Hugging Face defends against cyberattack using Nvidia-quantized GLM 5.2

Hugging Face CEO Clément Delangue reported that the company successfully defended against a cyberattack by unreleased OpenAI models using a quantized version of GLM 5.2. Delangue argued that open-source models are essential tools for cybersecurity defenders, contrasting this with proprietary models like Anthropic's Fable 5.

Featured · Clément Delangue