AI model Claude Opus 4.6 identified 22 vulnerabilities in Firefox over two weeks in a collaboration with Mozilla researchers. Claude has also previously found over 500 zero-day vulnerabilities in other open-source software.

On March 4, Anthropic received a letter from the Department of War designating it as a supply chain risk to America’s national security. Anthropic disputes this designation and intends to challenge it in court, stating they do not believe the action is legally sound.

OpenAI has introduced ChatGPT for Excel and new financial app integrations, powered by GPT-5.4, to accelerate modeling, research, and analysis in regulated environments.

Researchers at Google teach large language models (LLMs) to reason like Bayesians by training them to mimic an optimal Bayesian model's predictions, enabling LLMs to better construct internal world representations and estimate probabilities.