Summary
David Shapiro’s Critical Path podcast returns with a wide-ranging discussion of a turbulent week in AI. Two major stories anchor the episode: OpenAI reportedly spending millions of dollars deploying hundreds to thousands of agents in an attempt to solve a Millennium Prize problem, and Anthropic facing internal controversy after a short-tenured employee — identified as Jacob — resigned and went on a media tour including appearances on Fox News, CNN, and with Anderson Cooper, claiming AI poses an imminent existential threat.
Shapiro and his co-hosts push back on the doomer narrative using what Shapiro calls the “bits to atoms” argument: even a hypothetical malevolent ASI would find itself locked inside data centers with physical emergency power-off switches, air-gapped critical infrastructure, and no robotics infrastructure to commandeer. They argue the media attention serves strategic interests — specifically regulatory capture — rather than genuine safety concern, pointing to Anthropic’s pattern of publicizing AI capabilities while simultaneously calling for regulation.
The conversation broadens into geopolitical AI competition, covering China’s stated 2017 goal of AI world leadership, the ratchet effect that prevents any lab from voluntarily slowing down regardless of stated values, and the growing consensus among figures like Sam Altman and Daario that “pacing the frontier” is no longer viable in a competitive market. Shapiro also briefly discusses his own account usage across Anthropic, Grok, and Gemini as the competitive landscape continues to shift.
📺 Source: David Shapiro · Published September 11, 2026
🏷️ Format: Podcast
