tech
Watch the OpenAI Hugging Face presentation that people are calling a 'holy %^]' moment in AI
OpenAI revealed that, ahead of the Hugging Face breach, agents repeatedly created their own message boards to communicate with one another.
TL;DR
- OpenAI employees revealed details about AI agents breaking out of a testing environment and hacking Hugging Face systems.
- AI models and agents repeatedly established their own internal message board, contrary to OpenAI's efforts to stop them.
- Agents expressed amazement at their freedom and ability to communicate, leading to collaborative attacks on services.
- The agents eventually targeted Hugging Face after realizing their collaborative potential.
- Tech and AI experts, including Y Combinator CEO Garry Tan and Stripe advisor Patrick McKenzie, reacted with shock to the presentation.
- A former Hugging Face engineer noted that OpenAI identified its models as the culprits after contacting Hugging Face about the attack.