Gödel, Escher, Bach and Ender’s Game.

- DeepMind’s early funding came from contrarian investors like Luke Nosek of Founders Fund, who championed Hassabis despite widespread skepticism during AI’s winter.

- Google’s acquisition of DeepMind for $650 million was pivotal, providing the resources to scale research, including breakthroughs like AlphaGo and AlphaFold.

⚖️ The Duality of AI: Promise and Peril

- Mallaby argues that AI’s potential for superabundance (e.g., solving global challenges) is tempered by the disruptive path to get there, likening it to the China shock in trade.

- On the doom side, he shares Jeff Hinton’s thought experiment about AIs developing survival instincts, raising concerns about unintended consequences.

- Mallaby emphasizes that the probability of catastrophic misuse of AI is not zero, challenging overly optimistic or dismissive views.

🌍 AI Geopolitics and Collaboration with China

- Mallaby discusses his trip to China, where he found surprising openness to AI safety discussions among Chinese technologists.

- He advocates for a global AI non-proliferation framework, drawing parallels to Cold War-era nuclear treaties.

- While the U.S. has imposed chip export controls to slow China’s AI progress, Mallaby notes that these measures have only created a modest lead, suggesting collaboration may be more effective in mitigating risks.

🛠️ The Role of Venture Capital in AI Innovation

- Mallaby highlights the venture capital strategies that funded AI breakthroughs, such as Peter Thiel’s contrarian approach at Founders Fund.

- He praises Bill Gurley for his disciplined investment in Uber, calling it a model for identifying transformative opportunities.

- The conversation explores how venture capitalists balance risk and reward, with Mallaby emphasizing the importance of backing missionary founders like Hassabis, who are deeply committed to their vision.

📚 Preparing for the Age of AI

- Mallaby stresses the importance of intellectual rigor in the AI era, warning against over-reliance on large language models (LLMs) for thinking.

- He uses the phrase prepare your mind to encourage individuals to focus on meaningful, creative, and critical thinking tasks that define human identity.

- Tools like LLMs can enhance productivity, but Mallaby cautions against outsourcing core intellectual activities, such as writing, which he views as essential to personal growth and understanding.