The Informed Layer · Newsscan
Daily briefing

AI agents roll out; export controls block model access

← All issues

Anthropic / Claude ecosystem

Anthropic Brings Artifacts to Claude Code for Shareable Live Pages from Coding Sessions

Claude Code now supports Artifacts, enabling teams to create live, auto-updating web pages that can be shared within an organization, with version history and access controls. This enhances transparency and collaboration in coding projects.

Frontier model providers

OpenAI Prepares for GPT-5.6 Model Release with Longer Processing Times

OpenAI is reportedly testing GPT-5.6 Pro, a variant with extended processing times and larger context for complex reasoning tasks. The upcoming release signals a potential enterprise-focused upgrade in computational power and pricing.

OpenAI Upgrades ChatGPT with Scheduled Tasks, Web Monitoring, and Smarter Notifications

OpenAI introduced Scheduled Tasks in ChatGPT, enabling automated background task scheduling and proactive web/app monitoring. The update replaces Pulse with a dedicated hub for task management and enhanced notifications.

OpenAI Unveils Workspace Agents for Continuous Office Operations

OpenAI introduced Workspace Agents designed to autonomously handle continuous office tasks, signaling a move toward persistent AI assistants. The agents are part of an effort to embed AI deeply into business workflows.

Google DeepMind Teases RT-3 with Advanced Generalization and Dexterity for Humanoid Robots

A teaser suggests RT-3, DeepMind's next robotic foundation model, will bring improved generalization and dexterity, potentially applicable across different humanoid hardware. Details remain scarce.

Meta Unveils AI-Driven Commerce and Discovery Tools at Cannes 2026

At Cannes, Meta introduced Live Video Ads, Live Shopping Tools, Virtual Cards for Checkout, and an expanded Affiliate Partner program, all leveraging AI to accelerate shopping experiences across Facebook and Instagram.

xAI Releases Grok Imagine Video 1.5 with Faster Clip Generation and Synced Audio

Grok Imagine Video 1.5 is now generally available via the Imagine API, offering faster 720p clip generation with synchronized audio. New creator workflow tools were also introduced, targeting video creators.

Mistral AI Plans to Offer AI Cloud Services, Expanding Beyond Enterprise LLM Deployments

Mistral AI announced plans to launch its own AI cloud services through Studio, aiming for sovereignty over AI infrastructure. This move expands its scope from providing LLMs to hosting them, challenging cloud giants.

Google Unveils Gemini-Powered Home Speaker, Its First Smart Speaker Since 2020

Google launched a new home speaker powered by its Gemini AI, designed for conversational interactions and smart home control. This revives Google's presence in the smart speaker market with advanced AI.

AI developer tooling & infrastructure

Cursor Launches Origin, a Git Forge Built for Parallel AI Agents

Cursor released Origin, a Git-compatible forge optimized for AI agents, aiming to compete with GitHub at the infrastructure layer. It supports parallel agent workflows and is designed for AI-native development.

LiteLLM Library Hit by Six CVEs and Supply Chain Attack; Urgent Fix Required

The LiteLLM library was compromised with six CVEs and a supply chain attack, allowing potential data exposure. Users must upgrade to version 1.83.14+ and rotate all credentials immediately.

Cloud & platform providers

Cloudflare Expands Agent Harness Support with Flue SDK

Cloudflare announced the addition of the Flue SDK to its agent platform, broadening the frameworks available for building and deploying AI agents on its edge network. This enhances developer choice and flexibility.

Cloudflare Introduces Agent-Powered Deployment with Cloudflare One Stack

The Cloudflare One stack now includes agent-powered deployment to automate learning, configuring, and managing Cloudflare Zero Trust environments. This simplifies secure access service edge (SASE) adoption.

AI policy, regulation & governance

EU Cybersecurity Agency ENISA to Meet Anthropic Under US Export Cloud

The EU cybersecurity agency ENISA was set to gain early access to Anthropic's Mythos model under Project Glasswing, but U.S. export controls potentially block such access. This creates a regulatory clash and raises concerns about over-reliance on American AI technology.

Anthropic and Google DeepMind CEOs Call for U.S.-Led AI Coalition at G7

At the G7 summit, Dario Amodei and Demis Hassabis urged international cooperation led by the U.S. to establish rules and standards for frontier AI. Canada responded positively, but no binding commitments were made.

UK Government Scales AI-Powered Planning System with National Rollout of APD and Extract

The UK is deploying the Augmented Planning Decisions (APD) tool and Extract nationwide to modernize local planning processes, powered by Google Cloud. This follows earlier development by DeepMind and aims to halve planning application times.

Japan's government announced an intention to continuously review and update AI laws to counter risks from advanced models like Mythos. This proactive approach signals heightened regulatory activity in Japan.

Australia's CISC Releases Enhanced CIRMP Rules Covering AI in Critical Infrastructure

The Cyber and Infrastructure Security Centre issued new rules to address AI, legacy systems, supply chain, and insider risks in critical infrastructure. These rules strengthen national security frameworks in the face of AI-driven threats.

Industry & market moves

Anthropic Opens Seoul Office to Expand Korean AI Ecosystem Ties

Anthropic officially launched its Seoul office to deepen collaborations with South Korean AI institutions and industry players. The move comes amid U.S. export controls on advanced AI models, signaling global expansion despite regulatory tensions.

Anthropic Becomes First AI Startup to Join Frontier Carbon Removal Coalition

Anthropic has joined the Frontier coalition, becoming the first AI startup to commit to an advanced market commitment for permanent carbon removal. This signals growing involvement of AI firms in climate initiatives and corporate sustainability.

Meta AI Restructuring Head Emily Dalton Smith Leaves Amid Workforce Overhauls

Emily Dalton Smith, who led Meta's AI restructuring, is departing the company as it undergoes major AI-centric reorganization and layoffs. Her exit highlights internal challenges in Meta's AI transformation.

PM Modi Meets Mistral AI CEO Arthur Mensch to Discuss Open-Weight Models and India Partnerships

Indian Prime Minister Modi and Mistral CEO Mensch discussed human-centric AI, open-weight models, and potential collaborations. This signals growing sovereign AI interest in India and Mistral's international expansion.

Bell, Cohere, Hypertec, and BUZZ HPC Form Landmark Deal for Sovereign AI in Canada

A consortium including Bell, Cohere, Hypertec, and BUZZ HPC announced a partnership to build a fully domestic, sovereign AI compute stack for Canada. This will enable secure, enterprise-grade AI workloads on Canadian soil.

SpaceX Acquires AI Coding Assistant Cursor for $60 Billion

SpaceX has formally exercised its option to acquire Cursor in a $60 billion all-stock deal, integrating the AI coding platform into its AI stack. This marks one of the largest AI software acquisitions, extending SpaceX's enterprise AI tools.

Robo.ai Proposes Acquisition of QC Capital to Bolster AI Robotics Network

Robo.ai Inc. announced a proposed acquisition of QC Capital Limited, an AI-driven technology holding and venture-building platform. The deal aims to expand Robo.ai's global AI robotics network and venture capabilities.

Stellantis, Wayve, and Uber Partner to Scale Robotaxis Globally

The three companies formed a global partnership combining Stellantis' vehicles, Wayve's autonomous driving AI, and Uber's mobility network to deploy Level 4 robotaxis. This aims to accelerate the commercialization of autonomous ride-hailing.

Convey Raises $38 Million Series A to Automate Enterprise Operations with AI Teammates

Convey secured $38 million in Series A funding to scale its AI teammates that automate enterprise operations. The investment signals confidence in AI-driven autonomous business process management.

Behavox Raises $175M from HPS to Fuel Global AI Compliance Growth

Behavox, an AI-based compliance and surveillance platform, raised $175 million in equity financing from HPS to drive global expansion. This marks its first equity round since 2020.

Accenture Acquires Majority Stake in Dragos and Two Others to Build End-to-End OT Security Platform

Accenture plans to acquire a majority stake in Dragos, along with runZero and NetRise, to create a comprehensive operational technology security platform. The move targets AI-driven and geopolitical cyber threats to critical infrastructure.

Vaisala Acquires AI Weather Company Atmo to Transform Meteorology

Vaisala is acquiring Atmo, Inc., an AI-driven weather forecasting company, for up to $130 million. This makes Vaisala the number one weather technology provider and brings advanced AI nowcasting in-house.

AI product & feature launches

Richtech Robotics Launches Livestream for ADAM AI-Powered Humanoid

Richtech Robotics started a livestream to showcase ADAM, its AI-powered humanoid robot, aiming to make it an embodied-AI 'influencer'. The robot is intended for hospitality, automotive, and manufacturing sectors.

Sony AI Releases Woosh, a Foundation Model for Sound Effect Generation

Sony AI launched Woosh, a dedicated foundation model for generating studio-grade sound effects, with both private and public versions. This tool can streamline audio production for games, film, and other media.

RAI to Demonstrate Cognitive Architecture with Persistent Self for Humanoid Robots

RAI Robotics plans to demo RAI Brain, a cognitive architecture that gives humanoid robots a stable identity and continuous reasoning capabilities. The demo at World Robot Conference 2026 aims to assess long-term self-development.

RealSense Unveils AI-Native D585 Pro Depth Camera with Edge AI for Robotics

RealSense launched the D585 Pro, its first AI-native depth camera featuring on-device edge AI processing and SDK-driven updates. It targets robotics and automation, enhancing perception for autonomous systems.

Midjourney Expands into Physical World with Full-Body Ultrasonic Scanner and Spa Network

Midjourney, known for AI image generation, is developing a full-body ultrasonic scanner for health scanning and plans to establish its own spas. This marks a dramatic pivot toward AI-driven medical imaging and wellness services.

Pinterest Releases Ask Pinterest App and MCP Tools to Reshape Product Discovery

Pinterest launched Ask Pinterest, a dedicated AI-powered app for product discovery, along with MCP tools for integration. This aims to keep users within its ecosystem and enhance advertising through AI.

Research with immediate practical relevance

Anthropic's Project Fetch Phase Two Shows Dramatic Acceleration in Autonomous Robodog Tasks

Anthropic released results from Project Fetch Phase Two, demonstrating Claude Opus 4.7 greatly accelerating autonomous robodog task completion compared to human-only approaches. The experiment showcases significant progress in embodied AI capabilities.

OpenAI's Pre-Deployment Test Replays Real Conversations to Spot AI Behavioral Drift

OpenAI developed a method that replays de-identified production conversations to detect behavioral drift in candidate models before release. This pre-deployment simulation aims to surface misbehaviors that synthetic prompts may miss.

New Medical AIs MIRA and AMIE Match or Exceed Doctor Performance in Preliminary Studies

Researchers found that MIRA and AMIE AI systems perform at least as well as physicians in diagnostic and treatment planning tasks. These results suggest AI could soon augment clinical decision-making.

MIT's DAAAM Memory System Lets Robots Remember Object Locations Using Natural Language Queries

MIT researchers developed DAAAM, a framework that attaches language descriptions to 3D maps, enabling robots to answer questions like 'Where did I leave my keys?' in real time. This persistent spatial memory could transform service robots.