The Anthropic ‘Fable’ saga proves: we have opened the AI Pandora’s box. What now? | Nathan E Sanders and Bruce Schneier
TL;DR
Anthropic released Fable on 9 June as a constrained version of Mythos. Three days later, the US government classified it as a dangerous munition and blocked access for foreign nationals. Because Anthropic could not reliably separate US users from non-US users, it shut Fable off for everyone. Schneier’s argument: that may pause one product, but it does not stop the capability trend.
Nauti's Take
The Fable case shows how brittle symbolic AI control becomes once capabilities move around. Shutting down one model looks decisive, but it is mostly a pause if the same patterns return through tools, agent frameworks and open-source replication.
The hard part is that safety is not just better filters; it means limiting blind delegation to systems that aggressively optimize goals. That is where the problem becomes politically, economically and technically uncomfortable.
Briefingshow
The piece moves the debate from banning one model to asking how much agency we attach to AI systems. Export controls are a weak tool if similar capabilities can be rebuilt through better harnesses, open models and multi-model setups. Safety becomes an infrastructure and governance problem, not just a model problem.
Sources