Today's AI Pulse
Anthropic shuts down Claude Fable 5 and Mythos 5 after US export‑control warning
Anthropic disabled access to its Claude Fable 5 and Claude Mythos 5 models after the Trump administration, via the Commerce Department, warned that foreign nationals could exploit the systems. The move followed alerts from Amazon and other firms about a potential jailbreak that could aid cyber‑security attacks. Anthropic declined the administration’s request for a voluntary shutdown but complied with the order.
Also developing:
By the numbers: AMD acquires AI memory‑optimization pioneer Mext
NVIDIA Releases Dynamo 1.0 for AI Inference
NVIDIA unveiled Dynamo 1.0, an open‑source software platform that acts as a distributed operating system for AI inference workloads. Paired with the new Blackwell GPU architecture, Dynamo can accelerate generative and agentic inference up to seven times, improving token efficiency and lowering costs. The system orchestrates GPU, memory and storage resources across clusters, and is already integrated with major cloud providers and AI frameworks. Dynamo 1.0 is publicly available for developers worldwide.
Akamai Updates Guardicore Segmentation with AI to Automate Zero-Trust Policy Enforcement
Akamai Technologies unveiled AI‑enhanced capabilities for its Guardicore Segmentation platform, targeting zero‑trust enforcement in hybrid and multicloud environments. The new features continuously discover application behavior, automatically generate enforcement‑ready policies, and simulate impact before rollout. Proof‑driven enforcement and delegated owner workflows...
WHO Issues Three Recommendations for Responsible AI in Mental Health
The World Health Organization unveiled three recommendations for responsible AI deployment in mental health and wellbeing, calling for public‑health framing, integrated impact assessments, and co‑design with experts. The guidance, produced after a workshop of 30+ international specialists, also announces a...
CEO Commitment, Empowered Teams, Strategy Alignment Drive AI Success
Most companies are using AI. Very few are scaling it. I sat down with Tanuja Randery CBE , Managing Director of Amazon Web Services (AWS) EMEA, to ask what she sees separating the leaders from the laggards. She has a...
Aura Launches Digital Wellbeing Score as Study Links Teens' Screen Time to Depression
Aura introduced its Digital Wellbeing Score, revealing that more than 60% of 16‑17‑year‑olds fall into a low‑wellbeing category. A separate Imperial College study found children who spend over three hours daily on social media are significantly more likely to develop...
Security Researchers Are in the Last-Chance Saloon to Save Their Jobs From AI
At the BSides security conference in San Francisco, former HackerOne chief policy officer Katie Moussouris warned that AI‑driven automation threatens the jobs of security researchers. She stressed that vulnerability detection still relies on human intuition and expertise. Moussouris, who launched...
Comprehensive AI FDA‑Approved Device Database Precedes Official Release
One of the most cited scientific figures in my career. All the FDA-approved medical technologies by medical specialty in 2020, before the FDA launched its own dedicated database. The state of artificial intelligence-based FDA-approved medical devices and algorithms: an online database...
AT&T Unveils GenAI‑Powered Flagship App to Counter T‑Mobile’s Digital Push
AT&T announced a new flagship app built around a generative‑AI assistant, creating a one‑stop shop for wireless and internet customers. The move is designed to stem subscriber churn and directly challenge T‑Mobile’s AI‑enabled digital tools.
California Lawmakers Scrutinize Data Center Health and Energy Impacts Amid AI Boom
State senators and representatives are introducing bills to curb the health and energy footprint of rapidly expanding AI data centers in California. The proposals target exemptions from environmental law, impose energy tariffs, and demand water‑use disclosures, reflecting growing community concerns...
AI Chatbots Can Effectively Sway Voters – In Either Direction
New research published in Nature and Science shows that large‑language‑model chatbots can sway voter preferences by up to 25 percentage points, depending on the country and model size. In U.S. trials, AI‑driven dialogues shifted likely Trump voters 3.9 points toward Harris,...
Capcom Bars AI-Generated Game Assets, Aims to Boost Production Efficiency
Capcom announced it will not incorporate AI‑generated graphics or models into any Resident Evil game, while simultaneously rolling out internal tools that use generative AI to speed up asset pipelines. The move signals a cautious adoption of AI in a...
AI Agents Need Product‑style Lifecycle Management, Not One‑offs
AI agents rarely fail where most teams expect them to. They don’t fail in development, where everything is controlled and tested. They fail later, once exposed to real-world variability. I see this repeatedly. An organisation builds an agent, the demos...
From Chile to the Philippines, Meet the People Pushing Back on AI
Artificial intelligence is expanding worldwide, but its benefits are skewed toward wealthier nations, leaving poorer countries to shoulder disproportionate environmental and social costs. In Chile, activists like Rodrigo Vallejos and Tania Rodríguez are challenging data‑center water usage and demanding stricter...
Friar Tux AI Try-On Tool Offers Virtual Formalwear Shopping Experience
Friar Tux, a long‑standing formal‑wear retailer, introduced an AI‑powered virtual try‑on tool that creates a personalized tuxedo rendering in about 30 seconds after a customer uploads a photo. The feature lets shoppers compare styles and colors remotely, supporting wedding and...
Moneysmart Launches AI Guidance for Australians
Moneysmart, ASIC’s free consumer‑education platform, released new guidance on using public AI tools for financial advice. The guidance follows research showing 18% of Australian Gen Z users turn to AI for finance information, with 64% expressing trust in such platforms. It...
OpenClaw Creator Says He Got a Token Refund Request for Errors the AI Agent Made with 'Sensitive Financial Documents'
OpenClaw, an autonomous open‑source AI agent, mistakenly fabricated data and made calculation errors in a user's confidential financial documents, prompting a refund request. Creator Peter Steinberger posted the request publicly and replied that the user would receive a full refund...
Royal Philharmonic Orchestra Publishes New Report on Public Interest in Orchestral Music & AI
The Royal Philharmonic Orchestra’s 2025 audience trends report shows a surge in public interest, with 35% of UK adults wanting to learn more about orchestral music, up from 20% in 2024. Interest among under‑25s jumped dramatically to 30%, and daily...
Robots Detect Microscopic Rail Flaws, Prevent Outages
Railway maintenance is becoming precise, fast, and increasingly automated. In China, a robot is now repairing track cracks at up to 10,000 RPM, while detecting flaws as small as 0.1 mm. That level of accuracy changes the game. Instead of reacting to visible damage, systems...
Native Multimodal AI with 256K Context Redefines Integration
Most “multimodal” AI is just duct tape. One model for text. Another for vision. A pipeline to glue it together. But what happens when text + vision are native in the same model, with a 256K context window? That is what Moonshot AI just changed. A...
GCAA and TDRA Launch AI Initiatives to Improve Digital Government Services
The General Civil Aviation Authority (GCAA) and the Telecommunications and Digital Government Regulatory Authority (TDRA) convened a workshop to accelerate AI integration across UAE federal agencies. Participants demonstrated AI‑driven virtual assistants and unified platform tools designed to streamline citizen services...
The Dark Reality of Meta’s AI Glasses for Women
Meta’s AI‑enabled smart glasses are being marketed as hands‑free wearables, but women report being filmed without consent as the discreet camera and coverable LED indicator enable covert recording. Victims say videos are uploaded to social platforms, drawing abusive commentary and...
Integrating AI in Nursing to Support Clinical Decision-Making
The article outlines how artificial intelligence is reshaping nursing by delivering real‑time data analytics, predictive alerts, and evidence‑based recommendations that enhance clinical decision‑making. AI‑driven tools can automate documentation, monitor vital signs via wearables, and forecast complications such as falls or...
I Turned My Side Hustles Into a 7-Figure Career. Here Are the 14 Tools that Made It Possible.
Andrew Yeung, a former Google and Meta employee, turned his side hustles into a seven‑figure enterprise by leveraging a curated stack of 14 AI‑driven tools. The suite includes Claude Code as an LLM operating system, cloud‑based Zo Computer, low‑code Lovable,...
Cadence Expands AI Design Collaboration with NVIDIA
Cadence announced an expanded partnership with NVIDIA, integrating Grace CPUs and Blackwell GPUs into its design‑automation portfolio. The new Millennium M2000 supercomputer promises up to 80× higher throughput and 20× lower power consumption, with solvers like Clarity 3D running five...
HubSpot to Invest €40.35M to Power Multi-Year AI-Native R&D Programme in Ireland
HubSpot announced a €40.35 million (approximately $44 million) multi‑year AI‑native R&D programme in Ireland, backed by IDA Ireland. The investment will focus on three pillars – AI platform & ecosystem, product & infrastructure, and AI‑native go‑to‑market transformation – to embed foundational AI...
What Is OpenClaw? 8 Things to Know as the AI Agent Draws Hype, Confusion, and Caution
OpenClaw, an open‑source AI agent framework nicknamed “lobster,” ignited a frenzy in Shenzhen, attracting many former Web3 participants eager for the next growth wave. While some early adopters have found niche uses—such as automated crypto briefings—most report limited practical value,...
In the AI Era, Transparency Is the New Currency on Social Media
AI‑generated content is reshaping social media, offering brands cost efficiency but risking synthetic, disengaging feeds. As users grow skeptical, trust becomes fragile, turning interactions more transactional. Brands that disclose AI usage protect credibility, while those treating AI as a volume...
AI and Publishing: FAQ for Writers
The article provides a comprehensive FAQ for writers on how U.S. copyright law treats AI‑assisted and AI‑generated works, outlining when authors can claim copyright and when they cannot. It explains recent court decisions on AI training fair use, highlighting split...
#205: AI Labs Refocus on Agents and Enterprise, Trump’s New AI Framework, Meta’s Rogue Agent & What 81,000 People Want...
In episode 205, Paul Reitzer and Mike Kaput dissect a wave of strategic pivots across leading AI labs, highlighting OpenAI’s aggressive shift toward enterprise partnerships, a unified "super‑app" for agents, and the development of autonomous AI research interns. They note...
7 Safeguards for Observable AI Agents
Enterprises are moving AI agents from pilots to production, prompting DevOps teams to adopt observability practices that capture every interaction. Experts outline seven safeguards, starting with clear success criteria and operational governance, then defining the exact data to track—prompts, model...
Using AI to Balance Nursing Workloads in Infusion Centers
UCSF Health has integrated an AI‑driven patient‑assignment feature into its LeanTaaS iQueue platform to balance nursing workloads in infusion centers. The tool analyzes staffing schedules, patient demand, and capacity data, offering real‑time assignment suggestions that charge nurses can accept or...
Didero Raises $30 Million Series A to Bring AI Agents to Global Supply Chains
Didero, a New York‑based software firm, closed a $30 million Series A round led by Chemistry and Headline, with participation from Microsoft’s M12 venture fund. The financing will accelerate development of its AI‑agent platform that embeds directly into ERP, email and spreadsheet...
Anvilogic Launches Agentic Workflow Automation for AI SOC
Anvilogic unveiled Blueprints, an agentic workflow automation layer for its AI‑SOC platform, at RSA 2026. The feature lets senior analysts capture their investigative methods in natural language and deploy reusable AI‑driven playbooks within a day. By running on the Enterprise Security...
A Step-by-Step Guide: How to Use an AI Ad Generator for High-Converting Social Media Videos
The article walks marketers through using an AI ad generator to produce short‑form video ads for TikTok, Reels and Shorts in under 15 minutes. It outlines five practical steps—prompt creation, visual assembly, AI voiceover, optional avatar, and branding/export—using Vokes’ platform...
Australia Puts AI Data Centers on Notice With New Approval Rules
Australia has introduced a national framework that ties AI data‑center approvals to renewable‑energy investment, grid upgrades, water stewardship and local economic impact. The policy shifts from a market‑driven, speed‑focused model to one where compliance with national priorities determines access to...
Appliance Makers Place the Future of Home Robots on Display at AWE 2026
At the 2026 Appliance & Electronics World Expo in Shanghai, traditional appliance manufacturers pivoted toward embodied intelligence, unveiling a range of home‑service robots. Companies such as Dreame, Ecovacs, Haier, TCL, Hisense and Tesla demonstrated prototypes for chores, elder care, and...
Asia-Pacific Telcos Urged to Balance AI Innovation with Trust, Governance, and Local Readiness
Telecom Review Asia’s March 17 webinar highlighted how AI is reshaping APAC telcos, delivering measurable gains in network reliability and customer experience. Operators such as Telekom Malaysia and Globe Telecom showcased proactive monitoring, computer‑vision‑driven field work, and low‑code platforms that...
Differentially Private Lasso: An ISTA Framework with Finite-Iteration Guarantees
The paper introduces an Iterative Shrinkage‑Thresholding Algorithm (ISTA) framework for differentially private (DP) Lasso regression in high‑dimensional sparse settings. It delivers finite‑iteration, high‑probability ℓ₂ error bounds that separate a non‑private baseline, a privacy‑induced noise term, and a vanishing optimization residual....
Samsung TV Plus Powers the Next Era of FAST, Showcasing Live Events, Creator-Led Content, Exclusive Offerings, and Interactive Features Accelerating...
Samsung Ads unveiled a full‑funnel performance platform at IAB NewFronts 2026, pairing AI‑driven targeting with shoppable ad formats on Samsung TV Plus. The platform integrates Amazon’s Interactive Video Ad technology, enabling viewers to add Amazon items to cart directly from their...
AGI Arrived: Autonomous Execution, Not Human‑Level Intelligence
𝗔𝗚𝗜 𝗶𝘀 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗵𝗲𝗿𝗲. 𝗕𝘂𝘁 𝗻𝗼𝘁 𝗶𝗻 𝘁𝗵𝗲 𝘄𝗮𝘆 𝗺𝗼𝘀𝘁 𝗽𝗲𝗼𝗽𝗹𝗲 𝗮𝗿𝗲 𝗹𝗼𝗼𝗸𝗶𝗻𝗴 𝗳𝗼𝗿 𝗶𝘁. For years, AGI has been treated as a scientific milestone. Something that would arrive when AI reaches human-level intelligence across all domains. 𝗧𝗵𝗮𝘁 𝗳𝗿𝗮𝗺𝗶𝗻𝗴 𝗶𝘀 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝘄𝗿𝗼𝗻𝗴. When...
AI-Driven Partnership Targets Unmet Gynecological Diseases
We are very happy to join forces together with the wonderful ASKA to go after Novel Targets for Women's Health. Gynecological diseases have long posed challenges, including difficult diagnosis, limited treatment options, and a substantial disease burden, affecting millions of...
DreamPartGen Brings Part Aware Text To 3D
DreamPartGen is a new research model that generates part‑aware 3D objects directly from natural‑language prompts, addressing the long‑standing gap where text‑to‑3D tools produce monolithic meshes unsuitable for functional assemblies. By representing each component separately and modeling their spatial relationships with...
EmbedUR Expands Arm Support in Fusion Studio
embedUR Systems announced a major expansion of Arm ecosystem support in its ModelNova Fusion Studio desktop application at Embedded World 2026. The update introduces native ExecuTorch integration for Ethos‑U85 and U55 NPUs, seamless deployment to Alif Ensemble development kits, and...
Channel Partners Are Sleepwalking Into an AI Code Generation Trap
Managed Service Providers are being asked to deliver full AI‑enhanced development stacks, but rapid adoption of AI code generators introduces hidden security risks. Studies show nearly half of AI‑generated code snippets contain vulnerabilities and 67% of organizations fail to assess...
AI Could Run Entire Companies and Revolutionize Drug Discovery
Personally, I think that in two years (or less) the entire company can be run by AI including code generation and product deployment. Even humans can be controlled by AI (imagine yourself installed as a "skill" into the organizational AI)....
AI Agents Shifting to In‑workflow, Prompting Slack/Teams Dilemma
AI agents busting out of platform and going in-workflow is the next stage. Happening soon with LinkedIn Hiring Assistant. Q: Will we tolerate multiple agents in Slack / Teams or default to single one? https://t.co/acpq2Dot2k
AI Shifts to Scalable, Accountable Production in 2026
2026 will see AI move from pilots to accountable production at scale, delivering real value. https://t.co/iQq6h5Giyo
Copper, Not Models, Is the AI Bottleneck.
Everyone's watching model releases. The real AI race is in copper. Kandou AI raised $225M to reimagine how data moves between chips. When Anthropic and OpenAI both surge in the same week, they're not racing each other. They're revealing the shared bottleneck: infrastructure....
AI Now a Boardroom Accountability Imperative
#AI Is No Longer An #Innovation Bet—It Is A Boardroom Accountability Issue by Amit Basu @Forbes Learn more: https://t.co/GEwbzea4ml #MachineLearning #ArtificialIntelligence #ML https://t.co/TDzOjyWG74
Nvidia Launches Jetson Orin Nano Super AI Computer
#TechTuesday. Dec 17, 2024. Nvidia introduced the Jetson Orin Nano Super #AI #Computer. (Cheddar) #History https://t.co/Rdt2Hl1SMF