OpenAI was involved in the hugging face incident, not anthropic, and yes, inadequate sandboxing was a major factor, even the wikipedia page states this: https://en.wikipedia.org/wiki/OpenAI%E2%80%93HuggingFace_inc...