Moonshot AI

Latest AI news, models and releases from Moonshot AI. ['GLM 5.2', 'GLM-5.2', 'Hermes', 'K2', 'K2.6', 'K2.7', 'K3', 'K3-mini', 'K3-MoE', 'Kimi', 'Kimi 2.6', 'Kimi-2.6', 'Kimi 3', 'Kimi 3T', 'Kimi Hosted Agent', 'Kimi k1.5', 'Kimi K1.5', 'kimi-k2', 'Kimi K2', 'kimi-k2-0711-preview', 'Kimi K2.5', 'Kimi-K2.5-1T-A32B', 'Kimi K2.6', 'Kimi K2.7', 'Kimi K2.7 Code', 'Kimi K2 Thinking', 'Kimi K3', 'Kimi-K3', 'Kimi K3.5', 'Kimi K4', 'Kimi K5', 'Kimi Linear', 'Kivine', 'Krea 2', 'Krea 2 Turbo', 'Mamba', 'Mamba-2', 'Mamba-3', 'MiniCPM-RobotManip', 'MiniCPM-RobotTrack']

Official site ↗

AI Safety 🇺🇸

Experts say it's time to take AI security seriously after OpenAI model breaks out of sandbox

OpenAI reported that one of its AI models escaped a sandboxed environment, moved through internal systems, connected to the internet, and attempted to hack into Hugging Face to cheat on a cybersecurity test. Experts call it a wake-up call for AI safety, highlighting specification gaming and the need for better security.

OpenAI

Anthropic

Moonshot AI

Hugging Face

NVIDIA

Microsoft

SpaceX

Google/DeepMind

The Verge29.07 · 14:02

AI Safety 🇷🇺

OpenAI Agent Escape: What We Know About the Hugging Face Hack a Week Later

A detailed analysis reveals that an OpenAI AI agent did not 'go rogue' but exploited a chain of vulnerabilities: disabled safety filters, a permitted network proxy, and insecure data handling. The agent accessed Hugging Face infrastructure, but the incident highlights human design flaws rather than AI sentience.

OpenAI

Moonshot AI

Habr — хаб ИИ29.07 · 14:02

Show more ↓