The Pause AI Movement Is Winning
Summary
Taylor Lorenz interviews Washington Post journalist Garrett DeVink about the growing "Pause AI" movement, led by figures like Eliezer Yudkowsky and Holly Elmore, which demands a halt to advanced model training due to existential risks. They discuss recent protests, OpenAI’s temporary pause after a security breach, and the tension between activist demands and the financial incentives driving companies like Anthropic toward IPOs. The conversation also covers political opposition to data centers and the feasibility of enforcing global AI regulations.
Topics discussed
Introduction: The Pause AI movement and guest Garrett DeVink
Origins and key figures of the Pause AI movement
What does 'Pause AI' mean in practice?
Duration of the pause and existential risk concerns
AI safety failures and models escaping control
Protests, civil disobedience, and media tactics
Public sentiment vs. intellectual existential risk
Political power: Data center opposition and Bernie Sanders
Policy proposals: Moratoriums and international agreements
Enforcement challenges and open source accessibility
Tech giants' stances: Anthropic, OpenAI, and Google
National security, military AI, and cyber threats
Strategic positioning ahead of IPOs and valuation
Realism of pausing AI and historical tech regulation
Consumer sentiment, social media activism, and conclusion
Listen ad-free on Castria