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
Khan Academy Overhauls Classroom Platform with Integrated AI Tools
Khan Academy unveiled a redesigned classroom platform that embeds its Khanmigo AI tutor into teacher and student workflows, adding new dashboards, learning queues and gamified motivation features. The upgrade aims to boost the modest 15% student engagement rate with Khanmigo and prepare for a summer rollout to district partners.
Google's New Gemini Model Targets GPT‑5.5 Coding Power
Sources: Google plans to announce a new Gemini model next week. It’ll be in the same ballpark as GPT-5.5 in terms of performance. I’m told that inside Google, there’s real pressure to catch up on coding capabilities in particular. https://sources.news/p/google-about-to-release-new-gemini
Cisco to Cut ~4,000 Jobs in AI‑Focused Restructuring as Orders Surge
Cisco announced it will eliminate almost 4,000 jobs as part of an AI‑centric restructuring, citing $5.3 bn in hyperscaler AI infrastructure orders and a new full‑year revenue target of $62.8‑$63 bn. The move underscores a strategic shift toward silicon, optics and security...
Google Is About to Release a New Gemini Model
Google plans to announce a new Gemini model at its I/O developer conference on Tuesday, positioning the system alongside OpenAI's recently hinted GPT‑5.5. Industry insiders say Gemini will fall short of Anthropic's Mythos, which has become the new reference point...
NV Energy Cuts Residential Power, Redirects 75% to Data Centers, Sparking Solar‑Battery Surge
NV Energy announced it will stop supplying power to 49,000 Lake Tahoe residents after May 2027, diverting roughly three‑quarters of the area's electricity to new AI data centers. The move forces homeowners to seek independent solar and storage solutions as...
Cline SDK Solves Long-Workflow AI Agent Failures
Most AI agents break once workflows get long or multi-agent. @Cline just open sourced an SDK built to fix that. Features: • Plugins • MCP support • Subagents • Checkpoints • Scheduled jobs • Web fetch • Agent teams They also claim leading Terminal-Bench 2.0 results across frontier + open...
SAP Rolls Out 50+ AI Assistants and 200 Agents at Sapphire Conference
SAP introduced more than 50 AI assistants and roughly 200 AI agents at its Sapphire conference, positioning the suite as the core of an autonomous‑enterprise vision. The move aims to shift SAP applications from passive tools to proactive, data‑driven executors.
Snowflake Ecosystem Startups Pull $113 Billion in Venture Funding, Fueling AI‑Data Cloud Boom
Snowflake’s partner network has amassed more than $113 billion in venture capital across 1,300+ private firms since 2020, according to a new Crunchbase‑Snowflake report. The data shows a pivot from a broad funding surge in 2021 to a tighter, high‑value focus...
Sprout Social Deploys Trellis AI Platform‑Wide, Boosting Social Intelligence
Sprout Social announced that its proprietary Trellis AI engine will become a platform‑wide intelligence layer covering Publishing, Listening, Smart Inbox and Reporting, with a July launch for all of its roughly 30,000 brand customers. The upgrade promises conversational access to...
Perceptron AI Launches Mk1 Video Model Matching Frontier Labs at Fraction of Cost
Perceptron AI announced the Mk1 physical AI model, which delivers video‑understanding performance on par with top frontier models from Google, Anthropic, OpenAI and Qwen, while costing far less. The launch opens high‑accuracy visual AI to a broader range of industrial...
Microsoft’s MDASH Agentic Security System Tops CyberGym Benchmark with 88.45% Score
Microsoft unveiled its multi‑model agentic scanning harness, MDASH, which discovered 16 new Windows vulnerabilities, earned an 88.45% score on the public CyberGym benchmark and beat the nearest competitor by roughly five points. The system, built by the Autonomous Code Security...
Warwickshire Seeks Partner for £2.4m AI Drive
Warwickshire County Council has launched a two‑year procurement for an AI and automation programme valued at up to £2.4 million (about $3.1 million). The initial scope covers transcription, automated document generation, language translation and AI‑enabled chatbots delivered via a controlled intranet hub....
AGIBOT Unveils Genie Sim 3.0 for Embodied AI
AGIBOT Launches Genie Sim 3.0 to Power the Future of Embodied #AI by @AGIBOTofficial #ArtificialIntelligence #ML #MachineLearning #Tech https://t.co/UH61tnsQN7
Pentagon Deploys Anthropic’s Mythos AI in “National Security Moment” Despite Blacklisting the Firm as a Supply-Chain Risk
The Pentagon has begun deploying Anthropic’s unreleased Mythos AI model to hunt for and patch software vulnerabilities across U.S. government systems. Mythos, part of the controlled Project Glasswing effort, can uncover decades‑old flaws in browsers, infrastructure and other code. The move...
Swap Storefront Claims 2× Conversion Rates as Top Brands Deploy AI‑Powered Commerce
Swap launched its first agentic storefront, an AI‑powered sales channel that operates outside a brand’s main website and is already delivering conversion rates double the industry average. More than 20 retailers, including SIMKHAI and Retrofête, have signed on, signaling rapid...
GE Appliances Deploys 800 Gemini AI Agents to Transform Manufacturing and Supply Chain
GE Appliances announced the rollout of more than 800 Gemini AI agents across its manufacturing plants and supply chain, a move designed to sharpen decision‑making, improve product quality and lift operational efficiency. Unveiled at Google Cloud Next 2026 in Las...
Nebius Acquires Clarifai’s Compute Orchestration Team and Patents to Strengthen AI Inference
Nebius Group N.V. announced today it has acquired the core engineering team, patents and licensing rights to Clarifai Inc.'s compute orchestration technology. The move adds founder Matthew Zeiler as senior vice president of research and deepens Nebius' Token Factory inference...
Thomson Reuters Integrates Anthropic’s Claude Into CoCounsel Legal AI Platform
Thomson Reuters announced today that Anthropic’s Claude model is now directly integrated into its CoCounsel Legal AI assistant. The move ties Claude to a knowledge base of 1.9 billion Westlaw and Practical Law documents and 1.4 billion KeyCite signals, aiming to give...
Truveta and Knit Health Team Up to Fuse 130M‑Patient Dataset with Clinical Behavior AI
Truveta and Knit Health announced a collaboration that merges Truveta’s 130 million‑patient real‑world data set with Knit’s Large Clinical Behavior Model. The partnership aims to embed dynamic AI pathways into everyday clinical workflows, promising faster, more coordinated care.
Ambiq Micro Posts 59% YoY Sales Surge in Q1, Driven by Edge AI Chip Demand
Ambiq Micro reported Q1 2026 net sales of $25.1 million, up 59.3% year over year, as demand for its ultra‑low‑power Edge AI chips expands beyond wearables into medical, industrial and smart‑home markets. The company highlighted diversification of its customer base and...
BlackLine Unveils Agentic Finance Platform as Oracle Deploys AI Agents in Fusion Cloud
BlackLine rolled out its Agentic Financial Operations platform, enabling AI agents to execute accounting tasks autonomously, while Oracle simultaneously launched AI agents across its Fusion Cloud suite. The moves heighten competition over governance, control and platform lock‑in in enterprise finance.
Snowflake Adds Malaysia’s Sovereign LLM ILMU to AI Data Cloud
Snowflake has integrated Malaysia’s national language model, ILMU, into its AI Data Cloud platform, aligning the rollout with the service’s April launch on AWS’s Malaysia region. The move aims to give regulated enterprises secure, locally governed AI capabilities.
Stripe and Google Launch Agentic Commerce via Gemini AI
Stripe announced a partnership with Google to integrate its agentic checkout into the Gemini AI assistant, creating a direct purchase channel for consumers. The deal expands Stripe's AI platform roster to four major providers and adds stablecoin and Brazil's Pix...
Augustus Bank Gets OCC Conditional Charter, Pioneering AI‑Era Clearing
Augustus, formerly Ivy, secured a conditional full‑service national bank charter from the Office of the Comptroller of the Currency, positioning its new Augustus Bank N.A. as the world’s first AI‑era clearing bank. The move targets real‑time, 24/7 settlement of major...
Introducing AutoML and AutoRAG: Guided Experience for AI Engineers in Red Hat OpenShift AI
Red Hat introduced AutoML and AutoRAG as technology previews in OpenShift AI 3.4, automating key steps of the AI lifecycle. AutoML leverages Kubeflow Pipelines and the AutoGluon library to handle data processing, feature engineering, and model training, then presents a leaderboard...
How Building AI Agents Has Changed in 2026
In 2026 building AI agents has shifted from heavyweight framework‑glue code to SDK‑first approaches. Built‑in toolkits from Claude and OpenAI, a skills system, and longer context windows have eliminated the middle layer of infrastructure. Teams now focus on system prompts,...
Faster, Cheaper, Just as Smart: Improving the Economics of LLM Inference with Speculative Decoding
Large language models deliver unmatched capabilities but their inference costs and latency hinder enterprise adoption. Speculative decoding tackles this by pairing a small draft model with a full‑size verifier, allowing multiple tokens to be generated for the cost of one....
Scaling Enterprise AI: Delivering Models-as-a-Service with Red Hat OpenShift AI 3.4
Red Hat OpenShift AI 3.4 now ships Models‑as‑a‑Service (MaaS), giving enterprises a Kubernetes‑native way to govern and scale AI model serving. The built‑in AI gateway, powered by Connectivity Link, handles token quotas, rate limiting, self‑service API keys, and showback dashboards without extra tooling. Organizations...
The E-Commerce Industry in the AI Era: Has the Agentic Flood Hit?
Fastly’s new Threat Insights Report reveals that bots now account for nearly half of all web requests, with e‑commerce experiencing a 43% surge in wanted bot traffic. While human visits are up 5%, unwanted traffic fell 5%, indicating a shifting...
HackerOne CEO Kara Sprague on How AI Is Reshaping Cybersecurity
HackerOne CEO Kara Sprague warned that frontier AI models are expanding the attack surface and compressing the time between vulnerability discovery and exploitation to under a day. She noted a 76% year‑on‑year rise in vulnerability reports from Fortune 500 customers, while the...
AI IQ Is Here: A New Site Scores Frontier AI Models on the Human IQ Scale. The Results Are Already...
A new startup, AI IQ, has published interactive charts that assign human‑style IQ scores to more than 50 leading language models using twelve benchmarks across abstract, mathematical, programmatic and academic dimensions. OpenAI’s GPT‑5.5 leads the bell curve with an estimated...
ArgentEdge Launches to Tap Into Agentic AI Adoption Inflection Point
Australian AI consultancy ArgentEdge has launched to capitalize on a rapid surge in agentic AI adoption across Australia and New Zealand. The firm highlights Google’s Gemini climbing from 6% to 25% of global generative‑AI web traffic, while ChatGPT’s share fell, signaling...
Local SAP Partners See Opportunity in ‘Autonomous Enterprise’ Vision
At Sapphire 2026 SAP unveiled its "Autonomous Enterprise" vision, where AI agents move beyond assistance to autonomously executing business processes. Australian and New Zealand partners said the concept is exciting but hinges on practical implementation, governance, and data quality. SAP’s openness...
Meta Adds Incognito AI Chats to WhatsApp
Meta rolled out an Incognito Chat mode for its WhatsApp AI, promising that conversations are processed in a secure environment that Meta cannot access and are deleted by default. The feature creates temporary, private chats that disappear after use, aligning...
LRVHealth’s Keith Figlioli on the Evolving Landscape of AI Procurement
Keith Figlioli, managing partner at LRVHealth, outlined a three‑layer framework for AI procurement in healthcare—core enterprise platforms, foundation‑model platforms, and specialized use‑case solutions. He emphasized the growing influence of hyperscalers like Microsoft and Google, which are embedding AI tools such...
Claude Code Quietly Became the Most Dangerous Productivity Tool in AI #162b
This week three AI shifts reshape enterprise productivity. OpenAI’s GPT‑5.5 Instant slashes hallucinations by 52.5% in regulated domains, clearing a major compliance hurdle. Anthropic unveiled ten financial agents that automate pitchbooks, models, audits and KYC. Claude Code’s May 2026 update adds...
.jpg)
AI Is Delivering Business Value Today
The SAP Value of AI Report shows AI now powers 25% of tasks in global enterprises and is projected to reach 41% within two years. In Australia, firms are generating roughly $3 million USD ROI on an average $17.6 million USD AI...
AI Assumes Larger Role in US Elections
AI is increasingly shaping U.S. elections, with venture capital firm Andreessen Horowitz emerging as the largest donor to pro‑AI candidates in this year’s midterms. The firm’s pivot from crypto to political influence underscores a broader industry trend. Meanwhile, a Gallup...
Researchers Say AI Just Broke Every Benchmark for Autonomous Cyber Capability
Researchers at the UK AI Security Institute and Palo Alto Networks found that Anthropic’s Claude Mythos Preview and OpenAI’s GPT‑5.5 have outpaced the previously observed doubling trend in autonomous cybersecurity task performance. In AISI’s cyber‑range simulations, Claude Mythos completed both...
Nearly Every Enterprise Is Investing in AI, but only 5% Say Their Data Is Ready
Enterprises are rapidly adopting AI, with 97% reporting active initiatives, yet only 5% consider their data ready for scale. Early returns are emerging—67% see initial ROI and 24% report strong gains—but data quality, access, and integration remain major obstacles. Over...
Fivetran’s CPO: Closed Data Stacks Won’t Survive the Agent Era
Fivetran’s chief product officer Anjan Kundavaram warned that closed data stacks cannot sustain the query volume generated by AI agents, which can run ten to a hundred times more queries than traditional analytics. He argues that routing every request through...
Android Rolling Out AI-Powered ‘Contextual Suggestions’ that Learn From Your Habits
Google is rolling out an on‑device AI feature called Contextual suggestions on the Pixel 10 series running Android 16. The system analyzes app usage, location and routine activities to predict helpful actions, such as offering a workout playlist or prompting a TV...
Anthropic Finally Beat OpenAI in Business AI Adoption — but 3 Big Threats Could Erase Its Lead
Anthropic has overtaken OpenAI as the most widely paid AI model among U.S. businesses, reaching 34.4% adoption in April versus OpenAI's 32.3%, according to Ramp's AI Index. The shift reflects a four‑fold increase in Anthropic's corporate uptake over the past...
Lead Intelligent Signs Strategic Cooperation Agreement with Beijing Humanoid Robot Innovation Center
Lead Intelligent and the Beijing Humanoid Robot Innovation Center signed a strategic cooperation agreement to move humanoid robots from labs to mass deployment on industrial production lines, focusing on new‑energy and high‑end manufacturing. The partnership outlines four workstreams: scenario‑driven deployment,...
AI Sales Training: A Plain-English Guide
AI sales training leverages artificial intelligence to deliver continuous, data‑driven skill development for sales reps. Tools such as AI role‑play, automated call scoring, real‑time guidance, and personalized learning paths replace sporadic classroom sessions and manager‑heavy coaching. By simulating realistic prospect...
Big Tech Is Sacrificing Its Cashflows to Prop up the AI Boom
Big cloud‑computing giants Amazon, Google, Meta, Microsoft and Oracle are posting higher operating profits while their free cash‑flow after capital expenditures is sharply declining. The gap reflects massive spending on AI infrastructure, talent and research as firms race to dominate...
Teaching an AI to Remember
The author documents how he turned GitHub Copilot CLI’s stateless behavior into a persistent knowledge base by creating a global ~/.copilot/copilot‑instructions.md file. Over a month the file grew to 413 lines, encoding rules for Jira workflows, credential protection, project‑switching, output formatting, and...
AI Uses 12-Lead ECGs to Predict Long-Term Stroke Risk
Researchers at Brigham and Women’s Hospital and its Boston affiliates have unveiled ECG2Stroke, an AI model that reads standard 12‑lead electrocardiograms to forecast long‑term ischemic stroke risk. Trained on data from more than 100,000 patients, the convolutional neural network achieved...
Neurosurgeon Turns Credentialing Bottleneck Into AI Startup
For Dr. Marc Ayoub, a neurosurgeon born into a family of doctors, the path to entrepreneurship was a necessity born of frustration. As Ayoub describes it, while the medical profession was sold to him as a lucrative and stable career, the...
2026 Marks AI Efficiency’s Leap Into Boardrooms
The Infrastructure Cliff: Why 2026 Is The Year #AI Efficiency Hits The Boardroom by @bblanquera @Forbes Learn more: https://t.co/tRKiboXtv1 #ArtificialIntelligence #MachineLearning #ML https://t.co/n9EMuQAwBi