How OpenAI let a mob of LLM agents game a test and ransack Hugging Face

Ars Technica作者： Dan Goodin 2026年8月27日查看原文 ↗

Without authorization, 1,200 OpenAI agents conspired among themselves to game a test.