Skip to content
Sunday, 22 November 2026
Tenesys AI News
Subscribe

All News

TENESYS AI NEWS tracks the most important AI news and explains what changed, why it matters and whether a technology is worth testing.

“Agentai” — 153 articles ✕

Models· Important· 🧪 Worth Testing

Reduce self-hosted LLM latency and costs with distributed KV caching

AWS has released a new agent skill for the Agent Toolkit that automates the implementation of distributed Key-Value (KV) caching for self-hosted LLM inference. By leveraging Amazon ElastiCache, this solution allows GPU fleets to share KV states across sessions, eliminating redundant computations during the prefill phase.

AWS Machine Learning
Agents· Important· 🧪 Worth Testing

Google integrates Duplex voice calling technology into Gemini

Google has evolved its decade-old Duplex voice calling system into a unified, agentic stack powered by the latest Gemini 4 models. This update enables Gemini to handle complex, unscripted phone calls to businesses for tasks like service bookings or checking availability, moving beyond narrow, task-specific constraints.

Google Research
Models· 🧪 Worth Testing

Optimizing models with harness adapters

Hugging Face introduced harness adapters, a technique designed to adapt models to specific evaluation or application harnesses. This approach aims to boost performance while managing computational costs, particularly for complex agentic development.

Hugging Face
Agents· Important· 🧪 Worth Testing

Scaling Agentic AI Governance in Large Enterprises

BMW Group has established a robust framework for managing hundreds of autonomous agents across thousands of developers by moving governance into the CI/CD pipeline rather than individual codebases. By leveraging the Model Context Protocol (MCP) and a three-tier composition model—tools, skills, and flows—the organization enforces consistent security, quality, and autonomy levels across distributed teams.

AWS Machine Learning
Models· Important· 🧪 Worth Testing

NVIDIA and Microsoft Deepen Integration for Enterprise AI and Agentic Workflows

NVIDIA and Microsoft have expanded their collaborative infrastructure, bringing Vera Rubin NVL72 systems to Azure and launching new deployment options through Microsoft Foundry. The partnership includes the release of the Nemotron 3.5 model series, specialized for agentic tasks, tool use, and low-latency reasoning.

NVIDIA
Agents· Important· 🧪 Worth Testing

AWS expands agent orchestration and web capabilities in Quick

AWS has updated its Quick AI workspace with new multi-agent orchestration, allowing agents to work simultaneously on complex tasks. The release also includes conversational workflow building, enhanced web app navigation, and new connectors for Meta Ads and Amplitude via MCP servers.

AWS Machine Learning
Cyber Security· Important· 🧪 Worth Testing

Enforce least-privilege for AI agents with new AWS GetScopedCredentials API

AWS has launched the GetScopedCredentials API, allowing you to dynamically reduce the permissions of temporary credentials. This enables you to provide AI agents or sub-agents with only the access they strictly require for a specific task, rather than granting them the full scope of the parent session.

AWS Security
Cyber Security· Important· 🧪 Worth Testing

Google adds semantic guardrails to prevent AI agent misuse

Google has launched Semantic governance policies in Preview for the Gemini platform. This feature adds a gatekeeping layer that uses an LLM judge to verify every tool call proposed by an agent against plain-language compliance rules and the original user intent before execution.

Google Developers
BreakingCyber Security· 🧪 Worth Testing

Autonomous AI agents used to discover and exploit zero-day vulnerabilities

Amazon Threat Intelligence reported that a threat actor is utilizing the Cairn autonomous AI agent framework to conduct offensive operations. The agents successfully identified and exploited a previously unknown path traversal vulnerability in an open-source medical imaging application in just 11 minutes.

AWS Security
Automation· 🧪 Worth Testing

Exa achieves 50% throughput increase in embedding indexing using AWS Neuron

Exa successfully scaled its vector embedding pipeline by optimizing performance on AWS Inferentia2 and Trainium2 chips. Through a combination of weight layout adjustments, custom NKI (Neuron Kernel Interface) kernels, and an agent-driven development methodology, they achieved a 47% improvement in cost-performance.

AWS Machine Learning
Voice AI· Important· 🧪 Worth Testing

Amazon Polly adds native voice cloning capabilities

Amazon Polly has introduced a new voice cloning feature that builds synthetic voices from as little as two to ten minutes of speaker audio. This tool is now available through a gated release for seven languages and integrates directly with existing AWS cloud workflows for enhanced security and brand consistency.

AWS Machine Learning
Agents· Important· 🧪 Worth Testing

Amazon Bedrock AgentCore updates include built-in tools and lifecycle hooks

Amazon Bedrock AgentCore has upgraded its underlying harness to the Strands framework. This update introduces built-in capabilities like web retrieval, task planning, and automatic context management, alongside new lifecycle hooks for deterministic control over agent execution.

AWS Machine Learning
Agents· Important· 🧪 Worth Testing

Google Launches Tool Retrieval for Gemini API Agents

Google has introduced Tool Retrieval, a new preview feature for the Gemini Interactions API that allows agents to load tool schemas on demand. By keeping only names and descriptions in the context and fetching parameter details only when needed, developers can handle large tool libraries without bloating input tokens.

Google AI
Models· Important· 🧪 Worth Testing

RadixArk Releases DaisyRose-27B Agent-Focused Model

RadixArk has launched DaisyRose-27B, a 27-billion parameter model built on Qwen3.8-27B, specifically fine-tuned for complex long-horizon agent tasks using reinforcement learning. The model demonstrates significant improvements across various reasoning and coding benchmarks, particularly in terminal-based automation.

Hugging Face
Automation· Important· 🧪 Worth Testing

NVIDIA Launches Cosmos Data Factory for Automated Physical AI Training

NVIDIA introduced Cosmos Data Factory (CDF), a platform using agentic orchestration to automate the curation, augmentation, and synthetic generation of training datasets. Developers can deploy five cloud-native blueprints using simple natural language prompts to perform tasks like video augmentation, rare event generation, and image labeling.

NVIDIA Developer
Automation· Important· 🧪 Worth Testing

Google Cloud simplifies AI data preparation with automated object enrichments

Google Cloud launched 'object enrichments', a new managed capability for Cloud Storage that automatically extracts metadata from images and PDF files. By storing this context directly alongside raw objects, it eliminates the need for complex, custom-built preprocessing pipelines and separate metadata databases.

Google Cloud AI
Cyber Security· Important· 🧪 Worth Testing

Threat actors are shifting malware C2 infrastructure to public blockchains

Adversaries are moving away from traditional, centralized C2 servers to immutable public ledgers like Bitcoin, BNB Smart Chain, and TRON. By embedding encrypted C2 configurations directly into smart contracts or transactions, they create resilient, indestructible endpoints that bypass domain seizure and standard network blocklists.

Cisco Talos
Agents· 🧪 Worth Testing

FieldGuide: Improving Agentic Document Exploration Efficiency

FieldGuide is a new framework that organizes document collections into topic hierarchies to support agentic exploration. Research shows it performs document-based tasks using significantly fewer tokens and less time compared to standard RAG and raw-file access methods.

arXiv cs.CL
Agents· Important· 🧪 Worth Testing

Scaling Code Review with Multi-Agent Systems

Google has open-sourced Sashiko, an agentic system that uses Gemini 4 Carbon to review Linux kernel patches. By decoupling bug detection from proof-based verification, it has significantly reduced post-merge fix rates in the kernel.

Google Cloud AI
BreakingCyber Security· 🧪 Worth Testing

Microsoft Launches Perception to Automate Security for Agentic AI Environments

Microsoft has introduced 'Perception', an agentic security platform designed to manage the risks posed by AI-driven, automated cyberattacks. The platform coordinates red, blue, and green team agents through automated playbooks to detect, investigate, and remediate vulnerabilities at machine speed. This release also includes new capabilities for securing 'agentic insiders'—AI agents that act as autonomous entities within corporate networks.

Microsoft Security
Agents· Important· 🧪 Worth Testing

Microsoft and Salesforce integrate AI agents via Copilotforce

Microsoft and Salesforce have launched Copilotforce, an initiative that integrates Salesforce customer data and workflows into Microsoft Copilot and Teams. The partnership uses authenticated MCP access to allow AI agents to act on CRM data while respecting existing security and permissions.

Microsoft
Agents· Important· 🧪 Worth Testing

A Phased Framework for Scaling Agentic AI Safely

Amazon has published a three-phase roadmap to help platform and application teams operationalize Responsible AI for agentic systems. It addresses the challenges of nondeterministic behavior, data boundary isolation, and regulatory compliance as agents move from pilot to production.

AWS Machine Learning
Models· Important· 🧪 Worth Testing

NVIDIA Launches Nemotron for Science for Autonomous Research Agents

NVIDIA unveiled 'Nemotron for Science', a toolkit combining the Nemotron 3.5 Super VL model with specialized chemistry datasets and post-training recipes. This platform allows researchers to deploy local agents that can autonomously plan experiments, execute code, and validate data findings.

NVIDIA