Fireship
One man just liberated Fable... and now it’s illegal
Summary
AI summaries can be incomplete or wrong. Verify anything important against the original video.
Anthropic's release of Claude Fable 5 was followed by a government-mandated shutdown of the model due to security concerns, marking an unprecedented intervention in AI deployment.
Anthropic recently launched Claude Fable 5, a powerful AI model intended for general use, while restricting the more advanced and potentially dangerous 'Mythos' model behind a secure, partner-only program called Glasswing. An anonymous user successfully jailbroke Fable 5, enabling it to perform tasks the safety guardrails were designed to block, such as generating actionable cyber-attack procedures. Consequently, the US government issued an unprecedented export control directive requiring the immediate shutdown of Fable 5 and Mythos 5, impacting all users regardless of nationality or location. This event raises significant questions regarding the intersection of national security, corporate AI safety responsibilities, and the viability of government-enforced restrictions on deployed models.
Key claims
LockedKey Points
LockedWorth watching if: You want to understand the escalating tension between AI developers, regulators, and the open-source community regarding AI safety, security, and potential government intervention.
Sign in to unlock the full extract
Every claim, key point, and timestamp for this Fireship video — plus a daily email of every channel you follow.
Sign in with GoogleNo credit card. Free tier forever.