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

Unlocking the Power of Data: How We Built Text-to-SQL with Agentic RAG at Rocket Mortgage
Rocket Mortgage unveiled Rocket Analytics, an agentic Retrieval‑Augmented Generation (RAG) platform that translates natural‑language questions into SQL, runs the queries, and returns instant, visualizable results. The system builds a metadata‑only knowledge base using Amazon Titan embeddings stored in a FAISS vector store, then applies hybrid semantic‑keyword retrieval and Cohere’s re‑ranker to select relevant tables before prompting a large language model. Caching of semantics and prompts reduces latency and cost, while an orchestrator agent framework paves the way for cross‑domain insights. The tool cuts analysis time from days to seconds, democratizing data access across the enterprise.
The Scoop: Grammarly Apologizes for AI Tool that Mimics Writers Amid Legal Dispute
Grammarly announced it is shutting down the “Expert Review” AI feature after backlash and a class‑action lawsuit. The tool generated editing suggestions styled after real journalists, including Julia Angwin, who claims her name and reputation were used without permission. CEO...
TUS Launches AI-Powered Digital Platform for Professionals and Employers
The Technological University of Shannon (TUS) has launched ReSHAPE, an AI‑powered digital platform aimed at upskilling professionals and aiding employers in Ireland’s midlands. Developed in partnership with Munster Technological University and the University of Limerick, the service provides skills audits,...

Higgsfield Unveils Content-Scoring Tool to Help Hollywood Navigate the AI Likeness Minefield
Higgsfield has introduced a content‑scoring feature for its Team Plan that automatically scans AI‑generated video and images for visual similarities to celebrities, copyrighted characters, brand logos, and even a director’s signature style. The tool assigns a similarity score, flags potential...
Free Sentiment Analysis Tutorial Reaches 39,000 Professionals Tomorrow
Tomorrow, 39,000 professionals will receive a new, free analytics tutorial on a topic executives love - sentiment analysis. You can think of sentiment analysis as a technique where you use a computer to analyze free-form text data and summarize the contents...
How Boll & Branch Spotted Scammers Attempting Generative AI Fraud
Boll & Branch, a $300 million luxury bedding retailer, narrowly avoided a return‑fraud scheme when a scammer used generative AI to submit a fabricated photo of a ripped sheet. The AI‑generated image contained a visible watermark, which the seven‑person CX team flagged...
AI and Telecom Merge: NVIDIA‑Nokia Redefine Connectivity
The energy at MWC Barcelona is always incredible, but this year felt different. We are witnessing a fundamental shift in how networks are built and what they can do. One announcement that stood out to me was the deepening partnership...

Cargill’s CTO on AI, Innovation and Where R&D Is Going
Cargill’s chief technology officer Florian Schattenmann explains how the company’s food‑R&D is evolving from a hands‑on laboratory function to a strategic, ecosystem‑driven operation. He highlights the rapid adoption of artificial intelligence across product ideation, consumer co‑creation and process optimization, while...

Designers Deploy Production Code, Engineers No Longer Needed
what's happening inside Anthropic right now is genuinely fascinating: DESIGNERS are now shipping production code with no engineer in the loop at all these are non-technical people with zero coding background. engineers on the team said designers are making "large state management changes you...
Model Labs Can Be Profitable When Properly Managed
From Clouded Judgement: In Defense of Model Lab Profitability https://lnkd.in/d7xVmPGs

Digna Reports 12-Month Enterprise Deployment Without Traditional Data Quality Rules
digna announced a twelve‑month enterprise data‑warehouse deployment that operated without any traditional, manually coded data‑quality rules, relying instead on AI‑driven anomaly detection. The platform replaced thousands of null checks, threshold controls, and custom SQL assertions with statistical learning models that...

Maharashtra to Use AI to Boost Freshwater Fisheries Output; ₹1,024-Crore Scheme Launched
Maharashtra has launched a ₹1,024‑crore Chief Minister Matsyasampada Yojana to modernise freshwater fisheries using artificial intelligence. The AI platform, built with tech firm Marvel, will map thousands of ponds, collect precise production data and guide desilting efforts. By addressing structural...
AI Replaces Infrastructure with Alignment and Judgment
AI is shifting power from infrastructure to alignment. Scale no longer requires layers of management or massive teams; it requires clarity, taste, and strong judgement. Leadership is about directing systems while staying close to the creative core. This shift unlocks a...
Introducing The AI Second Brain
After the release of ChatGPT, Tiago Forte halted his Building a Second Brain cohorts, recognizing that traditional PKM methods were becoming obsolete. In February 2026, OpenAI and Anthropic unveiled agent‑harness models that can read files, use tools, and act on...

‘Grief and the AI Split’
The article reflects on AI‑assisted coding as the latest step in programming evolution, noting that seasoned developers still feel the same satisfaction when code works. It highlights a growing divide: skilled "craftsperson" programmers see productivity soar while maintaining quality, whereas...
Audit Your AI's Brand Knowledge Before Launch
Training the AIs is all of our jobs. Can it answer your prospects questions about your brand? Have you fed it the proof points it needs to recommend you? Is there any key information missing from it's training data?? It's...

AI Robot Leverages 3D Maps to Locate Lost Items
An AI search robot uses 3D maps and internet knowledge to find lost items 🤖 https://t.co/8s2JBU3308 https://t.co/pQCM9RdV1a
Upcoming Opportunities in AI (March)
The AI Consortium is recruiting judges for three upcoming hackathons and reviewers for the Seattle Startup Summit, which expects 1,200‑1,500 attendees and 80‑100 exhibiting companies. Membership applicants undergo a selective process, and accepted members gain access to conference committee roles,...
Are You Ready for AI Agents in Core Workflows?
Question for your next meeting: "If our competitors deployed AI agents into our core workflows tomorrow, would we know? And would we have a response ready?" The agent economy moved from PowerPoint to production this week.
Verizon Exec Co-Founds Modular AI Data Center Startup
Looks like @hansvestberg of $VZ is a co-founder of Digi Power X, which has "a proprietary modular data center system engineered for the rapid deployment of high-density AI computing infrastructure." https://t.co/SGeBmPoV7t

Scammers Use AI-Generated Images of Lost Dogs to Target Pet Owners
Scammers are exploiting AI image generators to fabricate photos of missing dogs, claiming the animals are injured and need costly emergency surgery. Victims in Florida and California have been pressured to send money via platforms like Zelle after receiving realistic...
Peacock's AI Andy Cohen Signals NBCU's Entertainment Platform Shift
Peacock's AI Andy Cohen is part of NBCU's strategy to add new experiences tied to its IP to keep fans on its own platform: “Don’t think of it as a streaming platform — think of it as an entertainment platform,”...
Winners Decide Differently as AI Outpaces Org Design
“Everyone has the same data — winners make different decisions” In Episode 33, @UNLEASHgroup founder Marc Coleman joins us to talk about: 🎧 • why nobody has budget for theory • AI capability is moving faster than org design https://t.co/qWwcEE02lS

AI Hiring Is Now a Legal Risk. Are You up to Speed?
A proposed class‑action lawsuit in California accuses Eightfold AI of using hidden candidate scores that violate consumer‑protection and fair‑employment laws. The complaint alleges the algorithm relied on sensitive data—social media, location, device activity—without informing or allowing candidates to contest the...
Peacock Debuts AI Andy Cohen for Personalized Bravo Gossip
Peacock Is Launching an AI-Generated Andy Cohen That Dishes Personalized Bravo Hot Goss and Vertical Video Clips https://t.co/B1zYs5YfF6 via @variety

Clinical Decision Support an Advancing Frontier for Palliative AI
Artificial intelligence is moving beyond paperwork in palliative care to assist clinical decision making. A 2025 market survey shows 53% of providers use AI for documentation, while 19% employ predictive alerts. PalliCare’s integration of athenaOne’s Patient Summary feature aggregates ambient...

DCX Launches ECDU Liquid Cooling Systems up to 2.6 MW
DCX introduced its Enterprise Coolant Distribution Unit (ECDU) family, offering three liquid‑cooling models that span 600 kW to 2.6 MW of real‑world capacity in standard rack and in‑row footprints. The flagship 1380/2600 units deliver up to 3,333 lpm flow and support over 100...

Perplexity Takes the Lead in Abandoning MCP
Perplexity’s macOS application has become the first major AI client to integrate Anthropic’s Model Context Protocol (MCP) for system‑level tasks, signaling a shift toward standardized model communication. In contrast, the popular OpenClaw tool still lacks MCP support, underscoring divergent adoption...

Solidigm Strikes Out in New AI Computer Vision Direction
Solidigm, traditionally a NAND fabber and SSD supplier, has launched Luceta, a generative‑AI computer‑vision platform aimed at manufacturing inspection. The software lets customers create and train models in weeks, achieving over 90 percent precision and enabling rapid edge deployment. While Luceta...

Milestone Systems Redefines the Open Platform for an AI-Native Era
Milestone Systems unveiled the XProtect App Platform, a Linux‑based containerized layer that lets customers add AI, analytics and access‑control apps to their VMS without system restarts. The platform runs alongside existing XProtect installations, enabling zero‑downtime updates and rapid customization. Milestone...
Leveraging AI, Automation, and Data Analytics in a Hospice Setting
Chapters Health System, the nation’s largest nonprofit hospice network, is deploying AI, automation, and data analytics to overhaul its revenue cycle management. An AI-driven chart review tool now scans clinical documentation in real time, automatically routing missing signatures or data...

MeQ Appoints Brad Swingruber as CEO to Lead Next-Gen AI Workforce Intelligence and Wellness Platform
meQ, a cloud‑based AI workforce intelligence firm, announced Brad Swingruber as its new CEO. Swingruber brings experience from ADP, Xyleme and other HR tech leaders to steer meQ’s next‑generation platform that detects burnout and turnover signals. Backed by Bow River...
14% U.S. Workers Experience AI‑induced Brain Fry, Marketers
A study of U.S. workers finds 14% report "brain fry" from AI overuse. Marketing had the highest rate at 26%. https://t.co/VUlgEfNh1Z via @MattGSouthern, @sejournal

Biologists Wielding AI Will Outpace Those Who Don’t
1/ AI won't replace you. But a biologist using AI will. Especially in bioinformatics, where the questions never stop coming. https://t.co/ckfZWjnei3

AI Finance Converges: Upstart Seeks Bank Charter as Santander Deploys Agentic Payments Across Latin America
Upstart, the AI‑driven lender from Silicon Valley, announced it will apply for a U.S. bank charter, allowing its machine‑learning underwriting to operate within a regulated balance sheet and expand into new loan segments. At the same time, Santander launched an...
Virgin Voyages Expands to 1,500 AI Agents to Improve Guest Experience
Virgin Voyages has expanded its internal AI agent fleet from 50 to over 1,500 in under four months, leveraging Google Cloud’s Gemini Enterprise platform. The agents support functions such as marketing, sales, crew training, and guest services, including tools like...
Musk Admits xAI Must Be Rebuilt From Scratch
so SpaceX shareholders picked up the pieces and now Musk says xAI needs to be rebuilt
AI Will Spawn Human Errand Jobs for Autonomous Agents
one big coming surprise in AI MORE jobs created for people. but with a twist the increase in jobs will be for human taskrabbits running physical world errands (with no context of why they're doing task)... for AI agents working for other humans...

AI Chips Are Pushing Everything Else Off TSMC's Most Advanced Production Lines
AI accelerators are set to dominate TSMC's most advanced N3 production line, with 86% of capacity earmarked for AI chips by 2027. Utilization is projected to exceed 100% in the second half of 2026, highlighting a severe capacity shortfall. TSMC’s...

Sema4.ai Announces Semantic Layer Capabilities at the Gartner Data & Analytics Summit 2026
Sema4.ai announced the general availability of its AI‑powered Semantic Layer at the Gartner Data & Analytics Summit 2026. The platform lets business users query databases, spreadsheets and documents using plain English, eliminating the need for SQL expertise. It couples a...

How AI Is Transforming the Modern Firewall for State and Local Government
State and local governments are adopting a hybrid‑mesh firewall model that spreads enforcement across branch offices, cloud workloads, data centers and remote users. AI is being embedded directly into these firewalls to govern generative‑AI usage, enforce data‑loss‑prevention, and automate rule‑set...
AI Robotics Unicorn Sharpa and NVIDIA Bridge the Simulation Gap for Dexterous Robot Training
Sharpa, a unicorn AI robotics firm, announced a collaboration with NVIDIA to close the simulation gap for training dexterous robots. The joint effort produced Tacmap, a high‑fidelity yet fast simulation framework that uses a shared geometric representation and will be...

AI-Generated Code Passes Far More Automated Tests than Human
A METR study found that AI‑generated pull requests often pass the SWE‑bench automated grader but are rejected by human maintainers at a much higher rate. Between 50% and two‑thirds of AI patches that clear automated tests would not be merged,...

Why RegTech AI Is Reshaping Privacy Compliance Teams
Privacy compliance has become a top operational challenge in 2026 as regulators such as the EU AI Act, DORA and California’s Automated Decision‑Making Technology rules move into active enforcement. 4CRisk.ai argues that artificial intelligence can dramatically accelerate compliance, especially as...

Claude’s /Btw Lets You Ask Quick Questions Discreetly
1/ TIL about /btw in Claude Code and I wish I had this weeks ago. You can ask Claude a quick question while it's still working on something else. The answer doesn't pollute your conversation history. https://t.co/FmPNX0EX6Y

Corvex Announced the Launch of Secure Model Weights
Corvex announced early availability of Secure Model Weights, a patent‑pending solution that encrypts AI model weights end‑to‑end and decrypts them only inside NVIDIA Hopper/Blackwell GPUs operating in Confidential Computing mode. The architecture combines GPU TEEs, Intel TDX, remote attestation, and...

Reply at NVIDIA GTC: Digital Twins and Physical AI Driving the Next Stage of Industrial Value Creation
Reply showcased at NVIDIA GTC how digital twins and physical AI can unlock new industrial value. The company demonstrated a self‑learning edge AI platform that validates and retrains models on connected robots, leveraging NVIDIA Omniverse and Isaac Sim. Joint presentations...

Azure Startup Credits Don't Apply to Claude via Azure AI Foundry, Reader Finds – After $1,600 Charge
Start‑up companies using Microsoft Azure credits discovered that the credits do not cover Anthropic's Claude model accessed through Azure AI Foundry, leading to surprise charges of $1,600 and higher. A moderator’s original forum post incorrectly claimed credits applied, but was...
ThunderSoft Showcases Edge AI Innovations at Embedded World 2026, Highlighting the Rise of Intelligent Edge Systems
ThunderSoft and its joint venture Thundercomm unveiled three edge‑AI solutions at embedded world 2026, including the TurboX IRB10 development kit for autonomous robots, an AI‑powered industrial inspection system, and the compact TurboX EB8 edge computing device. Built on Qualcomm’s Dragonwing IQ10 platform, these offerings...

Chief AI Officer on Course-Correcting when AI Moves Too Fast
Rob T. Lee, chief AI officer at SANS Institute, warns that moving too quickly with AI can unsettle established processes and provoke resistance. He recounts a recent incident where an AI‑generated microsite was built in hours, shocking the traditional web‑design...