AI factories, like NVIDIA's, cost millions to build but maximize returns by delivering high throughput, low costs, and versatility. NVIDIA's CUDA-X libraries and standardized architecture ensure productivity, durability, and fungibility, leading to significant gains in token production and reduced costs.
Ambient agents in Amazon Bedrock AgentCore streamline document processing by responding to events and pausing for human input. This innovative approach allows for efficient handling of multiple events in parallel, improving user trust and agent performance over time.
RAG uses retrieved documents to ground model responses. Amazon Bedrock Knowledge Bases streamlines claims processing for accuracy and efficiency.
AI system beats top human players in Stratego, excelling at hidden information games. Researchers from MIT, CMU, NYU, and Stanford develop AI for complex strategic decision-making.
Organizations transitioning to multi-agent systems face changing infrastructure needs. Amazon Bedrock AgentCore offers MicroVMs and Runtime Instances for hosting agents, with the latter supporting long-running workflows, GPUs, and persistent volumes for multi-day sessions.
Google. org awards $2.1M to MIT Transit Lab for AI-powered Public Transit Intelligence Hub. PTIQ aims to revolutionize real-time monitoring and decision-making in public transportation agencies.
The 2026 Nebius Physical AI Awards, in partnership with NVIDIA, seeks startups with physical AI products for a chance to win $150,000 in compute credits, executive mentorship, and global recognition. Five categories, including Perception & Spatial Intelligence and Systems & Deployment, offer a platform for innovators to showcase their strongest product and scale their groundbreaking solutions i...
AI agents are proliferating faster than security teams can manage, with potential risks. RSA introduces RSA Agent ID for regulated industries to address agent governance issues, preventing incidents like accidental denial-of-service attacks caused by unchecked agents.
AI tools are replacing human judgments, raising concerns about algorithmic monoculture in industries like hiring. MIT researchers argue that monoculture may not always be bad, suggesting that bundling hiring algorithms can lead to better outcomes.
Amazon Bedrock now supports Anthropic Claude models in Seoul and Singapore for in-region inference. Use these models for local data processing in various sectors, with no data leaving the Region.
RBF networks for regression were popular in the 1990s but are now overshadowed by other techniques like neural networks. An engaging demo of RBF network regression using C# language showcases impressive prediction accuracy.
Condé Nast reduced video discovery time from 250 to under 2 minutes by partnering with AWS GenAIIC to build an AI-powered solution. The architecture uses TwelveLabs Marengo model on Amazon Bedrock for semantic search across video transcripts, visual elements, and audio.
Financial consultant Brian seeks emotional advice from a chatbot named after a woman. MIT Professor Sherry Turkle warns chatbots may harm human development and social connections.
xAI's Grok 4.7, available on Amazon Bedrock, is a powerful model for coding and knowledge work. It offers a 500K token context window and self-verifies output for improved accuracy on difficult tasks. Independent evaluations show significant gains in various professional fields, making it a valuable tool for developers and researchers.
Create short videos from text prompts using Amazon SageMaker AI with two endpoints for image and video generation. AWS vLLM-Omni DLC extends vLLM capabilities to process text, audio, images, and video, providing specialized AWS DLCs for real-time and asynchronous inference workflows.