AI Leaders Can't Get Along. So How Will They Keep Us Safe? — With Kevin Roose

The race to develop artificial superintelligence is defined less by sober scientific progress and more by intense, personal rivalries among a small group of industry leaders. Despite their public warnings about existential risks, key figures from OpenAI, Anthropic, and Google DeepMind are trapped in a competitive cycle where mutual mistrust and the drive to lead fuel rapid, often reckless development. While the potential for transformative breakthroughs in medicine and science remains high, the lack of a unified safety framework and the tendency to prioritize speed over control create significant dangers. The current landscape, characterized by high-stakes corporate politics and a belief-driven culture, suggests that the industry is ill-equipped to govern a technology that could fundamentally surpass human intelligence, necessitating a shift toward greater caution and collective oversight to mitigate potential black swan events.