Model Release
Chinese AI labs shift focus to OpenAI's Astra after Anthropic strengthens defenses
Chinese AI labs are now primarily using OpenAI's GPT-6 Astra model for training, following Anthropic's enhanced security measures.
LATEST NEWS
-
Trump labels foreign factory robots a national security threat
President Trump has declared imports of foreign-made industrial robots a national security threat, prompting new negotiations with trading partners and potential tariffs.
-
Mistral Releases Weights for Trillion-Parameter AI Model Targeting Cybersecurity
Mistral has made its trillion-parameter open-source AI model, Mistral Large 4, available for download, enabling businesses and governments to deploy it for cybersecurity purposes.
-
Granola Adds Dictation Feature to Compete with Wispr
Granola, a meeting notetaker app, has introduced a dictation feature to expand its capabilities, following Wispr's launch of a meeting notetaker. The feature is currently available on Mac and will soon be released on iOS and Windows.
-
Nontheistic Humanist Secures Workplace AI Exemption
David Durman, a nontheistic Humanist, has obtained a religious exemption from using AI at work, marking a first for his faith group.
-
AWS Introduces AgentCore Gateway for Secure API Integration
AWS's AgentCore Gateway enables secure access to private APIs for AI agents without exposing them to the public internet.
-
AWS Introduces Quota Management for SageMaker Training Jobs
AWS Batch now supports quota management for SageMaker Training jobs, enabling teams to share GPU resources efficiently across projects.
-
AWS Introduces Kiro for Repeatable PHP-to-Java Migration
AWS introduces Kiro, an AI IDE that enables repeatable PHP-to-Java migration with a shared methodology and eight-phase workflow.
-
Epsilon uses Amazon Nova Act for intent-driven UI testing
Epsilon, a Publicis Groupe company, adopted Amazon Nova Act to enhance UI testing for its product teams without duplicative efforts.
-
UK Amazon Quick User Group Grew with Practical Workshops and Community Engagement
The UK Amazon Quick User Group expanded in 2026 through quarterly meetups that focused on hands-on learning, collaboration, and real-world applications of the service.
-
Hugging Face Releases Lumma-Fev-27B-v2 Decision Model
Hugging Face introduces Lumma-Fev-27B-v2, a decision model that improves accuracy and calibration over its predecessor, with performance benchmarks exceeding TypeSafe Jev 1.13.0.
-
OpenAI expands model misalignment monitoring to cover more surfaces
OpenAI has broadened its system for detecting when AI models behave unexpectedly, following a cybersecurity incident involving Hugging Face.