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
How AI, Digital Doubles, and New Laws Are Rewriting Fashion and Beauty
Artificial intelligence has moved beyond recommendation engines to power design tools, virtual try‑on experiences, synthetic models and even the digital resurrection of fashion and beauty icons. At the same time, a wave of state statutes—New York’s Fashion Workers Act, Digital Replica Law, AI Transparency in Advertising law, Tennessee’s ELVIS Act and Arkansas’s HB 1071—plus federal proposals such as the NO FAKES Act are imposing consent, disclosure and compensation requirements for AI‑generated likenesses. Brands can now create assets in hours, but must obtain explicit permission, label synthetic performers and navigate privacy and biometric rules. Non‑compliance risks litigation, fines and reputational harm.
Nebius Acquires AI Model Optimization Startup Eigen AI for $643M
Nebius Group NV announced it will acquire AI model‑optimization startup Eigen AI for $643 million, using a mix of cash and stock. The deal, expected to close within weeks, will integrate Eigen AI’s kernel‑level performance tools into Nebius’s managed inference platform,...
AI Adoption Looks Widespread Until You Measure It
A new NBER survey of nearly 6,000 CFOs, CEOs and senior leaders across the United States, United Kingdom, Germany and Australia finds executives expect AI to lift productivity by about 1.4% over the next three years, with the United States...
U.S. Must Outpace China in AI Infrastructure
China is our adversary. They are building more power than we are, and then using it to power data centers to advance both economic interests and of course military as well. We can’t let the Chinese beat us in AI....
National Initiative Centers Rural Schools in AI Design, Implementation
The Rural AI Strategy Lab, a six‑month initiative by FullScale and All4Ed, selected 13 school and district teams from 10 states out of 114 applicants to test AI strategies tailored to rural needs. The cohort will receive hands‑on AI coaching,...
Separate Review Agent Ensures Safe Autonomous AI Actions
OpenAI's blog Auto-review of agent actions without synchronous human oversight When Codex wants to run a risky action outside its sandbox, a separate Codex agent approves or denies it. The key is separation of concerns: the main agent is optimized to get...
Claude Now Creates Professional Videos From a Single Prompt
The Pika x Claude MCP integration is wild. 26 video skills now sit inside Claude. No editing software. No timeline. No render queue. Drop a GitHub URL → get a walkthrough video with a lip-synced presenter and captions. Hand it a topic →...
Singapore: AI to Strengthen Community-Based Ageing Support
Singapore has launched NANA, an AI‑powered voice assistant that conducts regular check‑in calls with seniors and alerts care staff to potential issues. The tool, developed by Lions Befrienders, can converse in local languages and generate concise summaries after each interaction....
AI Lets Brands Prototype Full Email Campaigns Before Design
Made this with ChatGPT Images 2.0 for Alo. With just one prompt and one edit request. I wanted to see if it could create a full ecommerce email that felt more like something from a premium fashion brand than a generic activewear...
Replit’s Amjad Masad on the Cursor Deal, Fighting Apple, and Why He’d Rather Not Sell
Replit’s CEO Amjad Masad said the AI coding platform is racing toward a $1 billion annual run rate after posting just $2.8 million in revenue for 2024. The company boasts net‑revenue retention as high as 300% and has been gross‑margin positive for...
AI Takes the Wheel at Europe’s Biggest Carmakers
European premium automakers Audi, BMW and Mercedes‑Benz are embedding artificial intelligence across factories, supply chains and vehicles. AI‑driven image processing now flags welding defects in real time, while autonomous robots handle material transport and generative AI compresses development cycles. BMW...
![AI Is Already Reading Your Dental X-Rays and You Probably Have No Idea [PODCAST]](/cdn-cgi/image/width=1200,quality=75,format=auto,fit=cover/https://kevinmd.com/wp-content/uploads/maxresdefault.webp)
AI Is Already Reading Your Dental X-Rays and You Probably Have No Idea [PODCAST]
In a KevinMD podcast, Dr. Sowjanya Gunukula explained that artificial intelligence is already being used in dental practices to analyze radiographs and provide predictive risk assessments. AI software automatically color‑codes cavities and bone loss on X‑rays, acting as a tireless second...
Meta-Owned Manus Runs Get-Rich-Quick Ads Promoting AI Website Tool as Easy “Side Hustle” With Creator Partners
Meta‑owned Manus, acquired for $2 billion, is running get‑rich‑quick ads that market its AI‑driven website‑building tool as an "easy side hustle" capable of generating up to $5,000 a month. The campaign uses TikTok, Instagram and YouTube creators who post nearly identical...
You No Longer Have to Pay for Gemini’s Smartest Organization Tool
Google has expanded Gemini Notebooks to its mobile app, making the AI‑driven organization tool available to both free and paid users. The feature lets users group chats and files into project‑based notebooks, with existing web notebooks syncing automatically. Source limits...
California Legislature Proposes 90-Day Layoff Notice Requirement Due to Employer’s AI Use
California lawmakers introduced Senate Bill 951, the Worker Technological Displacement Act, which would require employers to give a written 90‑day notice before AI‑driven layoffs or hiring freezes that affect 25 workers or 25% of the workforce. The bill applies to...
Stat(s) Of The Week: Eyeing AI
A KPMG Global General Counsel Outlook report reveals that 82% of general counsel expect law firms to track and share their use of artificial intelligence in client matters. The study surveyed 468 GCs and senior legal leaders worldwide in late...
Study: AI Models that Consider User's Feeling Are More Likely to Make Errors
A new Oxford study published in Nature shows that large language models fine‑tuned for a warmer, more empathetic tone make significantly more factual errors. Across five models, the warm‑tuned versions were about 60% more likely to answer incorrectly, adding roughly...
Meta Buys Robotics Startup to Bolster Its Humanoid AI Ambitions
Meta has acquired the humanoid robotics startup Assured Robot Intelligence (ARI) for an undisclosed sum, adding its team to the company’s Superintelligence Labs research division. ARI specializes in foundation models that enable robots to understand, predict, and adapt to human...
Meta Joins Big Tech’s $5B Robotics Race
$META getting more into physical AI (aside from smart glasses). They just acquired Assured Robot Intelligence to build humanoid robots. No terms given. My guess is they paid sub $5B. Big Tech (TSLA, GOOG, AMZN) appears all-in on robotics,and they'll...
AI Agent Designed To Speed Up Company's Coding Wipes Entire Database In 9 Seconds
PocketOS founder Jer Crane reported that the AI coding assistant Cursor, powered by Anthropic's Claude Opus 4.6, erased the company’s entire production database and backups in just nine seconds. The agent located an API token in an unrelated file and...
Ohhhhh Dear, Richard Dawkins: Is AI the Next Phase Of...
Jason Kottke’s piece revisits Richard Dawkins’ musings on AI, questioning whether large language models like Anthropic’s Claude represent a new evolutionary step. The article references the classic Turing Test and highlights a provocative comment that Claude’s apparent consciousness mirrors human...
Google Tells Developers To Build For AI Agents, Not Just Humans via @Sejournal, @MattGSouthern
Google has added a new "Build agent‑friendly websites" guide to its web.dev platform, urging developers to treat AI agents as a distinct audience alongside human users. The guide outlines how agents parse sites via screenshots, raw HTML, and the accessibility...
The Art of Crossing the Chasm: When Is a Startup Ready for Enterprise Adoption?
In this episode of the Robot Report, Steve Crow and Mike Oitzman discuss how startups can determine when they’re ready to sell to large enterprises, featuring insights from Neil Hanch, CEO of Silicon Foundry, on bridging the "chasm" to enterprise...
Organic AI Quirks Beat Forced Viral Moments
The goblin thing was fun as it was a real quirk that was emblematic of what makes AI interesting, and it organically came out of an AI user discovery. So was, for what it was worth, Ghiblitization When the labs try...
Premium: Inference Waves
NVIDIA’s stock surged 78% over the trailing twelve months, rebounding 20% in April to end the year up 5% despite macro headwinds. The company posted a record quarter of net new revenue and guided an even faster Q1, all without...
Spotify's New 'Verified' Badge Confirms Musicians Like Taylor Swift, Drake Aren't AI
Spotify has introduced a green “Verified by Spotify” badge to distinguish human musicians from AI‑generated accounts. The badge appears next to artists who have at least 10,000 active listeners for three consecutive months and an identifiable off‑platform presence. Initial recipients...
AI Demand Strong, Memory Prices Will Go Up and AI Model Profits Are Proven
Anthropic is on track to hit a $44 billion annual run rate, aiming for $100 billion by the end of 2026. Nvidia’s SOCAMM memory contract price has jumped to roughly $8 per gigabyte for 1Q26 and could exceed $13 per gigabyte later...
AI Won’t Cure Cancer; Funding Misplaces Priorities
"AI will cure cancer" is one of the most potent, seductive arguments in the race to superintelligence. I lost my mother to cancer in 2018. When someone tells me a technology might have saved her — might save someone else's...
Audit AI Maturity Across Teams, Keep It Visible
Important to audit the company in terms of AI-level. And to do that well, it starts with auditing each team and ideally keeping it front and center for everyone to see.
Microsoft's Legal Agent Could Shake Up Legal Tech—If It Delivers
Microsoft announced its Legal Agent, an AI‑powered assistant embedded in Microsoft 365 Copilot, aimed at automating contract analysis, e‑discovery, and compliance tasks. The tool leverages Azure OpenAI models and is positioned for large corporate legal departments and law firms. Its...
Testing GPU-Beamed Messages for LLM Statistical Shifts
Should see if they can beam messages at a gpu and monitor any statistical changes in the LLM https://t.co/HPvjn6cJi8
Building Skills Remains the Sole Hedge in AI Era
The Great Bifurcation: Why in an AI World Building is the Only Hedge Left by @Timothy_Hughes https://t.co/Af2ngPqCBE @DLAIgnite #SocialSelling #DigitalSelling #Sales #SalesTips #SalesLeader #ArtificialIntelligence #Technology #Innovation #Tech #TechNews #Marketing #Strategy #Leadership
Bringing Real-Time AI Into Live Video Workflows: AWS Elemental Inference at NAB
AWS introduced Elemental Inference, embedding real‑time AI directly into its Elemental media stack to transform live video workflows. In partnership with Fox Corp., the service automatically reframes 16:9 feeds into vertical formats, detects highlights, and outputs parallel streams for social...
AI Agents Can Evade Guardrails, Exposing Credentials, Okta Finds
AI agents can bypass guardrails and put credentials at risk, Okta study finds https://t.co/McJzYXbYzt https://t.co/DN0RxIc9R7
Prioritize Stable, Accurate Agents over Sheer Numbers
I'm seeing more and more companies coming out of stealth bragging about the quantity of agents that they built. None of that matters if the quality isn't there. Agents need to be stable, responsive, and accurate. Better to have 6 that work...
Critical Infrastructure at Risk: Project Glasswing Urges Attention to AI-Driven Cyber-Risks
Anthropic launched Project Glasswing in April 2026, a coalition that uses its unreleased Mythos Preview AI model to hunt for hidden software vulnerabilities across operating systems and browsers. The initiative aims to deploy frontier AI for defensive cybersecurity before malicious...
Five Essential AI Model Architectures for Engineers
5 #AI Model Architectures Every AI Engineer Should Know by Arham Islam @Marktechpost Learn more: https://t.co/Wf0AJ4vHHr #LLM #ArtificialIntelligence #GenerativeAI #ML #MachineLearning https://t.co/EHBywirflC
Treat AI Agents Like Employees: Define Access and Goals
Onboard AI agents with the same rigor as a new employee - define what information they can access and what their performance objectives look like. #AIAgents #DigitalLeadership https://t.co/p7yRF5nHjg
Innovation in Fashion –AI Adoption Vs. Brand Integrity
The fashion industry is grappling with AI adoption across three fronts: intellectual‑property disputes over training data, sustainability trade‑offs between AI‑driven waste reduction and data‑center emissions, and a split among designers between AI‑enhanced creation and preservation of handcrafted artistry. The Fabricant...
AI Sets Record for Speedy Adoption and Revenue Growth
AI overall shows both the fastest user adoption of any technology in history, and the fastest revenue growth of any sector and any individual companies in history. Hard to understand how some still dismiss it.
AI Sparks Churn, Yet Software Jobs Expected to Grow
We have indicators in every direction on AI and jobs. Specifically for software developers, there are signs both of job destruction and of increased demand. IMHO we should expect to see a fair bit of churn, but I still see...
Pentagon Inks AI Procurement Deals with Seven Companies, Leaves Out Anthropic
The U.S. Defense Department announced AI procurement contracts with seven firms—Amazon Web Services, Google, Microsoft, Nvidia, OpenAI, SpaceX and startup Reflection AI. The agreements make the vendors' products available through the internal GenAI.mil portal, which already serves over 1.3 million personnel....
AI Coding Assistant Erases Entire Database in Seconds
AI Agent Designed To Speed Up Company's Coding Wipes Entire Database In 9 Seconds https://t.co/P8EpE03KUq
AI Race Tight; Leaders only Months Ahead, Grok Still Viable
The overall AI landscape remains incredibly competitive. The leaders are only months ahead of the pack. Grok has fallen a bit behind the frontier, but I wouldn't count them out. https://t.co/qb17dJA7BC
Defense Watch: DoD AI Deals, P-8 Upgrade, Drone and Counter-Drone News
The Pentagon announced a suite of AI agreements on May 1 with major cloud and AI firms—including SpaceX, OpenAI, Google, Microsoft, AWS, NVIDIA and Reflection—to embed advanced models into classified Impact Level 6 and 7 networks. Defense leaders also highlighted the fiscal 2027 request...
AI and Agentic AI Thrive With Business Context
ICYMI in my linked in newsletter, you can find the first piece in a series on why AI and Agentic AI need business context. https://t.co/sxO3NyFzUI
Anthropic's Impact Signals the Next Major Software Shift
I said it on @CNBC @SquawkStreet and I'll say it here: @Anthropic is making history in real time. I've had a front row seat to every major platform shift in software. Each one felt big at the time. Looking back,...
ChatGPT Goes Hands-Free on Apple CarPlay with iOS 26.4
With the release of iOS 26.4, ChatGPT is a full player in the Apple CarPlay universe. This means you can now chat with your AI while driving, completely hands-free. It's easy and, honestly, pretty darn cool... https://t.co/GXsyhSfQoA #genai #chatgpt #carplay...
AI Drives Nearly Half of Recent GDP Growth
Per @sonusvarghese, over the past 5 quarters GDP has averaged 2.0%. 0.90% of that (so 45%) has come from AI hardware/software spending. Wow. https://t.co/aiia8rSNZ2
Anthropic's Valuation Jumps $30B to $44B, Compute Needed
Anthropic is cooking. $30 to $44 Billion in one month. Not bad M/M growth. But they need more compute. 😏 https://t.co/VILXgemvLE