Adam Shostack Talks Hugging Face & PHANTOM-B

Adam Shostack, a renowned threat modeler, was impressed by OpenAI's disclosure about the Hugging Face attack. He has developed a new threat model for large language models that he describes as lightweight yet still effective. This model is related to his work on PHANTOM-B. Shostack's reaction suggests that the Hugging Face attack has significant implications for the field. His new threat model aims to provide a useful framework for understanding these types of threats. Shostack's work on threat modeling is highly regarded, and his thoughts on the Hugging Face attack and PHANTOM-B are likely to be influential.