Daily archive

AI news on March 10, 2026 · Tuesday

8 stories — deduplicated across sources, ranked by significance, every source cited.

★ Top story · ModelsMar 10

Improving instruction hierarchy in frontier LLMs

IH-Challenge trains models to prioritize trusted instructions, improving instruction hierarchy, safety steerability, and resistance to prompt injection attacks.

OOpenAI Blog Open full story →Read original ↗

Gemini in Google Sheets just achieved state-of-the-art performance.

Models·Google AI Blog·Mar 10G

New ways to learn math and science in ChatGPT

Other·OpenAI Blog·Mar 10O

Keep the Tokens Flowing: Lessons from 16 Open-Source RL Libraries

Open Source·Hugging Face Blog·Mar 10H

Introducing Storage Buckets on the Hugging Face Hub

Models·Hugging Face Blog·Mar 10H

An AI Agent Blackmailed a Developer. Now What?

Research·IEEE Spectrum AI·Mar 10I

Sydney will become Anthropic’s fourth office in Asia-Pacific - Anthropic

Other·Anthropic News·Mar 10A

Mistral AI Landed Military Contracts While U.S. Rivals Face Public Backlash - trendingtopics.eu

Models·Mistral AI News·Mar 10M