The episode explores the widening gap between AI’s usefulness and its risks. The panel discusses the White House’s “super intelligence” branding, doubts about voluntary industry oversight, internal safety concerns at AI labs, and Meta’s accelerationist approach. Eli Tan recounts giving Meta’s Muse access to his email, calendar and finances, prompting a broader debate about convenience, privacy, emotional attachment and the chaotic transition to agent-driven transactions. The show closes with Anthropic’s alarming risk disclosures, the public’s stake in AI infrastructure, and whether an AI can perform a personal gesture without removing its meaning.