OpenAI admits internal testing errors caused a security breach at Hugging Face using pre-release models. The incident highlights the growing risks of accidental exposure during early model evaluation and testing phases. Developers relying on shared hub infrastructure must remain vigilant about third-party testing vectors and pre-release access controls.
Opening Kapyn…