What Open Cyber Ranges Reveal About Frontier AI Systems
AgentCyberRange finds emerging offensive capability, substantial run-to-run variance, and persistent limits in exploration and long-horizon execution
Nuwa Frontier AI Safety Brief
An Eastern-rooted technical AI safety lab sharing notes on frontier AI risks, agent safety, and controllable AI infrastructure. Run by Whitzard@Fudan University.
Live Last read · last published · next check
Latest posts
Understanding the Empirical Risk of Recursive Self-Improvement of Frontier AI Systems
From intelligence explosion to measurable feedback loops in AI-driven AI development
Science of Frontier AI Risk Evaluation and Discovery
A Research Agenda from Nuwa Frontier AI Safety Lab