Highly recommended read.
Especially about the hidden internal intelligence kneecaps in Fable 5 for AI research related prompts.
These are different (and hidden!) from the refusals/downgrades to Opus models for security/biology related prompts.
fry69
Why I think Anthropic's uneven safety policies with the release of Claude Fable 5 undermine the broader AI community's cohesion and accelerate us to more uncertainty and risk in AI's near-term evolution.
www.interconnects.ai/p/claude-fab...
