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
AI Music Is Reviving the Same Fights that Shaped the Player Piano
AI music platform Suno announced $300 million in annual recurring revenue and two million paying subscribers, signaling rapid market adoption. The service lets users generate songs from prompts, edit tracks, and even clone their own voices, attracting over 100 million free users. However, Suno faces lawsuits from artists and labels over unlicensed training data, echoing historic disputes from the player‑piano era. The debate centers on artistic integrity, labor displacement, and royalty structures in a rapidly automating music industry.
Anthropic Debuts Claude Design for Building Marketing Assets, Decks, and UIs
Anthropic Labs launched Claude Design, an AI‑powered platform that creates marketing assets, slide decks, and UI mockups using natural‑language prompts. Built on the upgraded Claude Opus 4.7 model, the tool lets users generate, edit, and collaborate on designs in real time,...
The Emerging Fifth Scaling Paradigm of AI
The AI landscape is currently defined by four overlapping scaling paradigms, each with its own data, alignment, compute, and orchestration requirements. Determining which paradigm a firm operates in guides where capital should be allocated—whether toward raw data, model alignment, inference...
Critical Permissions Should Never Be Auto‑Approved by LLMs
@boris_cherny Is there a way to get approve and auto accept edits back? I work on database systems my environment has very critical permissions that I just don't ever want an LLM to be able to use by itself regardless of...
Some Grocers Are Using AI to Cut Food Waste and Boost Profit Margins
Grocers are deploying AI platforms such as Flashfood to dynamically price perishable items, turning potential shrink into revenue. By targeting discounts to items near their best‑by dates, retailers like Kroger have cut waste by roughly 27% while driving extra traffic...
Ultrasense Systems Unveils Ultrasonic Tactile Platform for AI
UltraSense Systems introduced an ultrasound‑based tactile intelligence platform for Physical AI, featuring a protected sub‑surface sensing architecture. The system uses acoustic return signatures to detect contact, localize touch points, and infer force while keeping the core sensors shielded from wear....
Expert.ai and Microsoft Italy Together to Accelerate the Adoption of Agentic Architecture
Expert.ai and Microsoft Italy have partnered to place the EidenAI Suite on the Microsoft Azure Marketplace, enabling large Italian and global enterprises to deploy AI solutions faster and at scale. The collaboration aims to move organizations beyond pilot projects toward...
Hyperscale Data Announces Strategic Partnership with AGIBOT for AI Robotics
Hyperscale Data Inc., an AI‑focused data‑center operator anchored by Bitcoin mining, announced a strategic partnership between its wholly‑owned subsidiary Omnipresent Robotics and Singapore‑based AGIBOT PTE Ltd., a developer of intelligent robotics. The agreement calls for joint development and commercialization of...
Claude's New Prompt Length Limit Frustrates Users
I don't know what Claude did to Cowork's system prompt, but God, that's annoying. Some skills it could always do, but now it keeps on saying "are you sure, because it is too long to do?" However, you prompt it. I understand...
AI and Wearables Achieve 90% Accuracy in Predicting Athlete Injuries
AI-powered wearables are now able to predict injuries in athletes with roughly 90% accuracy, according to recent studies. The technology combines motion analysis, training load, sleep quality and recovery data, offering a proactive alternative to traditional reactive sports medicine.
LLM System Design Interview #27 - The Sequence Length Explosion Trap
In an Anthropic senior AI engineer interview, candidates are asked why a pure byte‑level tokenizer would cripple a Transformer’s compute budget. The answer lies not in linguistic semantics but in hardware efficiency: byte tokenization inflates token counts dramatically, turning a...
Dropbox Brings Its Files, Dash Search, and Reclaim Calendar Into ChatGPT with Three New Apps
Dropbox announced three new ChatGPT apps that embed its core file service, the enterprise search tool Dash, and the Reclaim AI calendar into OpenAI’s chat interface. The file app lets users preview, save, and share Dropbox documents without leaving the...
Why Human-in-the-Loop Quality and Simulation-Ready Data Assets Are Non-Negotiable for Safety-Critical AI
Safety‑critical AI systems—autonomous vehicles and robotics—depend on flawless data annotation, yet robotics lags behind due to heterogeneous sensors and no universal benchmarks. TELUS Digital, with a global community of over 1 million annotators, delivers more than 2 billion labels annually and offers...
UK Gov Announces First Startup to Get Funding From Its New Sovereign AI Unit
The UK government has announced the first startup to receive backing from its newly created Sovereign AI unit, a state‑run fund designed to operate like a venture‑capital firm. The initiative is intended to inject capital and strategic support into domestic...
Blackstone Invests $17 Million in TextQL to Automate Executive Data Queries
Blackstone Innovations Investments led a $17 million strategic round in TextQL, the AI‑driven analytics startup founded by Ding and Mark Hay. The deal targets the growing demand for instant, plain‑language answers to enterprise data questions, a market Blackstone sees as ripe...
Gaxos Labs Expands AI Platform with Music Generation, AI Chat and 3D Model Creation, Driving Increased Engagement Across Its AI...
Gaxos.ai announced that its Gaxos Labs platform now includes AI‑driven music generation, conversational AI chat, and 3‑D model creation. The new tools have already lifted user engagement across the company’s AI suite, signaling strong demand for multi‑modal creative utilities. By...
Google’s AI Mode Now Opens Web Pages Side-by-Side with Search Results and Lets Users Pull Context From Open Chrome Tabs
Google has added two AI Mode enhancements to Chrome desktop. The first introduces a side‑by‑side view that keeps the AI chat panel open while a clicked web page loads beside it, allowing users to ask follow‑up questions without losing context....
This Sam Altman-Backed $1.8 Billion Startup Bets AI Can Get Drugs Through Clinical Trials Faster
Formation Bio, backed by Sam Altman and top VCs, has raised $615 million at a $1.8 billion valuation to use AI for faster, cheaper clinical trials. The New York‑based firm plans to acquire a portfolio of about ten early‑stage drug candidates, many stalled...
Celonis and Oracle Deepen Ties, Deploy Process Intelligence on OCI for Fusion Cloud ERP
Celonis and Oracle announced an expanded partnership that makes the Celonis Process Intelligence Platform available on Oracle Cloud Infrastructure. The integration adds a dedicated AI‑agent context layer for Oracle Fusion Cloud ERP customers, promising real‑time process insights and autonomous decision‑making.
Cloudflare Unveils Code Mode MCP Server to Slash AI Agent Token Costs
Cloudflare announced a new Model Context Protocol (MCP) server built on its Code Mode platform, aimed at dramatically lowering the token consumption of large language model agents when they call Cloudflare's APIs. The move targets cost‑sensitive AI‑driven DevOps workflows and...
UPCEA Launches AI Hub to Serve as Trusted Source of Guidance for AI Adoption in Online and Professional Higher Education
The United Professional College and Education Association (UPCEA) has launched an AI Hub to serve as a centralized, trusted source of guidance for AI adoption in online and professional higher education. The platform offers applied research, primers, tools, webinars, labs,...
iDox.ai Launches Legal Service and Government Edition to Transform AI-Powered Redaction for Legal Workflows
iDox.ai unveiled its Legal Service and Government Edition, an AI‑powered redaction platform tailored for law firms, corporate legal departments, and FOIA professionals. The solution combines explainable redaction, case‑based file management, and compliance‑ready reporting to meet GDPR, HIPAA, and FOIA standards....
Top Bank Posts $4.6 B in M&A Advisory Fees as Stratgyk Launches AI Deal Platform
A leading investment bank reported $4.6 billion in M&A advisory fees for the prior year, underscoring the scale of advisory revenue in the sector. At the same time, Stratgyk introduced DealArchitect™, an AI‑first platform that promises end‑to‑end orchestration of mid‑market transactions...
LMA Survey Shows 93% of Lloyd’s Firms Adopt AI Frameworks in One Year
The Lloyd’s Market Association reports that 93% of firms across the Lloyd’s market now have or are building formal AI frameworks, up from half a year ago. The shift, driven by generative AI tools, emphasizes governance, data security and human...
Goodfire AI and the Billion Dollar Bet on Neural Network Interpretability: Why Reverse Engineering Foundation Models Matters for Health Tech...
Goodfire AI, a San Francisco public‑benefit AI lab, raised a $150 million Series B in February 2026, pushing its valuation to $1.25 billion. Its Ember platform gives developers programmatic access to neural‑network internals, claiming a 58% reduction in large‑language‑model hallucinations at roughly 90 × lower...
SK Networks Turns Walkerhill Into AI‑Driven Lifestyle Platform
South Korea’s SK Networks announced that Walkerhill Hotels & Resorts will be transformed into an AI‑driven lifestyle platform. The rollout adds a multilingual AI guide, AI lounge, and a fast‑track back‑office system, positioning the hotel as a flagship of SK’s...
Waymo Launches Autonomous Ride‑hailing Trials in London with Safety Drivers
Waymo has begun autonomous ride‑hailing trials on public roads in London, deploying a fleet of about 100 Jaguar I‑PACE vehicles with human safety drivers. The tests cover a 100‑square‑mile zone and are intended to validate the Waymo Driver’s performance before...
OpenAI Shifts to Business‑Focused AI Under CFO Sarah Friar, Targeting Profitability
OpenAI announced a strategic pivot toward business‑oriented AI products, driven by CFO Sarah Friar’s financial leadership. The company plans to launch a new high‑value model, double the share of revenue from corporate customers to 50% by year‑end, and trim consumer‑focused...
Firms Ramp Up AI Upskilling as Automation Threat Looms
Major corporations are fast‑tracking AI upskilling initiatives to shield workers from automation. Cisco mandates AI training for all employees, while PwC reports that its clients are seeing productivity gains after reskilling. The push coincides with bipartisan legislation offering tax credits...
Analytics Firm Bubblemaps Uncovers $300,000 Polymarket Profit From Biden Pardons via Blockchain Data
Paris‑based analytics company Bubblemaps identified a trader who earned roughly $316,000 by betting on four of President Biden's last‑minute pardons on Polymarket. Using AI‑driven blockchain forensics, the firm linked two accounts to a single Kraken wallet, prompting questions about insider...
AI Boom Collides with Skills Shortage in SA Workforce
South Africa faces a talent paradox as AI adoption accelerates while a deep skills shortage persists, according to Mercer’s 2026 Global Talent Trends report. The study, based on nearly 12,000 respondents, finds 52% cite talent scarcity and 50% point to...
Quantum Stocks Surge Over 50% as Nvidia Unveils Open‑Source AI Models
Nvidia's launch of open‑source AI models built to accelerate quantum workloads sent quantum‑computing stocks soaring. Xanadu Quantum Technologies led the charge, jumping 54% to a $7.9 billion market cap, while D‑Wave, IonQ and others posted double‑digit gains. The rally underscores growing...
Picsart Launches AI Playground to End Creator Subscription Fatigue
Picsart introduced AI Playground, a single‑prompt interface that bundles over 100 AI models from more than 24 providers. The tool replaces multiple subscriptions with a pay‑for‑use model, targeting the growing cost burden on independent creators.
Why AI-Powered Ordering Can Redefine the Restaurant’s Digital Experience
Restaurant digital ordering often feels like a static menu, leading to 60‑70% abandonment on complex orders. AI‑powered ordering turns the experience into a guided conversation, letting guests state intent such as budget or dietary needs and receiving a curated cart...
Palo Alto’s Helmut Reisinger Sees a Cyber Sea Change Ahead as AI Advances
Palo Alto Networks is deepening its AI‑centric security push by joining Anthropic’s Project Glasswing, an exclusive initiative that uses Claude Mythos to uncover zero‑day vulnerabilities. The company has accelerated its platformization strategy through recent acquisitions—Protect AI, Chronosphere and the soon‑to‑close...
AI‑Generated Val Kilmer Trailer Ignites Outrage Over Digital Resurrection
A trailer for the indie drama As Deep as the Grave debuted at CinemaCon featuring an AI‑rendered Val Kilmer, prompting a wave of criticism from actors and audiences. The filmmakers say they secured consent and compensation from Kilmer’s estate, but the...
Machine Learning Shows Knowledge Is Always Evolving
Machine learning reminds us that knowledge is provisional. Each new dataset has the power to reshape what we thought we knew.
LLM Operates Computer at Human Speed in New Demo
PURE SORCERY 🤯 check out this new codex demo... first time seeing an LLM use a computer as fast as a person. SO surreal. makes Claude’s computer use seem like it's in slow-mo. Video credit: @AriX https://t.co/Dx0596P0ej
AI’s Impact on Apparel Beyond Forecasting and Fit
The apparel sector is finally deploying AI in practical areas such as demand forecasting, size‑fit recommendation, pricing, search and internal workflow automation. These tools are helping brands cut excess inventory, reduce markdowns and lower costly returns by guiding customers to...
Agentic Swarms Boost Productivity—If Control Remains
Autonomous agentic swarms promise massive productivity gains if organizations can maintain control and compliance. https://t.co/XG4xwtjMEd
AI White‑Collar Automation Delayed Until Compute Surplus Hits
Uninformed gut feeling is that to the extent AI really can lead to large-scale automation of white collar work it won’t happen until there’s a glut of compute rather than the current environment. Double whammy on the way down in...
Can AI Help Us Finally Beat Cancer W/ Dr. Sanjay Juneja, TheOncDoc
In this episode of CareTalk, oncologist Dr. Sanjay Ganesha (TheOncDoc) discusses how advances in AI, immunotherapy, CAR‑T, and early‑detection blood tests are turning cancer from a death sentence into a manageable disease. He shares his personal journey from a vision‑loss...
Claude Will Require Government ID for User Verification
Claude users will soon need to use government ID to prove who they are and use the platform. https://t.co/LL7aqOItFS
OAuth Token Refresh Failure Halts OpenClaw; Need Re-Authentication
Guys how I fix my OpenClaw today? "⚠️ Agent failed before reply: OAuth token refresh failed for openai-codex: Failed to refresh OAuth token for openai-codex. Please try again or re-authenticate. Logs: openclaw logs --follow"
The System Behind Self-Driving: Waymo’s Dmitri Dolgov
In this episode, Waymo co‑CEO Dmitri Dolgov walks through the architecture that powers Waymo’s fully autonomous fleet, explaining how a multi‑sensor stack (LiDAR, radar, cameras) feeds a large foundation model that is refined into three specialized off‑board teachers—the driver, the...
Humanoid Robot Thrives in Environments Unsuitable for Humans
DR02: The Humanoid #Robot That Thrives Where Humans Can’t by @DeepRobotics_CN #Robotics #ArtificialIntelligence #Innovation #Technology https://t.co/hdoaTVx2Nh
Treat AI as Partner, Not Sole Decision-Maker
Organizations can get the best from AI by treating it less as a standalone decision-maker and more as a strategic partner: one that elevates insight and accelerates execution, while humans retain ownership of intent and outcome. https://t.co/omtdyAY01D
AI Is Accelerating Commerce, but Service-Led AI Is Where Brands Win
AI‑driven agentic commerce is surging, with AI‑referenced shopping traffic up 693% YoY in late 2025 and still 527% higher in early 2026. While social commerce still leads referral growth, brands are finding the biggest upside by embedding AI directly into...
Winning AI Firms Clean Data Before Scaling
"Garbage in, garbage out is as irrefutable as gravity." Scott Taylor. AI doesn't change physics. The companies winning with AI fixed the data first. Everyone else is paying NVIDIA to confirm their data is broken. https://t.co/o7kTniQ2Ev
Wowza Launches Video Intelligence Framework
Wowza announced the general availability of its Video Intelligence Framework, a solution that embeds AI directly into live‑streaming workflows. The framework extracts frames, runs inference, and outputs real‑time metadata, clips, alerts and event signals without separate pipelines. It works alongside...