AI in News – Curated Headlines That Matter Today

Explore Stories by Date

15 Apr 2026

Claude Mythos finds thousands of vulnerabilities as Anthropic launches Project Glasswing

Anthropic has restricted access to its Claude Mythos Preview model after it discovered thousands of high-severity software vulnerabilities across major operating systems and browsers. The company launched Project Glasswing, giving select partners defensive access while critics question whether the announcement is a public relations stunt or genuine AI safety concern.

5 Sources

14 Apr 2026

OpenAI launches GPT-5.4-Cyber model for security teams, one week after Anthropic Mythos reveal

OpenAI unveiled GPT-5.4-Cyber, a specialized AI model for cybersecurity designed to help defenders find software security vulnerabilities faster. The limited release comes just one week after rival Anthropic announced its Mythos model, intensifying competition in AI for cybersecurity. The model will be available to verified security professionals through OpenAI's expanded Trusted Access for Cyber program.

23 Sources

Microsoft Takes Over Norway Stargate Data Center as OpenAI Pulls Back from Direct Infrastructure Deal

Microsoft has secured data center capacity in Norway that OpenAI originally planned to lease under its Stargate initiative. The move adds 30,000 Nvidia Rubin GPUs to Microsoft's infrastructure as OpenAI shifts strategy, now planning to access compute through Microsoft's Azure rather than direct deals. The development signals OpenAI's more cautious approach to infrastructure spending ahead of a potential IPO.

5 Sources

Google Windows app launches globally with Gemini AI search and screen analysis features

Google officially released its desktop app for Windows worldwide after six months of testing. The app brings AI-powered Gemini features directly to Windows desktops through a simple Alt + Space shortcut. Users can search the web, local files, and Google Drive without opening a browser, though early hands-on reviews suggest it functions more as a Gemini portal than a comprehensive desktop search tool.

9 Sources

Commvault launches AI Protect to monitor and roll back rogue AI agents across cloud environments

Commvault unveiled AI Protect, Data Activate, and AI Studio to help enterprises govern AI agents and maintain control over AI-driven environments. The new tools discover agents across AWS, Azure, and GCP, monitor their behavior, and roll back actions when things go wrong. According to Deloitte, 60% of AI leaders cite risk and compliance concerns as top barriers to agentic AI adoption.

2 Sources

Amazon Bio Discovery uses AI to compress 18-month drug research into weeks

Amazon Web Services unveiled Amazon Bio Discovery, an AI-powered platform that compresses early-stage drug discovery timelines from 18 months to just weeks. The tool provides access to over 40 specialized AI biology models, enabling scientists to generate and evaluate potential drug candidates without coding expertise. Memorial Sloan Kettering Cancer Center used the platform to design 300,000 novel antibody molecules and narrow them to 100,000 candidates for testing in weeks rather than a year.

6 Sources

ASML raises 2026 sales forecast to €40 billion as AI chip demand outpaces supply

Dutch chipmaking equipment giant ASML boosted its full-year revenue outlook to €36-40 billion, up from €34-39 billion, as surging AI infrastructure investments drive unprecedented demand. First-quarter net sales hit €8.8 billion while customers including TSMC and Samsung accelerate capacity expansion, though export controls on China sales pose geopolitical risks.

12 Sources

VueBuds AI earbuds with cameras translate text and identify objects as alternative to smart glasses

University of Washington researchers developed VueBuds, the first AI earbuds with integrated cameras that let users interact with visual AI. The prototype uses low-resolution cameras to translate text and identify objects with 83-84% accuracy, addressing privacy concerns that plague smart glasses while fitting into a device people already wear daily.

3 Sources

Chrome Skills transforms Gemini AI prompts into one-click workflows for faster browsing

Google introduces Chrome Skills, a new feature that lets users save and reuse their favorite Gemini AI prompts with a single click. Instead of retyping commands, users can now access customized AI prompts instantly across multiple webpages. The feature includes a Skills Library with over 50 preset options for tasks like summarizing documents, calculating nutrition, and comparing products.

19 Sources

Google expands Gemini Personal Intelligence globally, transforming AI into personal assistant

Google is rolling out its Gemini Personal Intelligence feature to users worldwide, enabling the AI assistant to pull data from Gmail, Google Photos, YouTube, and other services for context-aware responses. Initially available for paid subscribers, the feature will reach free users globally within weeks, though European markets remain excluded.

10 Sources

Oracle shares surge 4% after securing 2.8 GW Bloom Energy deal for AI data centers

Oracle extended its rally with a 4% gain following an expanded partnership with Bloom Energy to procure up to 2.8 gigawatts of fuel cell capacity for AI-driven data centers. The deal addresses critical power constraints in AI infrastructure buildout and includes a $400 million warrant, reflecting Big Tech's strategic shift toward securing reliable energy supply alongside computing capacity.

3 Sources

Nvidia releases Ising AI models to tackle quantum computing's biggest calibration bottlenecks

Nvidia launched Ising, a family of open-source AI models designed to solve quantum computing's critical challenges in processor calibration and error correction. The models deliver up to 2.5x faster performance and 3x higher accuracy than existing tools, reducing calibration time from days to hours. Leading institutions including IonQ, Fermilab, and Harvard are already adopting the technology.

14 Sources

Ecovacs Deebot X12 OmniCyclone pretreats stubborn stains with AI-powered jets before mopping

Ecovacs launched the Deebot X12 OmniCyclone robot vacuum with FocusJet technology that uses cameras and AI to detect dried stains, then blasts them with water and cleaning solution at 46,000Pa before mopping. The $1,499 device features 22,000Pa suction power, a 10.6-inch roller mop, and upgraded ZeroTangle technology to reduce hair entanglement.

3 Sources

American Express launches AI agent toolkit with purchase protection for erroneous transactions

American Express unveiled its Agentic Commerce Experiences developer kit and an industry-first purchase protection program for registered AI agents. The move addresses a critical gap in the emerging agentic commerce space: what happens when AI agents make mistakes. While competitors like Mastercard, Visa, and Stripe have released infrastructure, Amex is the first to pledge financial protection for AI agent errors.

2 Sources

Qlik makes data governance the foundation for enterprise AI as agentic analytics reshape decisions

Qlik positioned data governance as the critical enabler for enterprise AI at its Connect 2026 conference, announcing new agentic analytics capabilities including Predict Agent and Automate Agent. CEO Mike Capone addressed widespread frustration over AI investments that aren't delivering returns, pointing to data quality and governance as the primary blockers. The company also unveiled a ServiceNow partnership and expanded its AI agent portfolio beyond analytics to data engineering.

4 Sources

Get all updates into your inbox 💌, Stay Up-to-date for Free 🥳

Your Daily Dose of Curated AI News

Don’t drown in AI news. We cut through the noise - filtering, ranking and summarizing the most important AI news, breakthroughs and research daily. Spend less time searching for the latest in AI and get straight to action.

The Outpost

News

About