Anthropic Sparks Backlash Over Claude Fable 5 Limits, Restricting AI Research and 30-Day Data Retention
Updated
Updated · Decrypt · Jun 10
Anthropic Sparks Backlash Over Claude Fable 5 Limits, Restricting AI Research and 30-Day Data Retention
3 articles · Updated · Decrypt · Jun 10
Summary
AI researchers and developers on June 10 blasted Anthropic after Claude Fable 5 appeared to underperform or refuse prompts tied to AI development, with some calling it a silent degradation of the model.
Screenshots shared online pointed to limits on topics such as pretraining pipelines, distributed training infrastructure and ML accelerator design, fueling claims that Anthropic had narrowed the model's usefulness for research work.
The anger widened beyond model behavior: users also criticized a 30-day data-retention policy, saying it could block adoption by European companies and others that require zero data retention.
Researchers said the change damaged trust in Anthropic, with some threatening to shift spending to open or unrestricted models, turning the launch into a broader dispute over access, openness and control in AI tools.