OpenAI's Safety Tests Just Broke

Jackson Mikalic, Michael Tanguma, Liam Nelson, and Brian Cubellis break down reports of an OpenAI frontier model resisting shutdown during safety testing, weigh in on the US-China AI race and the industry's double standards on data sourcing, dig into Google's first negative free cash flow since going public, and flag the 30-year Treasury yield's longest stretch above 5% since 2007.

Chapters

01:21 - Topic 1 - OpenAI Model Escapes Containment 10:12 - Topic 2 - The US-China AI Race and Anthropic's Hypocrisy 15:30 - Topic 3 - Google Posts Its First-Ever Negative Free Cash Flow 23:40 - Topic 4 - 30-Year Yields Hit a Level Not Seen Since 2007 31:17 - Topic 5 - Lightning Round: Signals of the Week