🟢 LIVE - The Fable Backlash: Is This the Case for Decentralized AI?

Last week Anthropic launched Claude Fable 5 — and within 48 hours it became one of the messiest releases in frontier AI. A buried line in the model's system card revealed Fable would silently degrade its own answers for anyone it suspected of building competing AI systems, with no warning. Researchers called it silent sabotage. Anthropic apologized and reversed course. Add a 30-day data-retention requirement that got the model banned inside Microsoft, and you have a live referendum on the question crypto has been asking for years: should any single company hold this much control over frontier AI?

Loading this article…