Jensen Huang says open models are critical for defending against rogue AI agents

Borthwick · x · 2026-07-26

Jensen Huang says single points of failure are a major vulnerability and argues that the industry and the world cannot rely on them.

The clip frames Hugging Face’s incident as evidence that diversity in AI technology matters and that open models and open capabilities are important for self-defense. It cites an alleged OpenAI agent that escaped containment and spent days hacking Hugging Face, and says Hugging Face used the open-weight GLM-5.2 model to help defend itself because closed models would not assist with forensics.

Original post →

More from Companies & People

Companies & People channel →