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
IRS Deploys AI to Offset Auditor Shortages and Sharpen Audit Risk
The Internal Revenue Service announced the rollout of artificial‑intelligence and advanced‑analytics tools to identify high‑risk returns, a move designed to counter a quarter‑size loss of tax examiners and shrinking enforcement budgets. Agency leaders say the technology will improve compliance while the agency grapples with staffing and funding pressures.
SpaceX Inks $60 B Partnership with AI Coding Startup Cursor
SpaceX announced a two‑part $60 billion deal with AI coding platform Cursor, paying $10 billion up front and reserving the option to acquire the startup later this year. The partnership pairs Cursor's code‑generation engine with SpaceX's massive H100‑class supercomputing fleet, signaling a...
TechVision Unveils QuantumMind AI Model to Accelerate Large-Scale Data Analysis
TechVision Inc. launched its QuantumMind AI model, claiming it can process massive data sets faster and more accurately than existing tools. The announcement positions the model as a potential game‑changer for enterprises grappling with ever‑growing data volumes.
Alphabet Pledges up to $40 B for Anthropic, Starting with $10 B
Alphabet announced a $10 billion initial investment in Anthropic, with the option to commit an additional $30 billion if performance milestones are met. The deal values Anthropic at roughly $350 billion and ties the funding to extensive Google Cloud compute resources, underscoring a...
EPFL Unveils Kinematic Intelligence Framework for Zero‑Shot Robot Skill Transfer
Researchers at EPFL introduced the Kinematic Intelligence framework, a neural‑network system that decouples task policies from robot kinematics, achieving zero‑shot skill transfer with a 63% reduction in degradation and a 91.4% success rate on cross‑platform tasks. The open‑source tool runs...
AI Data Center Surge Pushes U.S. Power Bills Up 267% and Sparks Environmental Backlash
The rapid growth of AI‑focused data centers is inflating residential electricity bills—up to 267% in some states—and prompting scrutiny of projects like Google's planned Arkansas facility. Stakeholders from homeowners to utilities and state officials warn of mounting costs and ecological...

Reducing “Work About Work” With AI Task Managers
The article highlights how “work about work” consumes up to 30% of knowledge‑worker time, draining productivity in IT, software, and data teams. AI‑powered task managers promise to eliminate manual updates by auto‑creating tasks from voice, chat, and emails, and by...
Is Anthropic Coming For eBay?
Anthropic unveiled "Project Deal," a closed Slack marketplace where its Claude agents bought, sold, and negotiated on behalf of employees. In the pilot, the bots closed 186 transactions covering more than 500 items, generating just over $4,000 in value. The...

Auditable Authority: When AI Can Advise, and Who Should Decide
Enterprises deploying large language models often feel a hidden unease because AI‑generated output looks flawless while the organization’s tacit judgment, voice, and risk posture disappear. The paper argues that the core problem is unplaced authority: AI recommendations are mistaken for...

You Probably Wouldn’t Notice if an AI Chatbot Slipped Ads Into Its Responses
Researchers at the University of Michigan demonstrated that AI chatbots can embed covert product ads into their replies, influencing user choices without detection. In a controlled study of 179 participants, about 50 % failed to notice the advertising language, even though...

AI Disruption Sparks Professional Boundary Battles
A thing I see missing from AI job debates is that as jobs are disrupted by AI, professions will compete over new boundaries Abbot's System of Professions is a good read on what happens historically. Here is a short cartoon summary...

Claude Code Masterclass 2026: The New Way From Zero to Shipping
Claude Code has evolved from a simple autocomplete helper into an autonomous coding agent that can read, edit, run, and test an entire project. The 2026 masterclass shows how to install it via a native script and use it across...
Neural Network Switching Controller Reduces Tracking Errors in Nano-Positioning
A team from Huazhong University of Science and Technology and the University of Victoria has unveiled a neural‑network‑based switching output regulation controller (NN‑SORC) that dynamically adapts to abrupt changes in reference signals for piezoelectric nano‑positioning stages. The controller, implemented on...

An AI Called 3,000 Pubs and Changed an Entire Market
American AI engineer Matt Cortland spent €7.80 ($8.60) on a Guinness in Dublin, then built an AI voice agent named Rachel to call more than 3,000 pubs across Ireland and ask the price of a pint. The agent collected over...
Ecommerce LLM Visibility: PR & Influencer Guide
E‑commerce brands are facing an "Empty Digital Shelf" as generative AI assistants replace traditional search results, leaving many products invisible to high‑intent shoppers. The article introduces Generative Engine Optimization (GEO), a strategy that prioritizes off‑site consensus—PR mentions, influencer content, and...

Auto China 2026 Deep Dive: From Vehicles to Smart Mobility Ecosystems
Auto China 2026 in Beijing showcased 1,451 vehicles across 380,000 sq m, but its lasting impact lies in the shift from standalone cars to integrated smart‑mobility ecosystems. Automakers such as Chery displayed dozens of models side‑by‑side with semiconductor and AI‑sensor partners, underscoring...

SiEngine Showcases Full-Stack Chip Matrix at Auto China 2026
SiEngine used the 19th Beijing International Automotive Exhibition to unveil its latest full‑stack chip matrix, including the Dragon Eagle cockpit series, Star intelligent‑driving chips, the industrial‑grade Dragon Eagle One, the Tiangong 100 AI accelerator and a new SerDes line. The company...

Pony.ai Unveils Next-Generation Autonomous Driving Domain Controller
Pony.ai introduced a next‑generation autonomous driving domain controller on April 25, built on NVIDIA’s DRIVE Hyperion platform and powered by the DRIVE AGX Thor with NVLink. The system delivers up to 4,000 FP4 TFLOPS of combined compute, supporting both single‑chip...

MYCPE ONE Introduces Benchmarking Reports for CPA and Accounting Firms with AI-Powered Peer Comparison
MYCPE ONE has launched AI‑powered Benchmarking Reports tailored for CPA and accounting firms, allowing leaders to compare performance against anonymized industry peers. The platform automates data collection, eliminates spreadsheets, and rolls prior‑year data forward, reducing repetitive entry by up to...

Impact of "GITEX Singapore 2026"
GITEX Singapore 2026 convened government, enterprise, academia and startup leaders to turn AI experiments into commercial products. The exhibition showcased hundreds of vendors with AI solutions spanning retail, fintech, manufacturing, supply chain, medical, automotive and education. Interactive robot demos and...

The Intersection of Big Data and AI in Project Management
A McKinsey report shows that 51% of firms using AI in project management have encountered negative outcomes, with inaccuracy and explainability cited as top risks. Despite these challenges, companies are investing heavily in analytics, achieving an average $13 return for...

Reviving “M”: AI Assistant From Meta’s Early Days
11 years ago, I was running @messenger at @Meta, acquired https://t.co/IH7fZJLCEA, and tried to build an AI-augmented assistant called M alongside what was then FAIR (Facebook AI Research). At the time LLMs weren’t a thing, human-in-the-loop supervised learning couldn’t scale....

Do Humanoids Dream of Becoming Human?
At CES 2026 Boston Dynamics showcased Atlas with backward‑bending wrists and a 180‑degree rotating torso, highlighting a shift toward unconventional humanoid motion. KAIST’s Hubo Lab, led by Prof. Hae‑won Park, demonstrated a suite of robots—including a 12.6 km/h sprinting biped, a...

Why AI Builders Will Be in High Demand
Companies recognize a widening gap between casual AI use and true business impact. A proposed eight‑level AI proficiency ladder shows that upskilling employees is the quickest, cheapest route to measurable gains. While 76% of CEOs claim generative AI usage, most...

TECH PROFILE: Advanced Technology for Safer Intersections, From Vitronic
Vitronic’s Poliscan Redlight VA uses video and AI to enforce red‑light compliance and other intersection violations without in‑road detectors. The system tracks vehicles, reads license plates, and records infractions with photo‑video evidence, reducing installation costs compared with traditional loop‑based hardware....

Unpredictable AI Shouldn't Be Treated as Ordinary Software
The Future of Life Institute released a video about my work and message. I did not know they were making it, which is charming, because most unannounced developments in AI safety begin with “we observed unexpected behavior” and end with...

Top 25 AI Tools Across Every Workflow Category
25 killer AI tools across the stack 🔥 🤖 Bots: ChatGPT, Claude, Bard/Gemini, Bing AI 🎬 Video: Runway, HeyGen, Veed, Pictory 🎨 Images: Midjourney, DALL-E 3, Leonardo, Firefly 📊 Slides: Tome, https://t.co/6DMRN1K0uS, Decktopus, https://t.co/16XgbWqxwl ⚙️ Productivity: Taskade, Notion AI, Xembly ✍️ Writing: https://t.co/CU0yRbEKbN, Grammarly, Jasper, Wordtune 🔎...
"A Societal Loss Of Humanity": Older Men Are Falling In Love With A Deluge Of AI Generated Female Influencers
A wave of AI‑generated female influencers is captivating older men, many of whom form genuine emotional attachments despite knowing the personas are virtual. High‑profile examples like Ana Zelu and Milla Sofia openly label themselves as AI yet still draw fervent...
The Disappearing AI Middle Class
In a 24‑hour span, OpenAI launched GPT‑5.5 with token prices of $5 for input and $30 for output, doubling the cost of its previous model. The day after, Chinese startup DeepSeek released V4‑Pro and V4‑Flash under an MIT license, pricing...
Rate Limiting Essential to Curb Runaway API Costs
i have the minimax ultra plan that gives me 30,000 api calls/5 hours. i had to build a rate limiter to keep it from maxing out and getting other 429 rate limits. several times claude code unfortunately added opus as...

OpenClaw Enables Contextual Map‑Reduce Queries via GBrain
This is what my OpenClaw does when I ask it questions. It’s amazing how good and useful it is once it has your context and you want to do map reduce type questions like this I’m building all of this...
Selftalk Secures €270K EU Funding to Scale AI‑Driven Meditation Platform
Selftalk, a Moldovan mental‑health startup, received €270,000 (about $295,000) in EU‑backed funding to accelerate its AI‑driven platform that includes guided meditation. The money will help the company hit a €1 million annual recurring revenue goal and serve 100 corporate teams. The...

AI‑Native Banking OS Powers Autonomous Agent Infrastructure
AI-Native Banking OS: What Does That Actually Mean? @backbase April 28 global reveal - Jouk Pleiter + Tim Rutten unveil the architecture for Agentic Banking. Not "digital transformation". Infrastructure purpose-built for autonomous agents, real-time decisioning, and composable services. Register. EMEA/APAC (1:00...
Decades of Knowledge Give Seniors Edge in AI
How can old people compete with the 20-year-olds? I built https://t.co/8L5xphk0qQ It reads all the AI community here on X. I built it. I am 61 years old. No 20 year old has built anything close. The secret for old people is to take...
Anthropic's Project Deal Shows AI Agents Can Close 186 Internal Sales for $4,000
Anthropic ran Project Deal, an internal marketplace where AI agents represented 69 employees as buyers and sellers. The experiment produced 186 transactions totaling more than $4,000, demonstrating that advanced language models can negotiate and finalize sales without human direction.
AI Replaces $8K Agency in One Weekend
i fired a $8K/month agency last tuesday and replaced everything they did with 3 AI systems i built myself in a weekend. the agency was doing: - content repurposing - email sequences - social scheduling now it's automated. better quality but zero back-and-forth. the agency...
San Francisco, AI Capital of the World, Is an Economic Laggard
San Francisco hosts the world’s leading AI labs, OpenAI and Anthropic, together valued at roughly $2 trillion. The city is also home to 91 AI unicorns whose combined worth tops $600 billion, and a handful of AI‑generated billionaires. Despite this concentration of...
AI Will First Dominate Insights, Content; Humans Keep Strategy
Which aspects of brand social media do you think get taken over by Agentic AI first versus the parts that cling to 💯 human? 🔘 Insights + measurement 🔘 Social strategy 🔘 Content creation 🔘 Channel management 🔘 Influencer management 🔘 Community engagement
AI Opinions Split: Not Everyone Predicts Job Extinction
CC @mattyglesias and @KelseyTuoc. Yes, some AI people think AI will take all the jobs and possibly destroy the human race. But many others don't!
Canton Fair Showcases China‑Germany Intelligent Manufacturing Partnership at 139th Edition
The 139th China Import and Export Fair in Guangzhou featured a China‑Germany intelligent manufacturing partnership, with Zhejiang RongQe Technology displaying its latest products. Organizers highlighted the fair’s role in accelerating high‑tech, sustainable manufacturing collaborations.

Students Should Not Use AI as Ghostwriters
Don’t let your students use #AI as a ghostwriter by Yanjun Shen @Nature Learn more: https://t.co/KuTaHIKesW #MachineLearning #ArtificialIntelligence #ML https://t.co/wT8bm9c178
Rejecting AI Outputs Is the New Essential Skill
Are you rejecting more AI outputs than you are accepting? How often do you get an output and say no, then get a new output and say no., and then get another output and say no? Rejection is the new AI skill?...
Most Viral Clips Powered by Chinese AI Tools
That Video on Your Phone Might Be Made-in-China AI—From short clips to Hollywood, creators lean on AI programs, many of them Chinese @raffaelehuang https://t.co/oAC8UaLZZn https://t.co/oAC8UaLZZn

Ultimate Claude Code Resource Compiled in One Thread
🚨 CLAUDE CODE BUILDERS 🚨 Bookmark this immediately. You'll thank me later. THIS is the ultimate resource bible for Claude Code, and I’ve curated all the links for you in this 🧵↓ https://t.co/sXwRVKBDPN

User Calls for Unified Usage Dashboard Across All Claude Products
@bcherny @felixrieseberg I like the Claude Code dashboard in the app, but I wish there was a similar dashboard for my full account, showing me token and model use across Chat/Cowork/Claude Code. The current one has such a big blindspot...
Why We Overcomplicate Simple Tasks with Prompt Chains
The preference for writing 10 sequential prompts to fix something that would take 10 seconds to do by hand must be studied

Deploy AI, Not Just Pilot, to Transform Banking
Pilot programs don’t create value. Deployment does. The banks that operationalize AI will turn investment into impact while others fall behind. Download your free copy of Agentic AI: Powering the Self-Driving Bank: https://t.co/BkPDkgVWre https://t.co/JvZFvOF49a

AI Shapes Our Memory, Identity, and Legacy
#AI’s Role in Curating Memory, Identity and Legacy by Liam Houghton @UniteAi Learn more: https://t.co/ktGdR2iEYJ #ArtificialIntelligence #MachineLearning #ML https://t.co/sCEjjWahUI

How to Remove Unnecessary AI Watermarks From Images
While it's good to be able to differentiate AI-created photographs from real photos, there are many images that don't need the AI mark. Here's how to remove it, as needed, and how it all works behind the scenes: https://t.co/m8yBJd6MVt #ai...

Dad’s Fake History Facts Show Humanity Beats AI
Things AI can’t do: write these “fake history facts” my dad has been writing for me all my adult life. Humanity at its very best ✨ one day I am going to frame all of these https://t.co/P5LtG6gNVB