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
OpenAI Partners with Customers Bank to Slash Loan Times and Boost Efficiency
Customers Bank, a $25.9 billion asset lender, has inked a multiyear partnership with OpenAI to embed AI engineers and automate lending, deposits and payments. The initiative promises to shrink commercial loan cycles from up to 45 days to about seven and lift the bank’s efficiency ratio from roughly 49% to the low‑40s.
Lovelace AI Launches Elemental Context Engine for Mission‑Critical Enterprise Use
Lovelace AI emerged from stealth today, introducing Elemental—a context‑engine platform that builds secure knowledge graphs for mission‑critical enterprise AI. Founder Andrew Moore says the system can answer investigative queries using only one‑thousandth the tokens of conventional models, targeting sectors where...
GoodData Unveils Agent Builder to Accelerate Agentic AI Development
GoodData released Agent Builder on April 22, a low‑code framework that transforms its semantic layer into deployable AI agents. The tool promises minutes‑level creation, governance and reuse, targeting both enterprise analysts and independent software vendors.
OpenAI Rolls Out $3‑$5 CPC Ads in ChatGPT, Targeting Performance Marketers
OpenAI introduced cost‑per‑click (CPC) advertising inside ChatGPT, letting brands bid $3‑$5 per click. The move pits the AI chat platform against Google Search and Meta, raising questions about intent, pricing and the viability of a single‑session marketing funnel.
AI Startups Claim 7 of Top 10 Seed Rounds, Avg Over $10 M
AI‑focused startups captured seven of the ten largest seed rounds recorded in the past few months, with average seed investments topping $10 million. The surge reflects capital‑intensive model training needs and a rush by venture firms to lock in early positions...
Box Launches Box Automate, AI‑Powered Workflow Service
Box announced Box Automate, an AI‑driven workflow automation service, expanding its SaaS portfolio and signaling a shift toward AI‑enhanced content management. The new offering aims to accelerate customer operations, though pricing and rollout details were not disclosed.
DigitalOcean Unveils Inference Engine, Cutting AI Costs Up to 67% and Boosting Speed 3‑Fold
DigitalOcean announced its Inference Engine, a suite of production‑grade AI services that includes an Inference Router designed to lower inference costs by up to 67% and accelerate token delivery threefold versus Amazon Bedrock. The launch targets developers building agentic systems...
Wingbeat Radar Signatures Let AI Sort Bees, Wasps and Other Insects
Researchers at Trinity College Dublin demonstrated that millimeter‑wave radar combined with machine‑learning can identify flying insects by their wing‑beat signatures. By extracting over 70 harmonic, spectral and temporal features from micro‑Doppler reflections, the AI model achieved 96% accuracy distinguishing bees...
Improving Access to Essential Medicines via Decision-Aware Machine Learning
A new wave of decision‑aware machine‑learning models is being applied to pharmaceutical supply chains in low‑income regions, blending demand forecasts with inventory and distribution constraints. Early pilots in Zambia and Rwanda report up to a 30% reduction in stock‑outs and...
Can Agentic AI (Finally) Modernize Core Technologies in Insurance?
Insurers have long struggled to modernize legacy core systems, citing complexity, cost and risk. Agentic AI—autonomous software agents that can reverse‑engineer code, generate configurations, run tests and orchestrate cutover—promises to reshape the modernization cost curve. McKinsey’s analysis shows productivity gains...
Using NVIDIA Aerial CUDA-Accelerated RAN on Red Hat OpenShift to Accelerate Development of AI-Native 5G and 6G RAN Solutions
Red Hat demonstrated NVIDIA Aerial CUDA‑Accelerated RAN running on OpenShift, using an NVIDIA GH200 Grace Hopper Superchip and ConnectX‑7 NIC in a single‑node edge configuration. The open‑source stack enables AI‑native 5G/6G development with zero‑touch GitOps provisioning and GPU‑accelerated PHY performance....
Where AI Will Create Value—And Where It Won’t
McKinsey warns that AI’s biggest business impact will come from reshaping offerings, models and market structures, not from incremental productivity gains. While nine‑in‑ten firms have deployed AI, 94% report little measurable value, echoing the historic Solow paradox. Early productivity wins...
How VEED Evolved Its AI Search Strategy
VEED, an AI video creation platform, is reshaping its SEO strategy to target AI‑generated search results, or AEO, after data shows 800 million weekly ChatGPT users and 2 billion monthly AI Overviews users. The company now monitors a new visibility score that...
Verisk Analytics Inc (VRSK) Q1 2026 Earnings Call Transcript
Verisk Analytics reported 6.6% organic constant‑currency revenue growth and 8.5% adjusted EBITDA growth for 2025, with adjusted EBITDA margin expanding to 56.2%. Subscription services now represent 84% of revenue, while transactional revenue slipped 6.5% amid low weather activity. The company...
Making AI Search Count (and Convert)
AI search has shifted from a ranking‑centric model to one where brands must earn citations from large language models. Georgie Kemp of VEED highlights Amplitude's AI Visibility add‑on, which now measures visibility scores, sentiment, and revenue impact of AI‑sourced traffic....
Manifest Raises $2 Million to Help Ad Agencies Eliminate Timesheets
Australian startup Manifest has closed a $2 million AUD pre‑seed round—about $1.3 million USD—led by Brand Fund and backed by Antler, Techstars and several ad‑industry veterans. The company’s AI‑powered platform automatically records agency workflows across tools, replacing traditional timesheets with real‑time visibility...
New Square AI Agent Automates Daily Tasks for Main Street Sellers
Square, a Block subsidiary, has launched Managerbot, an AI‑powered business agent, in open beta for U.S. non‑franchise food, retail and health‑beauty merchants. Integrated directly into the Square Dashboard, Managerbot goes beyond the conversational Square AI by monitoring operations, surfacing insights...
AI Gets a Museum; Its Story Cracks
Refik Anadol’s Dataland, billed as the world’s first AI‑generated art museum, is slated to open in June inside Frank Gehry’s Grand LA complex. At the same time, a new Google DeepMind paper contends that large language models will never achieve consciousness,...
How to Build Custom Reasoning Agents with a Fraction of the Compute
Researchers at JD.com introduced Reinforcement Learning with Verifiable Rewards and Self‑Distillation (RLSD), a hybrid training paradigm that merges reliable RL direction signals with dense token‑level credit from self‑distillation. Experiments on the Qwen3‑VL‑8B model show RLSD achieving a 56.18% average accuracy...
OmniOps Partners with Grafana Labs to Deliver Sovereign Observability in Saudi Arabia
OmniOps, Saudi Arabia’s leading AI‑infrastructure provider, joined Grafana Labs’ global partner program to deliver Grafana Cloud‑based observability services from within the Kingdom. The partnership enables enterprises to monitor metrics, logs, traces and AI workloads on a single open platform while...
Army’s Project ARIA Seeks to Accelerate AI Adoption Across the Force
The U.S. Army launched Project ARIA to fast‑track artificial‑intelligence integration across its force. The initiative focuses on three thrusts: a “model armory” that delivers AI tools to soldiers at the tactical edge, automating the Planning‑Programming‑Budgeting‑Execution (PPBE) process, and building a...
AI Is Changing Who Wins Research Grants
A Northwestern Innovation Institute study examined confidential grant proposals from two universities alongside all NIH and NSF awards from 2021‑2025, revealing a sharp rise in AI‑assisted writing after 2023. At the NIH, proposals with higher AI involvement secured more funding...
ASU+GSV 2026: Tactile Learning Tools Lay Foundation for AI Education
At the 2026 ASU+GSV Summit, LEGO Education’s head of education impact, Jenny Nash, advocated for tactile learning tools as a gateway to AI literacy. She argued that kindergarteners should prioritize hands‑on play with LEGO bricks rather than early screen time...
30 ChatGPT Prompts Every Marketer Needs to Save Right Now
An industry blog shares 30 curated ChatGPT prompts designed for marketers across product descriptions, social media, email, ads, and more. Each prompt embeds a proven copywriting framework—such as AIDA, PAS, or STAR—to steer the AI toward focused, high‑impact copy. The...
How Unstoppable AI Is Reshaping UK Legal Practice
AI is rapidly reshaping legal practice in the United Kingdom, where lawtech investment rose more than a third in 2025, setting a new record. A pragmatic regulatory framework and the absence of a U.S.-style unauthorized‑practice rule give UK firms an...
Goldman Stops Bankers Using Anthropic’s Claude in Hong Kong
Goldman Sachs has barred its Hong Kong bankers from accessing Anthropic’s Claude AI models, citing a strict reading of its contract after consulting the startup. The move reflects heightened US‑China tensions over artificial‑intelligence technology, as Western AI services are already blocked...
Indeed Rolls Out ChatGPT App to UK Jobseekers
Indeed has launched its ChatGPT app for UK jobseekers, extending the service that debuted in the United States in February to more than 50 countries. The integration lets candidates invoke a personalized job search by mentioning @Indeed in a ChatGPT...
Defence AI Beyond the Headlines
The US military leveraged AI‑enabled tools, including Palantir’s Maven system and Anthropic’s Claude, to strike over 1,000 targets in the first 24 hours of the Iran conflict, a tempo far beyond human capability. While AI accelerated data fusion, object classification...
SOCOM Adding AI, Autonomy ‘at Every Level,’ Commander Says
U.S. Special Operations Command (SOCOM) told the Senate Armed Services Committee that AI and autonomy are being embedded at every level of its operations. Admiral Frank “Mitch” Bradley highlighted AI’s role in battlefield sensing, target surveillance, and precision strike, as...
AI Gains Concentrate in Clean-Signal Tasks, Not Casual Use
Karpathy's framing of the AI debate is the cleanest I've seen: Two groups. Same industry. Opposite conclusions. → Group 1: judged AI on free/old models. Saw the failures. Wrote it off. → Group 2: uses frontier models for hard technical work. Progress feels...
Winning Agentic AI Requires Clean Data, Rigor, Not Flashy Bots
"If you’re trying to pick winners in agentic AI, don’t look for those with the cleverest agents. Instead, look to the companies with the cleanest data contracts, the best evaluation discipline, the most coherent identity model, and the least tolerance...
AI Model Detects Normally 'Invisible' Tissue Changes of Pancreatic Cancer at Stage 0
Researchers unveiled REDMOD, an AI radiomics framework that identifies stage 0 pancreatic ductal adenocarcinoma on routine CT scans. In a multi‑institutional study of 219 cancer cases and 1,243 controls, REDMOD flagged disease an average of 475 days before clinical diagnosis, achieving 73%...
Firm Leaders 'Determined' To Train Attorneys to Avoid Using AI-Produced Mistakes
Law firms are rolling out comprehensive training programs to ensure attorneys recognize and avoid errors generated by artificial‑intelligence tools. Partners like Robert Hays of King & Spalding say clients demand AI adoption, but only if it is used responsibly and...
Legaltech Connect: How Baker & Hostetler, Eversheds Prepare Staff for AI Adoption
Law firms Baker & Hostetler and Eversheds Sutherland are actively preparing their attorneys and support personnel for generative AI integration. They have rolled out firm‑wide training programs, pilot projects in contract analysis, and dedicated ethics oversight bodies. The initiatives aim...
Compare AI Automation Cost to Current Task Expense First
MyPOV - Every question about the cost of AI for automating a specific task must spur the immediate question - what does the task cost right now, before AI. Once vendors provide the AI cost - you still don't know...
Create Unit Vocabulary Tables Instantly with One Prompt
📝 Vocabulary lists organized by unit shouldn't take an entire afternoon. 🤖 One prompt can generate a structured table with terms and student-friendly definitions. 🔁 Then reply to add image cues or cross-curricular connections. https://t.co/s9XnH6TjMf
AI Strategy Pillars, New SMB Procurement Program Revealed in Canada’s Spring Economic Update
Canada’s Spring Economic Update unveiled six pillars for a forthcoming national AI strategy, emphasizing privacy safeguards, AI training, SME adoption, sovereign compute infrastructure, growth capital for Canadian champions, and international standards cooperation. The update also announced a Small and Medium...
Leaders Demand AI Guardrails, Yet Few Implement Them
92% of leaders agree AI agents need guardrails, but less than half have defined them. Time to move from experimentation to orchestrated, safe execution. Details here: https://t.co/iheEM08maa @Appian @HarvardBiz #AppianWorld https://t.co/9eHPgb09df
AI Updates Outpace Everyone Except Hardcore Nerds
It’s genuinely impossible to keep up with every AI tool’s platform updates. Iteration is happening so fast that anyone but hardcore AI nerds can never catch up.
Opinion: 5 AI Moves Leaders Must Make for Next School Year
The U.S. Department of Education’s new rule prioritizing AI in discretionary grants is pushing K‑12 districts to accelerate adoption for the 2025‑26 school year. Experts warn that success hinges on strategic leadership, not just tool selection, emphasizing problem‑first thinking, unified...
Integrate AI Into Workflows to Unlock Real Value
59% of companies have AI in production, but only 16% see high measurable value. The secret? Integration. Learn how @Appian is closing the AI success gap by embedding intelligence directly into workflows: https://t.co/Ei8vLflWQB #AppianWorld https://t.co/4q3TlGfBQm
AI Integration Hurts Dev Process; Replace Workflow Entirely
AI Bolted Onto Your Dev Workflow Made It Worse. IBM Bob Thinks the Fix Is Replacing the Workflow Entirely. https://t.co/NgS9s8qnYr
The Future of Software Development: Now with Less Software Development
At AI Dev 26x SF in San Francisco, more than 3,000 developers explored how AI will reshape software engineering. Speakers from AMD, AWS, and Actian showcased tools that automate code generation, enforce correctness, and support hybrid infrastructure. Andrew Ng argued that small, generalist teams could...
OpenAI and AWS Discuss Compute, Infra, Token Budgets
Met some interesting folks from OpenAI at AWS event today. Some good discussions around compute/infra/token budgets, etc.
AI Teammates Evolve From Tools to Boardroom Assets
From Tool to Asset: Why the AI Teammate is a Boardroom Priority by @Timothy_Hughes https://t.co/rgSzfVXvra @DLAIgnite #SocialSelling #DigitalSelling #Sales #SalesTips #SalesLeader #ArtificialIntelligence #Technology #Innovation #Tech #TechNews #Marketing #Strategy https://t.co/QZPJ0noiXY
The Future of Software Development: Now with Less Software Development
Over 3,000 developers gathered at AI Dev 26 x SF in San Francisco to debate how artificial intelligence will reshape software engineering. Speakers from AMD, AWS, Actian and DeepLearning.AI argued that imagination, speed and spec‑driven development are the new bottlenecks, with AI agents poised to...
Simplify Agent Orchestration with DB‑Stored Function Architecture
this is a great read on thinking through how to simplify agent orchestration to it’s core interestingly converges on an architecture very similar to babyagi 2, which stored functions (agent tools) in a db with input/output/dependencies/keys, and the started agent could...
AI Wearable Turns Daily Life Into Continuous Vlog
An AI wearable device that regularly captures footage throughout the day and compiles it into a full-day vlog.
The Future of Software Development: Now with Less Software Development
Over 3,000 developers gathered at AI Dev 26 x SF, hosted by Andrew Ng’s DeepLearning.AI, to debate software engineering’s future in an AI‑driven world. Speakers highlighted that AI moves the bottleneck from writing code to imagination, funding and speed, with...
AI, Face Photos May Predict Cancer Survival: Mass General Brigham Study
Mass General Brigham researchers validated FaceAge, an AI tool that estimates biological age from facial photographs, as a survival predictor for cancer patients. The study examined two routine photos per patient for 2,279 individuals, finding that changes in the Face...