AI companies are racing to build more powerful models, even as their leaders warn they may struggle to control what comes next. Former New York Times tech columnist Kevin Roose joins Zaid Admani to explain why he takes those warnings seriously and whether the companies have an incentive to exaggerate the risks. They discuss why Roose believes the labs could coordinate a slowdown, and his prediction that the government may eventually intervene. The conversation also turns to Meta’s new AI agent, Google’s response, and what intensifying competition means for consumers. Roose shares what he’ll be watching as the next generation of AI models arrives.