Urgent.News

What's breaking now, across thousands of outlets.

AI

OpenAI's research chief talks of ‘cultural reset’ after wild few weeks

Nature, Published online: 25 September 2026; doi:10.1038/d41586-026-03028-9 Mark Chen discusses how the Hugging Face cybersecurity incident prompted a pivot to safety — and giving AI models a sense of 'taste'.

We haven't written up this one. Nature has the full story — the link below goes straight to it.

Read the original at nature.com →

More in AI

The Silent Killer of AI Agents: Why Your Evaluation Metrics Are Lying to You

Originally published on tamiz.pro . The Metric Mirage You’ve shipped your AI agent. It aces the benchmark, clears every test case, and your dashboard glows green.

  • Evaluation metrics often lie, not reflecting true agent performance.
  • Proxy metrics like accuracy mask catastrophic real-world failures.
  • Shift focus to final-world state checks, not intermediate metrics.

More from Friday 25 September →