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
Cognex Unveils In-Sight 3900, AI Vision System for Edge Inspection
Cognex Corp. introduced the In‑Sight 3900, an embedded AI vision system built on Qualcomm Dragonwing chips that delivers high‑speed, high‑accuracy inspection at the edge without a separate computer. The launch aims to let manufacturers run demanding quality checks without sacrificing line throughput.
RedCloud Hires Ex-Microsoft Exec Vikram Sharma as CRO to Drive AI Joint‑venture Expansion
RedCloud Holdings plc appointed Vikram Sharma, a former Microsoft senior leader, as Chief Revenue Officer for Infrastructure. Sharma will lead AI‑focused joint‑venture expansion in Southeast Asia and West Africa, building on a $30 million licensing agreement in Saudi Arabia and a...
LinkSquares Unveils First All‑Agentic CLM Platform to Speed Real‑Estate Deals
LinkSquares announced the launch of an all‑agentic, AI‑native contract lifecycle management platform that drafts, redlines and executes real‑estate contracts in minutes. Early‑access users report cutting workflow time from hours to two minutes, promising faster deal cycles and tighter legal oversight.
Viz.ai Secures ISO/IEC 42001 Certification, Setting Benchmark for Agentic AI in Healthcare
Viz.ai announced it has achieved ISO/IEC 42001 certification, the first international standard for AI Management Systems, positioning the company as a pioneer in responsible agentic AI. The audit covers the full AI lifecycle and reinforces confidence for the 2,000 hospitals...
Google Extends NotebookLM to Higher‑Ed Students in Classroom
Google has opened its generative‑AI NotebookLM tool inside Google Classroom for higher‑education students aged 18 and over. The rollout, which began on April 27, lets learners build personal AI‑driven notebooks that generate audio, video, flashcards and visual diagrams from course...
OpenAI’s Symphony Codex Orchestrator Boosts Pull‑Request Throughput Sixfold
OpenAI released Symphony, an open‑source specification that lets its Codex agents pull tickets from Linear and run autonomously until merged. Internal teams reported a sixfold jump in merged pull requests within three weeks, highlighting a new way to eliminate human...
Johnson Controls Unveils Air‑Cooled Chiller Guide for 1‑GW AI Factories
Johnson Controls launched its second AI Factory Reference Design Guide, focusing on air‑cooled chillers that can support up to a 1‑GW AI data center. The blueprint offers sizing for 220 MW compute clusters and promises industry‑leading energy and water efficiency, addressing...
HCLTech Names Avishek Chattopadhyay Global CIO to Drive AI‑Led Transformation
HCLTech announced that Avishek Chattopadhyay will serve as its Global Chief Information Officer, overseeing an AI‑focused overhaul of the firm’s technology landscape. The move positions a two‑decade veteran to align tech and business priorities across a $14.7 billion revenue operation.
Arista Networks Beats Q1 Forecast, Projects $11.5B Revenue on AI Surge
Arista Networks reported first‑quarter 2026 revenue of $2.71 billion, a 35.1% year‑over‑year increase that topped its $2.6 billion guidance. The company now expects full‑year revenue of $11.5 billion, driven by AI‑related sales, while warning that wafer and chip shortages could pressure margins.
Can Companies Insure Against AI’s Growing Risks?
Artificial intelligence is generating a surge of legal and financial liabilities, from fabricated legal citations and biased hiring tools to deep‑fake privacy breaches. High‑profile cases—including Anthropic’s $1.5 billion copyright settlement and a $240 million verdict against Tesla’s autopilot—show that exposure can be...
Pentagon Leaders Love Agentic AI. But It’s Giving Cyber Criminals Nation-State-Like Powers
Pentagon officials report that the GenAI.mil platform, now equipped with Google Gemini and under evaluation with Anthropic’s Mythos, is compressing multi‑week defense tasks into a few hours. While the agency touts the productivity boost, Mythos remains on a national‑security blacklist,...
Space Data Centers: Inevitable Future Backed by Tech Titans
We have been discussing the future of space based data centers at OODA for years. Some very technically savvy, space systems engineering friends of mine are convinced data centers in space will never work. They present long lists of engineering...
Visa Unveils AI‑Powered Dispute Suite to Slash Fraud Losses for Banks and Merchants
Visa introduced a six‑product AI‑driven dispute resolution suite designed to reduce fraud losses and operational costs for banks, merchants and payment service providers. The suite leverages predictive analytics and generative AI, targeting the 106 million disputes Visa processed in 2025, a...
Moritz Raises $9 Million Seed Round Backed by Y Combinator and 20 Unicorn Founders
Moritz, an AI‑driven legal‑tech startup founded by former OpenAI counsel Pamir Ehsas and ML engineer Stefan Mandaric, closed a $9 million seed round in four days. The round was led by Y Combinator and included Urban Innovation Fund, 20VC, Inception Fund and...
25‑nm Ferroelectric Memory Chip Beats Power‑Loss Limits, Paving Way for Cooler AI Devices
Scientists at the Institute of Science Tokyo have demonstrated a 25‑nanometer ferroelectric tunnel junction memory cell that consumes less power as it gets smaller. The breakthrough overturns a long‑standing belief that miniaturization inevitably raises energy cost, opening a path to...
Zalando Attributes 24% Q1 Revenue Jump to AI Boost From ABOUT YOU Deal
Zalando posted a 23.8% rise in first‑quarter revenue to €3bn ($3.2bn), saying AI‑driven personalization and logistics improvements from its ABOUT YOU acquisition powered the growth. The German‑based fashion marketplace also lifted GMV 21.7% and adjusted EBIT 38.7% year‑on‑year.
Codex's GPU Slowdown Makes Tasks Take Ages
I want to love Codex, but holy mother of god did they cover the GPUs in molasses? I swear to god it takes 25 min to do something that Claude Code could do in 5 minutes (and yes, I'm using...
ServiceNow Expands AI Control Tower with Governance Tools
.@ServiceNow Knowledge 2026: AI Control Tower, Action Fabric, Autonomous Workforce and more https://t.co/Y5Tk3LEjiR ServiceNow expanded its AI Control Tower with new integrations, observability and governance tools and features to measure the value of AI agents. #Knowledge2026 https://t.co/FLvCVxEYmZ
Sierra Secures $950 Million Funding, Valued at $15 Billion to Scale AI‑Driven CX Platform
Sierra, the San Francisco‑based AI startup, announced a $950 million financing round that lifts its post‑money valuation above $15 billion. The round, led by Tiger Global and Alphabet’s GV, is earmarked for expanding its enterprise AI customer‑experience platform and accelerating global adoption.
AI Aids Learning but Weakens Deep Understanding in Universities
"While AI assistance may ease initial learning, it appears to undermine the effortful processes needed for robust learning. These results have important implications of how generative AI tools should be integrated into higher education." https://t.co/Rsq9ugWvgc https://t.co/0Aimpfd9ob
Claude Feels Like a Trusted AI Friend, Now Scalable
Met some legends and friends today at Code with Claude. Of all the AIs, Claude still feels the most like a trusted friend and I’m glad that now @AnthropicAI has the compute to scale. https://t.co/MpADxwZ06D
White House Weighs Pre‑Release Reviews After Anthropic’s Mythos Triggers Security Alarm
The White House is evaluating a pre‑release review process for high‑risk AI models after Anthropic’s newly released Mythos model raised security alarms. Officials say the move aims to curb potential misuse while preserving innovation, marking a rare federal step into...
AI Web Clicks Cost 45× More Tokens than APIs
Using AI to click around on a website burns 45x as many tokens as just using APIs https://t.co/plSUDqrslx https://t.co/x9aNER0B7g
Google Adds Multimodal Support and Page‑level Citations to Gemini API File Search
Google announced three upgrades to its Gemini API File Search: multimodal data handling, custom metadata filtering, and page‑level citations. The changes let developers build retrieval‑augmented generation (RAG) systems that process text and images together while providing precise source references.
Kodiak AI Inc (KDK) Q1 2026 Earnings Call Transcript
Kodiak AI reported Q1 2026 results, announcing a $100 million equity PIPE that lifts cash to roughly $185 million and funds its scaling plan. The company expanded its driverless fleet to 28 trucks, delivering 15,600 loads and 23,500 paid driverless hours, while...
Seeing MCP
The author highlights a rapid wave of MCP (Machine‑Centric Platform) API expansion that dwarfs earlier REST, GraphQL, gRPC, and Kafka growth. Existing documentation and discovery tools lag, leaving enterprises without clear visibility for governance. A prototype called “See” now generates...
As Execution Becomes Commoditised, Trust Becomes the Differentiator
As AI tools make media execution faster and cheaper, agencies risk becoming interchangeable. Industry leaders argue that the decisive edge now lies in trust, strategic alignment, and deep partnership rather than price or speed. By focusing on big‑picture objectives and...
Unity Software Inc (U) Q1 2026 Earnings Call Transcript
Unity Software reported a 35% year‑over‑year revenue increase in Q1 2026, propelled by an 80% surge in Vector AI revenue and strong growth in both Grow and Create segments. Adjusted EBITDA rose 65% to $138 million, delivering a 27% margin—the best...
Does Kazakhstan’s Power-Generating Capacity Match Its AI Ambitions?
Kazakhstan’s Ministry of Artificial Intelligence and Digital Development signed an MOU with the JMOT04 consortium to build a Tier IV data center costing up to $1.5 billion, plus a 250 MW gas‑fired power plant estimated at $400 million. The project is a cornerstone of...
Infrastructure for AI Agents: What Platform Teams Need to Build Now
AI agents are moving from code‑assistant roles to autonomous platform citizens, demanding infrastructure that operates at machine speed. Traditional cloud platforms built around human‑centric ticketing and manual approvals become bottlenecks, causing delays that AI‑driven workflows cannot tolerate. The article advocates...
Rewiring Europe: Turning Pressure Into Performance
Europe is at a productivity inflection point, pressured by energy shortages, US tech advances, and Chinese competition. McKinsey partners Jan Mischke and Ruth Heuss argue that only a handful of firms are delivering meaningful productivity gains, with AI adoption still...
OpenAI Is Under Criminal Investigation — Why Chatbots Don’t Always Follow the Law
Prosecutors in Florida have launched a criminal investigation into OpenAI after a murder suspect allegedly used ChatGPT to plan a mass shooting at Florida State University. While no charges have been filed against the company, the probe highlights the difficulty...
Why Developer Portals Matter More in the Age of AI Agents
Internal developer portals (IDPs) are not obsolete; they are evolving into infrastructure that serves both human developers and AI agents. While AI assistants can answer queries, they lack the enterprise context that IDPs provide, such as service catalogs, compliance policies,...
The Eval Signal That Predicts 3x Agent Retention
Amplitude analyzed 20K+ power users of its Global Agent and discovered that early agent experiences drive dramatically different retention outcomes. For new users, a positive first‑week session—especially when the user saves the agent’s output—triples long‑term retention, while established users retain...
Google Preps ‘Gemini Agent’ as Your ’24/7 Digital Partner’
Google is upgrading its Gemini Agent, an experimental AI feature introduced with Gemini 3, into a 24/7 digital partner dubbed “Remy.” The new version can perform multi‑step actions such as sending emails, sharing files, and making purchases by leveraging personal context,...
AI-Generated Spencer Pratt Ad Sees Gavin Newsom Eat Cake and Competitor Karen Bass as The Joker | Video
Conservative creators released a 99‑second AI‑generated ad supporting Los Angeles mayoral hopeful Spencer Pratt, portraying Gov. Gavin Newsom eating cake, former Vice President Kamala Harris drinking vodka, and Mayor Karen Bass as the Joker. The video, built with ByteDance’s Seedance generative‑AI...
Memory & Architecture Design Crucial for Inference Era
Good discussion here with @MicronTech Chief Business Officer Sumit Sadana, @AMD CTO Mark Papermaster on memory and architecture design needs in the inference era. https://t.co/H5d1zt1KbM
Estate Agency Leads to Be Contacted by AI Voice System
Estate‑Agents.ai has launched an AI‑driven prospecting platform that uses synthetic voice calls to reach homeowners in an agency’s existing database. The system engages former enquiries and dormant contacts in natural‑language conversations to gauge intent to sell within six months. Responses...
AI Isn’t Taking Graduate’s Jobs, but It Will Reshape Entry-Level Roles
AI is set to reshape, not eliminate, entry‑level graduate roles, with 87% of employers expecting changes to tasks and skill requirements. About 40% of firms anticipate no graduate positions will be replaced in the next three years, while another 40%...
Elon Musk’s Last-Ditch Effort to Control OpenAI: Recruit Sam Altman to Tesla
Elon Musk tried to lure OpenAI CEO Sam Altman to Tesla in early 2018, offering him a seat on Tesla’s board and the chance to lead a "world‑class AI lab" within the automaker. The recruitment pitch surfaced during the Musk...
One Click in Claude Unlocks Vast Data Insights
this is super cool. one click in Claude and you get a ton of fascinating data
5G Advanced Fuels AI, US Leads 6G Future
Awesome time with @CTIA sharing how 5G Advanced is already enabling new AI-powered experiences and why continued US leadership in spectrum and AI-native infrastructure will shape the future of 6G and intelligent connectivity. https://t.co/nBex3bIVcA
‘Tokenmaxxing’: Is Your Company Measuring AI the Wrong Way?
Employers are increasingly tracking AI "tokens" – the basic units of language‑model processing – as a proxy for employee productivity. Experts warn that "token‑maxxing" merely measures tool intensity, not work quality, and can drive hidden costs as token prices remain...
AI Mode Adds Entity Filter Checkboxes for Refined Searches
AI Mode now has these checkboxes to modify your search with the entities displayed https://t.co/pEw3uxl68l
How AI Coding Is Driving Efficiency Gains for Datacom
Datacom is integrating AI‑powered coding assistants such as GitHub Copilot, Cursor and Augment, achieving 20‑30% productivity gains across its digital engineering and SaaS teams. By embedding AI agents that act as business analysts, developers and QA testers, the firm reports...
Law Firm’s AI Tool Translates Legalese Into Gen Z or ‘Sassy’ English
Manak Solicitors launched the Legalese Translator, an AI‑driven tool that rewrites contracts, tenancy agreements and court rulings into plain English using five distinct tones—including Gen Z slang, Boomer, sassy, corporate speak and TL;DR. Built on a siloed ChatGPT model, the service...
Enterprise AI Is Consolidating Fast and Most Companies Are Unprepared #161b
Anthropic’s Claude is shifting from a chatbot to enterprise infrastructure, requiring new operating models like prompt version control, spend telemetry, narrow permissions, and human‑review gates. Reuters cites that about 80 % of Anthropic’s revenue now comes from enterprise customers, with a...
Enterprise AI Cultivates Development Application for Blacktown City Council
Blacktown City Council has deployed Enterprise AI’s Development Application Information System (DAISY), an AI‑driven platform built on Microsoft Azure, to streamline its planning workflow. The solution combines DAISY Assist, which offers site‑specific guidance to applicants, and DAISY Assess, which automates...
A Groq-Powered Agentic Research Assistant with LangGraph, Tool Calling, Sub-Agents, and Agentic Memory: Lets Built It
The tutorial demonstrates how to build a Groq‑powered, agentic research assistant using LangGraph and LangChain. By configuring Groq’s free OpenAI‑compatible endpoint, the workflow runs the 70B Llama‑3.3‑versatile model for multi‑step tool reasoning. It integrates web search, file handling, Python execution,...
When AI Meets Local Debt
Two recent studies examine how AI investment reshapes local borrowing costs, finding opposite effects in the United States and five OECD countries. In U.S. counties, higher AI‑related job postings are linked to lower municipal bond yields, especially for longer‑maturity and...