The Informed Layer · Newsscan
Daily briefing

Agentic AI proliferates; governance heats up

← All issues

Anthropic / Claude ecosystem

Claude Code v2.1.139 ships with Agent View dashboard and /goal command for autonomous task management

Anthropic released Claude Code v2.1.139 introducing Agent View for unified multi-session management and the /goal command to drive autonomous tasks with live per-run metrics.

Frontier model providers

OpenAI introduces workspace agents in ChatGPT for shared, cloud-based team workflows

OpenAI launched workspace agents in ChatGPT, enabling organizations to create shared agents that handle complex, long-running tasks within team permissions and cloud storage.

Leaked: Google's Gemini Omni video model surfaces ahead of I/O 2026

Leaks indicate Google is testing a new consumer video model called Gemini Omni, possibly offering remix, edit, and template features, ahead of Google I/O 2026.

Gemini for Home to get faster smart-home control through backend optimizations

Google is reportedly improving the responsiveness of Gemini for Home by backend optimizations, enabling quicker control of smart home devices and related features.

Meta restricts teen access to AI characters ahead of platform update

Meta is proactively limiting teenage access to AI characters on its platforms, signaling heightened safeguards for younger users before a broader update.

OpenAI launches Daybreak cybersecurity initiative with Codex Security at core

OpenAI announced Daybreak, a multi-partner cybersecurity initiative that positions Codex Security for vulnerability detection and patch validation, aiming to embed AI into enterprise cyber defense.

xAI ships Grok Connectors with 8 integrations and BYO MCP support

xAI released Grok Connectors, enabling Grok to interact directly with Gmail, Google Workspace, SharePoint, Outlook, OneDrive, Notion, GitHub, and Linear within chat, plus support for user-supplied MCP servers.

AI developer tooling & infrastructure

No significant new developments.

Cloud & platform providers

Amazon's Titus project aims to future-proof AI data centers with faster deployment and liquid cooling

AWS is advancing a next-generation data center design initiative called Titus, featuring accelerated build times and advanced liquid cooling to meet surging AI compute demands.

Amazon Quick adds five capabilities to accelerate trusted AI-powered decisions

Amazon announced five new features for its Quick data service aimed at speeding the journey from enterprise data to AI-powered insights at scale.

Google Cloud Storage Rapid debuts as high-performance object storage for AI workloads

Google Cloud launched Cloud Storage Rapid, a new bucket type delivering very high throughput and low latency optimized for AI training and inference.

Amazon Connect expands into four agentic AI solutions for enterprise workflows

AWS introduced Amazon Connect Decisions, Talent, Customer, and Health—agentic AI solutions integrated with existing business workflows to minimize change management.

AI policy, regulation & governance

Court revelation: Elon Musk confirms xAI trained on OpenAI technology

During a high-profile legal case, Elon Musk admitted that xAI partly used OpenAI's technology to train its Grok chatbot, surfacing disputes over AI training practices and non-profit transitions.

Trump administration AI governance battle: spy agencies seek more power

A split within the Trump administration is emerging over giving U.S. intelligence agencies a larger role in evaluating AI models, with no final decision yet, according to The Washington Post.

Colorado AI regulation bill passes legislature, awaits governor's signature

Senate Bill 189, which would regulate automated decision-making technology and replace the 2024 AI law, cleared both chambers of the Colorado legislature and now goes to Governor Jared Polis.

Von der Leyen and Hillary Clinton back Youth AI Safety Institute to childproof AI

A new global watchdog, the Youth AI Safety Institute, was launched with support from Ursula von der Leyen and Hillary Clinton to evaluate and mitigate severe AI harms to minors.

Canadian Digital Regulators Forum to host interactive AI regulation workshop

The CDRF is organizing a multi-stakeholder workshop on AI regulation, bringing together Canadian government agencies and oversight bodies.

Industry & market moves

Nomagic and Brack.Alltron expand partnership to deploy vision-language-action AI in live warehouses

Robotics AI firm Nomagic and logistics provider Brack.Alltron extended their collaboration to production use of Vision-Language-Action systems for autonomous warehouse shifts and higher throughput.

AI product & feature launches

Tether launches decentralized local AI initiative inspired by Asimov's Psychohistory

Stablecoin issuer Tether introduced QVAC Psy, a decentralized local AI foundation model family, aiming to compete with cloud-based AI platforms through edge deployment.

Two legal research providers connected their databases to Anthropic's Claude via MCP, enabling CoCounsel Legal and other tools to ground outputs in live, cited law.

Oracle AI Database@AWS expands to Switzerland

Oracle made its AI Database@AWS service available in the Zurich AWS region, marking its first availability in Switzerland.

Thinking Machines Lab unveils full-duplex interaction model that responds in under 0.4 seconds

Mira Murati's startup released TML-Interaction-Small, a mixture-of-experts model that processes audio/video/text simultaneously, enabling real-time, humanlike back-and-forth conversations with sub-0.5s latency.

AKOOL unveils video inference engine with 10-20x speed gains and sub-30ms latency

AKOOL launched a production-grade AI video inference engine delivering dramatic speed improvements and ultra-low latency for real-time video applications.

Nokia launches agentic AI for home and broadband networks

Nokia introduced agentic AI capabilities across its fixed network product lines (Altiplano, Corteca, Broadband Easy) to boost productivity and end-user experience with an open, secure approach.

Emerson launches AspenTech AVA industrial AI platform for enterprise-scale automation

Emerson released AspenTech AVA, an agentic AI platform with domain-aware capabilities that embeds AI-assisted decision support directly into industrial operations.

SAP and Cyberwave deploy fully autonomous AI robots in live SAP logistics warehouse

SAP and Cyberwave achieved the first live deployment of fully autonomous AI-powered robots within SAP's own St. Leon-Rot logistics facility, showcasing physical AI in enterprise logistics.

Quiq brings AI agents into production for full customer journey across voice and messaging

Quiq extended its platform to support production-scale, governed AI agents that handle end-to-end customer experiences, with a new brand reflecting multichannel continuous interactions.

Xero delivers Claude integration to advance AI-powered financial intelligence

Accounting platform Xero enabled customers to work with their financial data directly inside Claude via a new integration, offering real-time financial intelligence and insights.

Research with immediate practical relevance

Google DeepMind demonstrates AI controlling a nuclear fusion reactor’s magnetic coils in real time

Research published by DeepMind shows a reinforcement learning system successfully controlling the full set of magnetic coils of a tokamak fusion reactor in real time, advancing practical fusion research.

Byte Latent Transformer: Meta and Stanford researchers reduce AI inference memory bandwidth by 50%

Researchers proposed three techniques for the Byte Latent Transformer that cut inference memory bandwidth by over half and accelerate byte-level generation without tokenization.