AI News

Latest AI engineering news, updated daily.

In-depth tutorials and guides. Go to Blog →

Ai Agents

Frontier Agents Score Below 50% on SRE Task Benchmark

IBM Research and Artificial Analysis launched ITBench-AA, revealing that top frontier AI models score below 50% on complex enterprise SRE tasks.

Frontier Models · Site Reliability Engineering · Enterprise Ai

Ai Engineering

$8.2M Seed Backs Human Archive's Gig-Worker Robotics Dataset

Human Archive has raised $8.2 million to build a multimodal robotics dataset by paying Indian gig workers $1 per hour to record physical service tasks.

Robotics Data · Multimodal Datasets · Seed Funding

Ai Agents

Starlette BadHost Flaw Enables Auth Bypass in Python AI Agents

A critical HTTP Host header vulnerability in the Starlette framework allows attackers to bypass middleware authentication across the Python AI agent ecosystem.

Starlette Vulnerability · Python Security · Authentication Bypass

Ai Agents

Hugging Face Defines the Scaffold vs Harness Agent Architecture

Hugging Face has published a new technical glossary formalizing the structural differences between an AI agent's scaffolding and its execution harness.

Hugging Face · Agent Architecture · Technical Glossary

Ai Agents

Anthropic Moves Claude Mythos Toward Public Agent Access

Anthropic's autonomous vulnerability discovery model, Claude Mythos, has appeared in Claude Code, suggesting an upcoming public release for the restricted tier.

Anthropic · Claude Mythos · Autonomous Agents

Ai Agents

Android XR Launches With Gemini 3.5 Wearable Agent Support

Google's Android XR platform introduces a two-tier hardware strategy for smart glasses, relying on Gemini 3.5 to process multimodal agentic workflows.

Android Xr · Google Gemini · Multimodal Ai

Ai Engineering

DharmaOCR 7B Proves Domain Alignment Beats Parameter Scaling

Dharma-AI has released two specialized OCR models, demonstrating that targeted training history outpaces general-purpose frontier models on structured tasks.

Optical Character Recognition · Domain Alignment · Model Specialization

Ai Engineering

NVIDIA Nemotron-Labs-Diffusion Yields 6x TPF Over Qwen3-8B

NVIDIA has released the Nemotron-Labs-Diffusion model family, introducing a joint autoregressive and diffusion training objective to accelerate text generation.

Diffusion Models · Text Generation · Model Optimization

Ai Coding

Cursor Cloud Agents Shift to Isolated VMs and Durable Execution

Cursor has transitioned its AI agents to isolated cloud virtual machines with decoupled states and durable execution to handle multi-hour tasks.

Cloud Infrastructure · Durable Execution · Agentic Workflows

Ai Agents

AI Therapy App The Path Hits 95 on Vera-MH Safety Benchmark

Founded by Tony Robbins and Calm alumni, The Path launched with $14.3 million in seed funding to build specialized foundation models for clinical AI therapy.

Ai Therapy · Mental Health Tech · Foundation Models

Ai Engineering

SageMaker Endpoints Now Expose Native OpenAI Completions API

AWS updated Amazon SageMaker AI to expose standard OpenAI API paths, allowing seamless migration of agent workloads without custom SigV4 wrappers.

Aws Sagemaker · Openai Compatibility · Api Integration

Ai Coding

Google AI Studio Generates Native Kotlin Apps via Text Prompts

Google AI Studio now allows developers to build, test, and deploy native Kotlin Android applications entirely through natural language text prompts.

Google Ai Studio · Kotlin Development · Vibe Coding

Ai Engineering

Third-Party Devices Gain Gemini Home APIs and Reference Specs

Google released the Gemini built-in Program and Home APIs, allowing hardware partners to build devices with Gemini 3.5 and advanced computer vision.

Google Gemini · Home Apis · Hardware Integration

Ai Engineering

$45B Colossus Deal Secures 220K GPUs for Claude Inference

Anthropic will pay SpaceX $1.25 billion per month to lease the Colossus data center, securing 300 megawatts of capacity for Claude AI inference workloads.

Infrastructure · Gpu Clusters · Cloud Computing

Ai Engineering

Hark Secures $700M Series A for Universal AI Interface

Brett Adcock's AI laboratory Hark secured $700 million at a $6 billion valuation to develop a vertically integrated software and hardware platform.

Venture Capital · Artificial General Intelligence · Universal Interface

Ai Agents

Local Vision Agent IrisGo Automates Desktop Workflows via NPU

Backed by Andrew Ng, IrisGo has launched an ambient desktop agent that uses local computer vision to observe and automate cross-application workflows.

Computer Vision · Workflow Automation · Local Ai

Ai Agents

AWS Bedrock Agents Automate Pen-Testing and Incident Triage

AWS has launched two persistent Frontier Agents for automated penetration testing and SRE investigations built on Bedrock AgentCore and MCP.

Aws Bedrock · Automated Pentesting · Incident Triage

Ai Engineering

Apache 2.0 Gets 218B Command A+ as Cohere Acquires Reliant AI

Cohere expanded its sovereign AI strategy by open-sourcing the 218-billion parameter Command A+ model and acquiring biopharma startup Reliant AI.

Large Language Models · Open Source Ai · Sovereign Ai

Ai Agents

Volvo EX60 Routes External Camera Feeds to Gemini AI

Google and Volvo are integrating a specialized automotive version of Gemini into the EX60 SUV to process real-time external camera feeds for parking compliance.

Autonomous Vehicles · Computer Vision · Multimodal Ai

Ai Engineering

Project Canvas Renders Vector UI via Gemini 3.0 Ultra

Google introduced Project Canvas at I/O 2026, an AI design platform powered by Gemini 3.0 Ultra that generates editable multi-page marketing materials.

Generative Design · Vector Graphics · Multimodal Ai