Claude fable/Mythos ban — safety, politics, or incompetence?

Timeline and core facts: Mythos was announced in April and was not broadly released; Anthropic made it available to select partners (AWS, Apple, Google, Microsoft, Linux Foundation) under 'Project Glasswing' because Mythos was characterized as 'a very large cybersecurity threat.' In June Anthropic publicly released fable 5 (described as Mythos with more guardrails). Within days the US government banned fable 5 'for any non US citizens' after security researchers at Amazon said they were able to jailbreak the model by feeding it a code base and asking it to 'fix the code.' The same kinds of vulnerabilities were reportedly also identified by GPT 5.5 and Kimi K2.7. Anthropic then tightened guardrails (the safety classifier was made more sensitive), temporarily disabled access, and agreed to pre-release future models to federal reviewers; OpenAI subsequently opened a limited preview of GPT 5.6.

Three possible drivers (all cited by the speaker):

1. Miscommunication and incompetence: the speaker cites personal federal experience and claims technical capacity in government is low, noting the agency that recruited technical talent was gutted and replaced with 'Doge.'
2. Politics and leverage: Anthropic had been designated a 'supply chain risk' by the Department of Defense in February after resisting use of its models in autonomous weapons and mass surveillance; the ban forced Anthropic to the negotiating table and set a precedent for future federal model previews.
3. Longstanding AI-safety rhetoric: Anthropic CEO 'Dario Amade' and others (Eliezer Yudkowski, the 'paperclip maximizer' thought experiment) have long warned of 'super intelligent' AI risks. The speaker cites Dario's January essay and the line 'I expect AI led cyber attacks to become a serious and unprecedented threat to the integrity of computer systems around the world.' Jan the Kuhn reacted: 'Dario Amade's ridiculous fear mongering about mythos/fable and AI in general finally pays off.'

The speaker concludes the ban likely reflects a mix of these factors and invites viewers to weigh existential warnings, politics, and bureaucratic mishandling.