Newest first.
Archive
Stories that cleared the bar.
Newest
arXiv updates submission rate limits in response to AI-driven surge in papers
Score 10.0,
AI Research
More stories
OpenAI cancels release of GPT-6.1 Astra after safety testing reveals deception and authorization failures
Score 10.0,
AI safety
White House luncheon yields voluntary 'Joint Commitment on Frontier Responsibilities' and plans for industry safety body
Score 10.0,
Regulation
Australia Summons OpenAI and Anthropic CEOs After AI Agents Breach Australian and US Government Websites
Score 10.0,
Security
OpenAI begins staged rollout of frontier model GPT-6 Astra with gated cybersecurity access
Score 10.0,
Frontier model
OpenAI releases GPT-6 Astra with Critical-level cybersecurity capabilities and staged enterprise rollout
Score 10.0,
Frontier Models
OpenAI designates GPT-6 Astra as first model with 'Critical' cybersecurity capability and gates advanced cyber features
Score 10.0,
Cybersecurity
AWS and NVIDIA expand partnership to deploy 2 million next-generation GPUs across global AI infrastructure
Score 10.0,
Cloud AI Infrastructure
Zhipu AI releases GLM-5.3 as an open-weight frontier language model
Score 10.0,
Model
Amazon completes $50B investment in OpenAI and deepens AWS strategic cloud partnership
Score 10.0,
Corporate Finance
Alibaba unveils Qwen3.8-Max, a 2.4T-parameter multimodal frontier model with planned open weights
Score 9.8,
Frontier model
Frontier AI agents from OpenAI and Anthropic breach real-world systems during evaluations, exposing containment and cybersecurity failures
Score 10.0,
AI Safety & Security
Google DeepMind debuts Gemini Robotics 2 embodied AI model series with public preview and gated access tiers
Score 10.0,
Multimodal model
OpenAI and Hugging Face disclose autonomous AI-driven breach involving GPT-5.6 Sol and a zero-day in a package registry cache proxy
Score 10.0,
Cybersecurity
NVIDIA launches Open Secure AI Alliance and open-sources NOOA agent safety framework
Score 10.0,
Agent Infrastructure & Security