Today's AI Pulse
Anthropic shuts down Claude Fable 5 and Mythos 5 after U.S. export‑control warning
Anthropic disabled access to its Claude Fable 5 and Claude Mythos 5 models after the U.S. Commerce Department warned that foreign nationals could exploit the systems. The warning followed alerts from Amazon and other tech firms about a potential jailbreak that could facilitate cyber‑security attacks. Anthropic declined the administration’s request for a voluntary shutdown.
Also developing:
By the numbers: Meta scraps $2B Manus acquisition amid China divestiture order
You Can Make an App for That
The Verge reports that Anthropic’s Claude Code upgrade in late 2025 turned AI code generation into a reliable development partner, allowing users to build functional software for roughly $20 a month. The shift means non‑technical professionals can now create custom apps without learning traditional programming. This democratization challenges the long‑standing dominance of professional developers and enterprise‑grade tooling. The article frames the change as the start of a personal software revolution.
Welcome to the Age of the Underdog AI Model
Krea, a 37‑person design‑tools startup, announced its first generative AI model, Krea 2, as it repositions itself as a full‑fledged AI research lab. The company closed an $83 million Series B round, valuing it at $500 million, putting it in stark contrast with AI...
Building an AI CoE: Why You Need One and How to Make It Work
Artificial intelligence has become mainstream, with nearly nine in ten enterprises using it in at least one business function, but only a third have moved beyond pilot projects. Research shows AI delivers measurable innovation benefits, yet only 39% see a...
SKT, South Korea Ministry Target Defence AI
SK Telecom (SKT) has signed a partnership with South Korea’s Ministry of National Defence to create sovereign AI models for military administration and operations. The collaboration will leverage SKT’s A.X family of large language models, GPU‑as‑a‑Service infrastructure, and defence data...
MinIO Launches MemKV, Promising 95% Higher GPU Utilization for Enterprise AI
MinIO unveiled MemKV, a petabyte‑scale context memory store that it says can improve GPU utilization by more than 95% and halve the cost per token for AI inference. The product targets the recompute tax that slows large‑language‑model services in enterprise...
AI Will Write the Government Software. Who Writes the Spec?
Government technology leaders are confronting a new dilemma: spend $4 million on a traditional custom‑software contract or adopt AI that writes code from a plain‑language specification. The AI‑first approach promises faster delivery and lower costs, but it forces agencies to own...
Introducing Semaphore for AI Agents
In this episode, Pete introduces Semaphore for AI Agents, an open‑source CLI that lets AI coding assistants interact directly with Semaphore’s CI/CD platform via natural language. The tool provides commands for pipeline diagnostics, flaky‑test detection, CI insights, and workflow analysis,...
Microsoft Research: AI Can Generate Realistic Command-Line and Process Telemetry
Microsoft Research unveiled an AI system that converts attacker tactics from frameworks like MITRE ATT&CK into realistic command‑line and process telemetry. The approach uses prompt‑engineered generation, multi‑agent workflows, and reinforcement‑learning‑with‑verifiable‑rewards to synthesize logs that closely mimic real attack behavior. Evaluations show...
METiS TechBio's $270 M AI‑Drug Delivery IPO Sets HKEX Record
METiS TechBio listed on the Hong Kong Stock Exchange, raising HK$2.11 bn ($270 m) and becoming the world’s first publicly traded AI‑powered drug‑delivery company. The offering was oversubscribed more than 6,900 times, drawing $148 m from 18 cornerstone investors including BlackRock. The IPO...
Amazon Swaps Rufus for Alexa‑Powered Shopping Assistant, Rolling Out to All US Shoppers
Amazon announced that its Rufus AI shopping helper is being retired and replaced by Alexa for Shopping, a generative‑AI assistant available to every U.S. shopper within a week. The free service, which requires no Prime membership, promises personalized recommendations, price‑tracking...
AI-Generated Ads Fuel Controversy in Los Angeles Mayoral Race
AI‑produced campaign videos have racked up over 13 million views in the Los Angeles mayoral race, prompting fierce debate over their impact and fairness. Candidates Karen Bass, Nithya Raman and Spencer Pratt are scrambling to balance viral digital tactics with traditional...
Fanuc Shares Surge 16% After Partnering with Google on AI-Driven Robots
Fanuc Corp. saw its shares climb up to 16% to a record intraday level after unveiling a partnership with Alphabet’s Google to embed Gemini Enterprise AI into its robot arms. The move signals a major push to combine advanced cloud...
Atos and Backbase Team Up to Deploy AI-Native Banking Platforms Across 10 Regions
France's Atos and Dutch fintech firm Backbase have signed a Master Collaboration Agreement to deliver AI-native banking platforms in ten international markets. The deal combines Backbase's Banking OS with Atos' AI, sovereign‑cloud and integration expertise, aiming to help banks meet...
Physical AI Moves Closer to Factory Floors as Companies Test Humanoid Robots
British startup Humanoid will install up to 2,000 wheeled humanoid robots at Schaeffler’s global factories by 2032, with the first deployment slated for late 2026 to mid‑2027 at the German sites in Herzogenaurach and Schweinfurt. The initial phase focuses on...
Pax Silica: Inside the Trump Administration’s Tech Strategy with US Under Secretary of State for Economic Affairs Jacob Helberg
In this episode, Under Secretary of State for Economic Affairs Jacob Helberg outlines the Pax Silica initiative, a 14‑nation coalition aimed at securing the AI supply chain through forward‑deployed industrial zones, starting with a 4,000‑acre economic security zone in the...
Speed Up Your Video Redaction With Amped Replay
Amped Replay is a forensic video player designed for non‑technical police staff, offering drag‑and‑drop support for over 300 formats and offline AI‑assisted redaction of faces, bodies, vehicles and licence plates. The tool combines traditional pixel‑based tracking with machine‑learning detection, allowing...
HTN Now Panel Discuss AI in Primary Care, Covering Challenges, Adoption, Governance, Literacy and Successes
HTN Now hosted a webinar where a panel of experts from Tandem Health, a GP practice, and Nottingham & Nottinghamshire ICB discussed AI adoption in primary care. They highlighted the rapid rollout of Tandem’s medical AI assistant—now in roughly 97%...
The AI Banking Revolution: Why Trust Has Become More Important Than Ever
AI is reshaping online banking, with 70‑80% of banks deploying generative AI for fraud detection, customer service, and credit scoring. The technology could generate an extra $200‑$340 billion in annual revenue, but it also accelerates the disappearance of entry‑level roles. Over...
Transforming Benefits with AI: A New Era of Human-Centric Solutions
Artificial intelligence is reshaping employee benefits by automating core administrative tasks and delivering personalized health guidance. AI platforms can process enrollment, claims, and inquiries at scale, freeing HR teams to focus on strategy. The technology also uses predictive analytics to...
Most Organizations Now Use AI Agents for Sensitive Security Tasks
A new Semperis study finds that 93% of global enterprises already use or plan to deploy AI agents for sensitive security functions such as password resets and VPN access. At the same time, 92% have these agents installed on machines...
Protein Industries Canada Commits $5.1M to AI-Enabled Crop Insurance Framework Led by Agi3 and Aon
Protein Industries Canada has pledged roughly $1.0 million USD toward a two‑year, $1.33 million‑USD pea genomic‑selection initiative. The project will build a national database that integrates genetic, phenotypic, environmental and pedigree information for Canadian pulse breeders. Led by GIFS Ag Tech Enterprise with partners...
Mistral Pitches Security-Focused Model As Mythos Alternative
Mistral AI, a French startup, is promoting a security‑focused generative model as a European alternative to Anthropic’s Claude Mythos. The company says its AI can scan code for vulnerabilities in high‑risk sectors such as banking and the military, and it...
The Future of AI Healthcare Lies in a Solid Infrastructure Backbone
AI is reshaping healthcare with higher diagnostic accuracy and faster administrative tasks, but its success hinges on robust infrastructure. Edge AI reduces latency, especially in regions like Africa where network reliability is limited. Reliable power, tier‑IV data centers, and tailored...
New Ways to Agentically Build and Edit Dashboards
Sentry has launched a beta feature that lets users create and edit dashboards through an AI‑powered agent chat, replacing the older Insights pages with fully customizable, clonable dashboards. The new capability is free for organizations with AI features enabled and...
Agentic AI Is Redefining Human Connection and Leadership Now – HR Can’t Miss the Opportunity
A recent SAP SuccessFactors roundtable revealed that employees are increasingly treating agentic AI as a peer, with 22% preferring AI for celebration, 30% seeking AI for stress relief, and 40% turning to AI for emotional support. HR leaders, however, still...
UNICCON Group Showcases Indigenous AI and Autonomous Systems Drones Through Its Defence Subsidiary, Babasky Technologies.
UNICCON Group, via its defence subsidiary Babasky Technologies, has entered a collaboration with Nigeria's Defence Headquarters to develop and field indigenous AI‑driven autonomous drones and related security systems. The partnership builds on UNICCON’s prior AI work, including the Omeife humanoid...
The Creative Risk of Letting AI Do All the Work
MIT Sloan professor Sinan Aral warns that while AI can boost output, it often outperforms human‑AI teams, prompting managers to replace workers with automation. His experiments show that AI‑driven teams produce more ads faster, but the content becomes homogenized—a phenomenon...
White Circle’s AI Control Platform Closes $11M in Funding From Industry Giants After Founder Goes Viral for Exposing Safety Flaw...
White Circle, an AI monitoring and safety platform, closed an $11 million seed round backed by executives from OpenAI, Anthropic, Mistral, Hugging Face, Datadog and other AI heavyweights. The company, founded by Denis Shilov—who went viral in 2024 for bypassing safety...
Why Context Management Is the Missing Foundation Behind Every Successful AI Agent
Enterprises are racing to deploy autonomous AI agents, but nearly half of these projects stall before reaching production. The article identifies fragmented data sources—not model performance—as the primary cause of hallucinations, inconsistent answers, and audit failures. It argues that prompt...
AWS Introduces Claude Platform to Global Customers
Amazon Web Services has launched the general availability of Anthropic’s Claude Platform on its cloud, allowing customers to access Claude’s APIs, console and beta features through existing AWS accounts. The integration uses AWS Identity and Access Management for authentication, consolidates...
Introducing “The Architecture Speaks”
Arm has unveiled “The Architecture Speaks,” an experimental generative‑AI chatbot that answers technical questions about the Arm Architecture Reference Manual (ARM). The 17,000‑page manual is notoriously dense, and the tool provides concise answers plus direct HTML links to the relevant...
Flash Getting Stacked High-Bandwidth Version
Sandisk, in partnership with SK Hynix, has unveiled a high‑bandwidth flash (HBF) memory stack that mirrors the footprint of HBM but delivers far greater capacity—up to 3 TB per 16‑die stack—and 1.6 TB/s read bandwidth. The NAND‑based architecture targets AI inference workloads, keeping...
AI Hardware Enters Golden Age, Driving Infrastructure Boom
We may be entering the golden age of AI hardware. Hedge funds are heavily rotating into semiconductors, networking and infrastructure plays as the AI buildout expands far beyond software alone. The important shift is this. AI is no longer just an application...
Software Developers Shift to AI Code Reviewers
A Harness survey of 700 developers across five countries shows AI coding tools improve traditional productivity metrics but also create hidden work. While 89% report faster cycles and higher output, 81% now spend more time reviewing AI‑generated code, with 31%...
Care Homes Record Significant Reduction in Falls After Adopting AI Smart Lighting
Care homes in Lancashire and South Cumbria reported a 32% drop in resident falls after deploying Nobi’s AI‑enabled smart lighting across roughly 800 rooms. The technology also slashed average incident response times from over 11 minutes to under three minutes...
The Architecture Shift Behind Reliable Enterprise AI - with Ravi Marwaha of Arango
In this episode, Ravi Marwaha, COO and Chief Technology Product Officer at Arango, explains why AI pilots often succeed in isolated settings but fail to scale enterprise‑wide. He argues that the real bottleneck is fragmented, inconsistent context—not model performance—and that...
AI Agents Now Control Real Devices via Cloud
AI agents are starting to operate real devices like infrastructure. I’m streaming a headless Android phone from a data center to my Mac using scrcpy over Tailscale, then exposing controls to my AI claw through Peekaboo. Now the claw can navigate apps...
Nepal's Digital Marketplace, UAE's Agentic State, and the Singaporean Government Responds to Mythos
This week’s roundup highlights a wave of digital‑government reforms across Asia. Indonesia is amending its human‑rights law to strengthen the “right to be forgotten,” giving citizens a formal path to erase outdated personal data. Singapore’s Monetary Authority convened top bank...
AI Can Pinpoint You From a Reflection's Pixels
GeoSpy AI can locate a photo from a few pixels in a reflection. No metadata. No GPS tag. No obvious landmark. Just a tiny visual clue. And it runs in real time. This is the kind of AI demo that feels impressive for...
Baker McKenzie Announces Global Rollout of Legal AI Platform Legora
Baker McKenzie is launching its AI platform Legora globally across six practice groups. The phased rollout will initially support large‑scale document review, data extraction, drafting, redlining and checklist preparation. With more than 3,500 lawyers in 74 offices, the firm aims to...
The AI Chat Ad Frontier: What LLMs Change About Brand Safety And Control
OpenAI has launched ChatGPT ads, marking the first true advertising channel inside large language models. Unlike traditional digital media, the ad environment is generated in real time, with the surrounding conversational context evolving dynamically as users interact. This shift introduces...
Modern Retail+ Research: Marketers’ AI Use Rises, but Tech Skills Stall
Modern Retail’s annual survey shows AI adoption among marketers accelerating dramatically, rising from 44% in 2022 to 86% in 2025. While investment in AI tools is now near‑universal, the same respondents report that internal technical expertise has not kept pace....
Microsoft Build Signals AI Shifting to Production Impact
#MicrosoftBuild continues to be one of the most important events to watch for understanding where enterprise AI, developer platforms, cloud infrastructure, and intelligent applications are heading next. The presence of @satyanadella , alongside sessions focused on GitHub Copilot, AI deployment, and...
Inflation Erodes Debt as AI Reshapes US Economy
New free article: Inflating Away the Debt and the AI-fication of the US Economy $AIQ $SPX Link in reply below. 👇 https://t.co/gdRWS9brVh
Inside the AI Studio Where Mahabharat Is Engineered
Galleri5’s AI‑driven studio produced a new Mahabharat series for Jio Hotstar and Star Plus, reaching over 50 million viewers. The company claims a 30‑minute episode can be created in four‑to‑five days for $18‑$48k, far below the traditional Indian film budget of more than...
US Clears Chinese Firms for Nvidia Chip, Deliveries Stalled
Reuters: The U.S. has cleared around 10 Chinese firms to buy Nvidia’s second-most powerful AI chip, but no delivery has been made, three people familiar with the matter said, leaving a major deal in limbo as CEO Jensen Huang seeks...
ThinkyMachines Launches Ultra‑responsive, Humanlike AI Model
.@ThinkyMachines drops a new, highly responsive model designed for humanlike interactions in real time https://t.co/uYA1xfRzJz @SiliconANGLE @Mike_Wheatley “But Thinking Machines will need to be quick, for besides Amazon, it’s likely that OpenAI and Anthropic are also…” #nextGenApps #AIWars
Observe.AI Unveils Companion Agent to Boost Real‑Time Sales Call Performance
Observe.AI introduced Companion Agent, an AI‑powered multi‑agent interface that assists sales reps before, during and after customer calls. The tool promises faster call preparation, higher first‑call resolution and a measurable lift in pipeline speed, addressing the gap between automation and...
Anthropic Beats OpenAI in Verified Business Customers, Ramp Data Shows
Anthropic now leads OpenAI in the share of verified business customers, with 34.4% of Ramp‑tracked firms paying for its services compared with 32.3% for OpenAI. The shift, driven by rapid adoption in finance and tech, could reshape the AI tools...
Acres.com Unveils AI Agent That Cuts Land Research From Days to Seconds
Acres.com announced the launch of Acres Intelligence, the first AI agent built for land‑acquisition teams, capable of turning multi‑day zoning and feasibility research into seconds‑long answers. The tool leverages the company’s data on more than 150 million parcels to let developers,...