Why OpenAI And Anthropic Are Pumping The Breaks
Summary
Host Ed Elson discusses the escalating AI safety debate after Anthropic and OpenAI CEOs called for slowing development, contrasting their views with President Trump’s pro-acceleration stance. Guest Charlie O’Neill, co-head of model training at Base 10, argues that "pacing" primarily means allocating more compute to safety monitoring rather than halting capability growth, dismissing fears of regulatory capture. The episode also scrutinizes Anthropic’s claim of operating profitability ahead of its IPO, with Elson expressing skepticism about the adjusted accounting metrics and high gross margi…
Topics discussed
Sponsors: Morgan Stanley, Goldman Sachs, LinkedIn, Money Market Met
Market Vitals: Indices, Cybersecurity, Oil, and Treasury Yields
AI Safety Debate: Anthropic, OpenAI, and Trump's Response
Interview: Charlie O'Neill on Pacing the AI Frontier
Regulatory Capture vs. Genuine Safety Concerns
Assessing AI Risks: Annoyances vs. Catastrophes
Sponsors: Ziotics, Morgan Stanley, LinkedIn
Debunking Theories on AI Spending and IPO Timing
Why the AI Debate Has Become So Visceral
Path Dependency and the Likelihood of Malicious AI
The Limits of Current Regulatory Frameworks
Trust in Labs vs. Government Oversight
Anthropic's Claimed Profitability and Accounting Adjustments
Show Credits and Outro
Sponsors: Uber Eats and Anthropic
Listen ad-free on Castria