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
When AI Transparency Backfires
Researchers from UNSW and Wharton reveal that popular explainable‑AI tools, especially partial dependence plots, can be deliberately manipulated to mask bias while model decisions remain unchanged. By inserting synthetic feature combinations that rarely occur in real data, models can produce neutral‑looking visualizations yet continue discriminatory outcomes. The study, based on insurance pricing models, highlights a gap between perceived transparency and actual fairness. Executives relying on these dashboards risk regulatory scrutiny and reputational harm.
AI Elevates Generalists; Judgment and Vision Become Priceless
"AI will also shift the power balance to generalists and to people who can use AI to teach themselves new skills to contribute value. "In our companies, we're seeing the blending of the designer, product manager and engineer. Now what's...
AI and Crypto Empower Agents, Automate Online Shopping
What our lives will look like when the crypto x AI future is in full bloom, according to Ali Yahya of @a16zcrypto: 1. having AI eliminate the tedium involved with buying online 2. telling your agent what you're looking for and...
AI and Robotics Leaders to Headline Automate 2026
The Association for Advancing Automation announced the keynote lineup for Automate 2026, the largest North American robotics and automation trade show, scheduled for June 22‑25 in Chicago. Industry heavyweights from Siemens Digital Industries and Standard Bots will lead sessions on AI‑driven automation,...
Anthropic Previews Orbit: Proactive Assistant for Knowledge Workers
Jarvis will need to be proactive -> Anthropic is working on Orbit, its upcoming proactive assistant "Across recent web and mobile builds, more references and supporting scaffolding have surfaced, though for now the tool only manifests as a toggle in the...
Leaders Win by Mastering Attention, Not Doing More
What's the leader's role in AI? As your AI agents begin to read between the lines of your life and your business, your role is not to do more. It is to see more clearly. To decide what matters. To...
AI Risk and the Very Large Hammer
The article argues that AI risk is undeniable but policy should target the specific capabilities where misuse occurs, not blanket regulation. It critiques the emerging independent verification organization (IVO) model, noting its reliance on state‑backed enforcement and potential conflicts of...
VVV Locks Yield AI Compute Tokens as Demand Soars
DIEM & VVV tokenomics... • 1 DIEM = $1/day of daily renewing AI compute credits, spendable on any model from Qwen to Opus 4.7 to Grok to Nano Banana via Venice (app or api) • As demand for AI compute rises, DIEM...
Some Acts Lose Meaning When Made AI‑Legible
Something @bhalligan posted yesterday has stayed with me: not everything in our lives or in our companies should become legible to AI. I think I resisted this at first because I love this idea of being totally AI-pilled. On second...
InsideTracker AI Study Shows Sustained Improvements in 43 Biomarkers Across 20,000 Users
InsideTracker’s AI‑driven health platform was linked to statistically significant, sustained improvements in 43 blood and fitness biomarkers among a real‑world cohort of 20,000 users over more than four years. The peer‑reviewed study validates the platform’s personalized recommendation engine and marks...
AI Has a PR Problem
The AI industry faces a growing public relations crisis as recent polling shows a starkly negative perception of the technology. A March 2026 NBC News poll found only 26% of U.S. voters view AI positively, while 46% hold a negative...
Anthropic Launches 10 AI Agents for Finance Tasks
Work in finance? -> Anthropic unveils 10 new AI agents for the financial sector, including for drafting pitch decks, reviewing financial statements, and escalating compliance cases "Anthropic PBC unveiled a set of new artificial intelligence agents designed to handle a...
From Answers to Influence: How AI Is Reshaping Social Platforms’ Role in Search
AI is fundamentally reshaping how social platforms serve as search gateways, moving from simple answer delivery to influence‑driven discovery. MIDiA’s Q3 2025 consumer survey shows 52% of generative‑AI users rely on these tools to locate information. Platforms are embedding large language...
Google’s Back‑to‑Back 2026 Core Updates Slash Organic Traffic for 55% of Sites
Google rolled out two major core algorithm updates in March and April 2026, triggering traffic declines for more than half of tracked websites. The March update targeted large‑scale AI‑generated content, while the April update emphasized “information gain,” rewarding pages that...
Treasury Warns AI Model Could Hack Bank Accounts, Urges Immediate Action
Treasury Secretary Scott Bessent told Fox News that Americans should be worried about AI-driven hacks to their bank accounts, citing Anthropic’s Claude Mythos model that can locate thousands of software flaws. The warning follows an emergency April meeting with Federal Reserve...
Tokenopoly Debuts First AI Real‑Estate Agent for Tokenized U.S. Homes
Tokenopoly introduced an autonomous AI real‑estate broker, OpenClaw, at the Consensus 2026 conference. The skill lets AI agents buy, sell and manage tokenized U.S. homes on Ethereum without human input, marking a new frontier for PropTech and DeFi integration.
LinkSquares Unveils First All‑Agentic CLM Platform, Automating Drafting, Redlining and Execution
LinkSquares announced the launch of an all‑agentic contract lifecycle management platform that automates drafting, redlining and workflow execution. Built on its LinkAI engine, the solution claims to cut hours‑long tasks to minutes, marking a shift from AI‑assisted to AI‑driven contract...
Medline Launches Mpower Digital Control Tower, Promising 50% Workflow Boost for Hospitals
Medline announced the launch of Mpower™, an AI‑driven digital control tower that gives health systems real‑time supply‑chain visibility. Early adopters report more than a 50% efficiency gain in order substitution workflows and a 1‑2% rise in unadjusted fill‑rates, positioning the...
IBM Study Finds 59% of CEOs Expect CHRO Influence to Rise with AI
IBM’s Institute for Business Value surveyed 2,000 global CEOs and found 59% anticipate a stronger CHRO role as artificial intelligence becomes a core strategic driver. The study also notes a surge in Chief AI Officer appointments, underscoring a broader re‑engineering...
ABB Robotics Launches PickMaster Lite to Simplify & Accelerate Robotic Picking
ABB Robotics introduced PickMaster Lite, a streamlined version of its PickMaster suite aimed at simplifying high‑speed, vision‑guided robotic picking for packaging OEMs and system integrators. The software reduces engineering effort by 30% and cuts commissioning time by 25% through pre‑configured...
U.S. Data Center Boom Pushes Electrical Equipment Market to $65B, Sparks Power Supply Strain
Wood Mackenzie projects the U.S. data‑center electrical equipment market will swell to $65 billion by 2030, with transformer demand climbing to over 9,000 units. The surge, driven by hyperscalers’ AI‑fuelled expansion, is forcing utilities, cities and regulators to confront unprecedented power‑supply...
IShares Semiconductor ETF Soars 40% in April on AI-Driven Chip Rally
The iShares Semiconductor ETF (SOXX) posted a 40% gain in April, far outpacing the S&P 500. The rally was powered by strong earnings from Nvidia, Intel, TSMC and other chip leaders, as hyperscalers and AI workloads drove unprecedented demand for...
Dealing With AI Pressures: Thriving With An Abundance Of Knowledge
Law firms are scrambling to acquire AI technologies to satisfy increasingly demanding clients. However, many clients lack clear expectations, and firms often purchase tools they don’t understand, resulting in wasted budgets and underutilized solutions. The mismatch creates frustration on both...
ServiceNow Appoints Former Philips Exec Srikanth Akkiraju as Chief Transformation Officer
ServiceNow announced the hiring of Srikanth Akkiraju, a former Philips senior leader, as Chief Transformation Officer. The role will focus on guiding Fortune 500 customers through AI‑led operating‑model redesigns and large‑scale automation initiatives, underscoring ServiceNow’s push into deeper enterprise transformation...
Customer Contact Week Europe Amsterdam 2026
Customer Contact Week (CCW) Europe 2026 will take place in Amsterdam from October 5‑7, gathering senior CX and contact‑center leaders at the waterfront Passenger Terminal. The conference features over 250 speakers across 125 sessions, covering AI, automation, journey design, data‑driven decision‑making,...
South Korea and Taiwan Set Record Highs as AI Chip Rally Fuels Asian Markets
South Korea's KOSPI and Taiwan's TAIEX surged to fresh all‑time highs on Monday, with the KOSPI up 5.1% to 6,937 and the TAIEX climbing 7% as AI‑related semiconductor earnings lifted market sentiment. The rally was powered by SK Hynix’s 12%...
Observe by Snowflake: AI-Powered Observability at Scale for the Data Cloud
Snowflake's Observe platform unveiled an AI‑driven CLI that lets engineers and autonomous agents query and act on telemetry via reusable workflows. The update adds native read/write support for Apache Iceberg, enabling observability data to reside in low‑cost cloud storage and...
AWS Exec: Revamping MAP Program Is ‘Game-Changing’ For Partners
AWS has revamped its Migration Acceleration Program (MAP) to include funding for new AI workload builds, expanding the program beyond pure migrations. The update adds performance‑based rewards such as faster pre‑sales benefits, delivery‑excellence badges, and visibility via the Solution Matching...
Luma Uni-1 Is Now Available via Partner Nodes
Luma’s Uni‑1 image model is now accessible in ComfyUI via Partner Nodes, allowing users to run the model directly in the popular open‑source workflow or on Comfy Cloud. Unlike diffusion‑based generators, Uni‑1 is a decoder‑only autoregressive transformer that treats text...
Panthalassa Raises $140 Million Series B to Deploy Ocean‑Based AI Compute Nodes
Panthalassa announced a $140 million Series B round led by Peter Thiel, aimed at completing a pilot factory near Portland and launching its Ocean‑3 AI inference nodes at sea. The financing brings together a roster of high‑profile investors and targets the...
Your AI Calendar Agent Is Failing Because You Haven’t Told It What to Do
In a recent workshop, participants struggled with AI calendar agents that scheduled unwanted meetings and missed important ones because their prompts were overly vague, often limited to “Help me manage my calendar.” The author argues that, like a doorman, an...
TDK Ventures' $500M AI Play Targets Inference Chips and Rugged Robots
Nicolas Sauvage is steering TDK Ventures, the corporate venture arm of Japanese electronics giant TDK, with a $500 million portfolio that backs low‑profile AI infrastructure such as inference chips, solid‑state grid transformers and warehouse robots. His four‑year horizon strategy aims to...
China's AI‑Enabled ‘Wolf Pack’ Drone Swarms Target Taiwan Conflict
A Foundation for Defense of Democracies report reveals China is developing AI‑powered quadruped “wolf pack” drones designed for scouting, supply and combat support in a possible Taiwan invasion. The analysis warns the systems could lower the political and military costs...
‘There Were a Lot of Naysayers’: Half of T-Mobile’s Customer Calls Are Now AI — And That’s Just the Beginning
T‑Mobile has moved half of its customer‑service calls to voice AI, handling more than 200,000 AI‑driven interactions each day. The company’s AI strategy focuses on proactive problem‑solving, aiming to resolve issues before customers need to call. Director of AI engineering...
UAE Cyber Security Council, Cisco and Open Innovation AI Launch National AI Test and Validation Lab
The UAE Cyber Security Council, Cisco and Open Innovation AI announced the creation of a National AI Test and Validation Lab in Abu Dhabi. The facility will certify AI models for security, safety and compliance, marking the first government‑backed AI...
Verisk Integrates Insurance Analytics & Gen AI with Anthropic’s Claude via MCP Connectors
Verisk has integrated its regulated insurance analytics with Anthropic’s Claude Enterprise using its Model Context Protocol (MCP) connectors. The partnership adds a conversational layer that lets underwriters and restoration contractors query proprietary data—such as ISO loss‑cost trends and pricing intelligence—through...
Accelerating Gemma 4: Faster Inference with Multi-Token Prediction Drafters
Google announced Multi‑Token Prediction (MTP) drafters for its Gemma 4 family, delivering up to three‑fold faster inference without sacrificing output quality. The drafters use speculative decoding, pairing a lightweight draft model with the full‑size Gemma 4 target to predict several tokens in...
AI Agents Emerge as Enterprise Workforce Layer
AI agents are becoming a new enterprise workforce layer. @IrenaCronin and I write this newsletter every week. AI agents are becoming a new enterprise workforce layer because they can move beyond simple chatbot tasks and operate inside real business workflows. They can...
AI Empowers All Practices to Achieve Healthcare Equity
#AI's #Healthcare Promise: Creating Healthcare Equity By Empowering All Practices by Sameer Bhat @Forbes Learn more: https://t.co/fg209YO96e #MedTech #HealthTech #Tech #TechForGood https://t.co/c6Bg9MFeq6
Jamie Dimon Blesses the Trillion-Dollar AI Capex Boom
JPMorgan Chase chief Jamie Dimon stood beside Anthropic CEO Dario Amodei in New York, declaring the ongoing AI build‑out worth every dollar. He highlighted a projected AI capital‑expenditure wave that could exceed $1 trillion by next year, underscoring its role in...
AI and Crypto Together Combat Deepfake Misinformation
🤖🥸 "AI can help address the problems of AI deep fakes, and help us reason about he provenance of media and information, so crypto and AI are good counterweights for one another." -- @alive_eth https://t.co/5tGKOqyEEC
Distribution, Not Models, Is AI’s Biggest Bottleneck
OpenAI: $10B JV with TPG, Brookfield, Bain, SoftBank. Anthropic: $1.5B JV with Blackstone, Goldman, Hellman & Friedman. Same day. Same bet. AI's bottleneck isn't the model. It's distribution. And no consulting firm made either cap table. https://t.co/sxE6VUEoAG
AI Agents Expose a Costly Blind Spot in Fraud Prevention
Merchants are now seeing legitimate AI‑driven purchases flagged as fraud as AI shopping agents mimic bot behavior. Traditional fraud models, which rely on human‑centric signals, struggle to distinguish authorized agents from malicious bots. The resulting false declines generate lost revenue...
AI Must Grasp Architectural Intent, Not Just Render Images
Scot from Bureau on our podcast last week: "Generic AI image tools don't listen. They don't even know what my product is." The most useful AI in AEC understands architectural intent. Otherwise, it's just pretty images. Full convo: https://t.co/5LFVbNQITz
IBM Targets Day 0 Customer Engagement Across AI, Cloud, Quantum
Meet the customers where the are om Day 0 is the goal of @IBM says @ArvindKrishna - on the three innovation vectors - AI. Hybrid Cloud and Quantum. #IBMThink https://t.co/MpxGStKX9H
The Coming Hackastrophe
The Atlantic and NYT highlight a looming shift as AI models such as Claude Mythos enable bots to discover and exploit software flaws at scale. Experts warn that within a year or two these tools could make most existing applications...
OpenAI Might Follow Netscape’s Decline Without AI ‘Homepage’
"Does OpenAI go to zero the way Netscape did?" @fubini asked me this. We both interviewed at Netscape the same day in '95. His answer: "The “homepage” for AI hasn't shown up yet." Deployed EP4 is live. Link in comments. https://t.co/NzofcUJFwa
Google Funds AI Agents to Own Business Outcomes
Google just dropped a $750M fund and a new enterprise agent platform. The vision: AI that owns business outcomes, not just tasks. The problem nobody solved: reliability, accountability, compliance, security. Ambition is not a product.
AI Is a Thinking Challenge, Not a Workplace Tool
The biggest AI mistake: treating it like a workplace problem. It’s not. It’s a thinking problem. Kids won’t “discover” AI someday. They’ll grow up inside it. The future won’t belong to kids who press the right button. It will belong to kids who ask better...
Bob Demonstrates Agentic‑Era SDLC Automation at IBM Think
At the 2nd keynote of day 01 at #Think Neel Sundaresan, General Manager, Automation and AL and Sanah Pallithotungal, Product Manager, IBM Bob gave us the demo of how Bob enables SDLC in Agentic era. #IBMThink #Think #Think2026 https://t.co/1kxeKLtKGo