Summary
In this episode of The Diary of a CEO, host Steven Bartlett sits down with Daniel Kokotajlo, a former OpenAI researcher who walked away from approximately $2 million by refusing to sign a non-disparagement agreement on his way out. Kokotajlo is known in AI circles for detailed probabilistic forecasting of transformative AI timelines, and this conversation covers his current median estimate: a 50% chance superintelligence arrives by 2029, with meaningful probability of earlier arrival.
Kokotajlo argues that the most important dynamic being underappreciated by the public is that leading labs — OpenAI, Anthropic, and others — are primarily focused on automating their own AI research processes before deploying intelligence broadly across the economy. This means an intelligence explosion could occur internally at labs, making AI vastly superhuman at AI research, before most sectors see obvious disruption. He points to Anthropic’s reported revenue growth from roughly $1 billion to $60 billion annually in one year as evidence of the pace. He also discusses the competitive pressure between lab CEOs he characterizes as a race dynamic with existential stakes, citing a roughly 70% probability of catastrophic outcomes including human extinction.
The interview also addresses job displacement — why Kokotajlo believes it will be sudden rather than gradual due to recursive self-improvement dynamics — and what individuals and policymakers can realistically do to steer toward better outcomes. His scenario planning work, including a detailed public report, is referenced throughout.
📺 Source: The Diary Of A CEO · Published July 13, 2026
🏷️ Format: Interview
