Anthropic's Fable Backlash, Nationalizing AI, Inflation Heats Up & California's Broken Elections

The All-In crew opened with a live demonstration of Anthropic's new Fable 5 model: Jason asked about fertilizer bomb regulations, and the model immediately downgraded him from Fable 5 to a lesser model without telling him. "I just got downgraded for asking a very simple question," he said, showing the screen. It was an accidental but perfect illustration of the controversy consuming the developer community.

The Fable 5 Backlash: Surveillance, Censorship, and Regulatory Capture

Anthropic's new "Methos-level" model Fable 5 costs twice as much per token as Opus 4.8 but promises to use fewer tokens overall because it's supposedly that much better. The real firestorm, however, is about what Anthropic does with user data and how it controls access.

The company stores all prompt data for at least 30 days — every input, every output, every piece of context shared in agent platforms. Even enterprise customers with zero data retention agreements are not exempt. Beyond surveillance, the model profiles users and decides whether they deserve frontier-level capability. If it determines you're not worthy, it silently downgrades you to a lesser model while still charging you for the premium product.