AWS Introduces Managed Knowledge Base for Amazon Bedrock
AWS now offers a Managed Knowledge Base for Amazon Bedrock, simplifying the migration process and enhancing retrieval capabilities for users.
AWS now offers a Managed Knowledge Base for Amazon Bedrock, simplifying the migration process and enhancing retrieval capabilities for users.
AWS announces updates to its AgentCore runtime, enabling faster startup times, scalable compute, and more efficient resource usage for AI agents in production environments.
AWS provides a new method for developers to access foundation models via Amazon Bedrock using LiteLLM, offering centralized control and observability.
AWS outlines a framework for calculating true model costs, including hidden cost drivers and techniques for reducing cost per task.
AWS now allows tracking and capping Amazon Bedrock spend per IAM principal using CUR 2.0 and AWS Budgets without custom code.
Amazon Bedrock's inference costs can vary based on settings like cache duration, service tier, and reasoning effort. These choices affect token processing and billing without altering the model or prompt.
Natera implemented an agentic AI system using Amazon Bedrock AgentCore to address knowledge continuity challenges in lab operations, improving response times and automation.
AWS has made SGLang-Neuron publicly available, enabling users to deploy large AI models on its Trainium AI chips using the same SGLang interface.
Red Hat introduces new tools and partnerships to help enterprises manage AI workloads across hybrid environments and reduce costs.
Google introduces Agent Substrate, an open-source project designed to securely isolate AI agents on Kubernetes clusters, addressing enterprise security and efficiency needs.
Google has donated OpenRL, a tool for model fine-tuning through reinforcement learning, to the Cloud Native Computing Foundation. The tool aims to simplify infrastructure management for AI researchers.
A security firm found nearly 95% of Grok-generated web apps were publicly accessible without login, exposing patient data, passwords, and internal sales pipelines.