Today's AI Pulse
Pinterest inks $4B AWS cloud pact to power AI visual search through 2031
Pinterest has signed a $4 billion multi‑year agreement with Amazon Web Services, its largest infrastructure deal to date, to run AI‑driven visual search and discovery tools through 2031. The partnership will expand use of AWS Trainium chips for large‑language and vision‑language models and shift Pinterest’s workloads to a Kubernetes‑based architecture on Graviton processors.
Also developing:
Waymo and Waze Launch Pothole‑Patch Pilot Using Autonomous Fleet Data in Five U.S. Cities
Waymo and Waze announced a joint pilot that will feed pothole detections from Waymo’s robotaxi sensors into the Waze for Cities platform. The program starts in Austin, Atlanta, Los Angeles, Phoenix and the San Francisco Bay Area, where Waymo has already logged roughly 500 potholes. City officials say the real‑time data could speed repairs and improve road safety.
Meta’s Muse Spark AI Asks Users for Health Data, Prompting Privacy and Safety Alarm
Meta launched Muse Spark, an AI assistant that invites users to paste fitness‑tracker, glucose‑monitor or lab‑report numbers for analysis. Early testing revealed inaccurate medical guidance and raised alarms about data privacy, HIPAA compliance and potential misuse of health information.
Pentagon Requests $54.6 B for Drone Warfare Unit, 243‑Fold Budget Jump
The Department of Defense has asked Congress for a $54.6 billion budget for the Defense Autonomous Warfare Group (DAWG) in FY 2027, a 243‑fold rise from the $225 million allocated in FY 2026. The surge reflects a strategic pivot toward low‑cost, expendable drones after...
OpenAI and Anthropic Just Previewed Judgement Day
OpenAI released a 13‑page “Industrial Policy for the Intelligence Age” paper outlining a future where AI drives four‑day work weeks, robot taxes and a public wealth fund. Days later Anthropic announced its Mythos model, which uncovered a decades‑old operating‑system vulnerability...
MercadoLibre Boosts AI Spending as Revenue Jumps 47% YoY
MercadoLibre said its AI investments are accelerating revenue, delivering a 47% year‑over‑year Q4 growth on a constant‑currency basis. The e‑commerce leader is expanding AI across logistics, payments and seller tools while grappling with short‑term margin pressure and a 35% stock...
TriNet Acquires Cocoon to Add AI-Enabled Leave Management SaaS for SMBs
TriNet (NYSE: TNET) announced a definitive agreement to acquire Cocoon, a leading leave‑management SaaS, adding AI‑enabled compliance tools for small‑ and medium‑size business clients. The deal, expected to close in April 2026, expands TriNet’s HR suite without altering its 2026...
China Accelerates Humanoid Rollout with UniX AI's Panther and New 10,000‑Unit Factory
Chinese robotics firms are moving from prototypes to commercial scale. UniX AI introduced Panther, a dual‑arm service robot for homes, while Leju Robotics and Dongfang Precision opened a factory capable of producing a humanoid every 30 minutes, targeting 10,000 units...
Oracle Taps Lucinity Tech to Enhance AI-Driven Compliance Tools
Oracle is integrating Lucinity’s AI‑native investigation technology into its Financial Crime and Compliance Management (FCCM) platform, embedding AI agents into the Oracle AI Investigator tool. The partnership aims to automate manual investigation tasks, surface relevant context, and guide decision‑making for...
Anthropic Flags Claude Mythos AI as Potential Cyber‑attack Catalyst, Limits Release
Anthropic announced a limited preview of its Claude Mythos AI model, warning that the system can discover high‑severity vulnerabilities across major operating systems and browsers. The company is restricting access to a 40‑company consortium while regulators and security experts debate...
AI's Growing Backlash Calls for Ethical Governance
⚠️ AI was supposed to feel exciting. For many people, it feels unsettling, and the backlash is growing. In this video, I break down: ✅ Why AI content is triggering frustration in creativity and culture ✅ The economic concern, massive...
Amazon's AWS AI Revenue Hits $15 B Run Rate as $200 B Capex Push Fuels B2B Growth
Amazon announced that its cloud division, AWS, now generates a $15 billion annualized AI revenue run rate. The company also unveiled a $200 billion capital‑expenditure plan through 2026 focused on AI infrastructure, signaling a decisive bet on enterprise AI growth.
Meta Expands CoreWeave AI Compute Deal to $21 Billion, Total Commitment Hits $35 B
Meta announced on April 9 a $21 billion expansion of its AI cloud agreement with CoreWeave, bringing total spend to $35 billion through 2032. The deal, focused on inference workloads and early access to Nvidia’s Vera Rubin chips, lifts CoreWeave’s share price and diversifies...
In This Critical Part of Audits, the Accountant’s Role Is Shrinking Fast
Accounting giants are rapidly shifting routine audit testing—payroll, expense vouching, and contract verification—to AI agents. KPMG will launch a summer pilot and aim for full deployment of orchestration agents by next year, while EY, PwC, and Deloitte experiment with similar...
Build Your Own Copilot Legal Assistant — No Tech Skills Required
Microsoft’s Copilot Agent Builder lets law firms create custom AI assistants without any coding. By defining a name, purpose, instructions, and uploading firm‑specific documents, attorneys can build agents—such as contract reviewers or client‑communication helpers—in roughly 15 minutes. The tool is...
Autonomous Retail Concept Scales up at Circa Resort
VenHub Global Inc. is teaming with Las Vegas‑based Circa Resort & Casino to debut what the partners call the world’s most advanced autonomous smart‑store. The 66‑foot, three‑unit installation will launch this spring, offering 24/7, staff‑free retail powered by AI‑driven robotics...
Wipe Out a 'Civilization'? Minor Stuff Compared with What Just Happened in AI
Anthropic announced it will not launch Claude Mythos, its most powerful AI model, after internal tests showed the system could autonomously exploit a decades‑old operating‑system flaw and gain control of any server. The company discovered a 17‑year‑old vulnerability in a widely...
More Consumers Tapping AI when Looking for a Deal
Nearly 70% of shoppers now turn to AI tools such as ChatGPT, Gemini, and Claude to locate discounts, according to the 2026 Shopper Study by XCCommerce and SmartBrief. An even larger 83% say savings and rewards drive loyalty to a...
The Company’s AI Agent Refused to Do Math, So I Had It Close My Account Instead
A Fortune 500 company’s AI chat agent could not perform a simple addition, leading the author to terminate the account after a frustrating 15‑minute exchange. The incident occurred while the author tried to verify a spending requirement for a bonus, only...
Control AI Inference Costs: Build Your Own Infrastructure
Reuters reported this week that @AnthropicAI is exploring building its own AI chips. The plans are early. No design, no dedicated team. May not happen at all. That is not the story. The story is why a company at $30 billion in annualized...
U.S. Firm Develops Interceptor Drone with AI Sound Targeting
Talon Avionics of Boise unveiled SECTR, an autonomous counter‑drone platform that combines AI‑driven acoustic sensing with radar to detect, classify and engage hostile drones in under one second. The modular station can launch up to 100 lightweight interceptors, each weighing...
Re: Machine Learning Based Screening of Potential Paper Mill Publications in Cancer Research: Methodological and Cross Sectional Study
In a rapid response to a BMJ study, Professor Min Dai argues that the machine‑learning model used to flag potential paper‑mill articles in cancer research suffers from serious methodological flaws. The training set, drawn largely from the Retraction Watch database,...
Meta Pulls Internal AI Leaderboard After Data-Leak Concerns
Meta quietly retired its internal AI usage leaderboard, dubbed “Claudeonomics,” after internal metrics about employee token consumption began circulating and raised data‑leak concerns. The tool ranked staff by the amount of data processed through generative‑AI models, awarding badges to top...
AI Finds Patterns; Humans Decide What Matters
AI can surface patterns, but it cannot decide which ones matter. That responsibility remains human.
Brittle AI Detectors Fuel Witch Hunt, Erode Trust
For those who follow me on LinkedIn, I have published some thought on the "witch hunt" for authors using AI and why brittle AI detectors (often from profit-oriented companies) are just fueling the epidemic of mistrust. https://t.co/gC0zP9fLH5 https://t.co/awI9AEJbSB
Ultisim Launches Data Fusion Plane, Bringing Optimal AI Infrastructure to Mid-Market Enterprise and Government Market
UltiSim unveiled its Data Fusion Plane, a sovereign AI operating system aimed at mid‑size enterprises and government agencies. The platform lets organizations connect, transform, and query siloed data without ever moving it outside the firewall, avoiding vendor lock‑in and eliminating...
AI Engineering Summit Highlights OpenClaw Founder’s Insights
This whole segment of the AI Engineering summit that @swyx ran with the founder of @OpenClaw @steipete and others is quite excellent. This link should take you to the right beginning over on YouTube: https://t.co/STwp32gvhw
Students Build AI Robot That Tracks and Shoots Moving Balls
Students Design #AI-Powered #Robot That Picks and Shoots Balls in Motion by @lukas_m_ziegler #Robotics #EmergingTech #TechForGood #Innovation #Technology https://t.co/NVNAG7ZvIA
Clarivate Names Mitsubishi Electric to AI50 List
Mitsubishi Electric has been named a Clarivate AI50 2026 organization, recognizing its leadership in AI‑related inventions. The award, based on Clarivate’s four‑factor invention‑strength model, places Mitsubishi among six Japanese firms in the top 50 AI patent holders. This follows Mitsubishi’s...
PeakMetrics Raises Series A to Bring AI-Powered Narrative Intelligence to Organizations Worldwide
PeakMetrics announced a $6 million Series A round led by Moneta Ventures, bringing its total capital to $16.3 million. The company reported a three‑fold year‑over‑year revenue increase as demand spikes for tools that can identify AI‑generated and manipulated online narratives. Its AI‑driven platform...
New Bid to Scale AI-Driven Cotton Programme
US agricultural data platform FarmRaise has partnered with AI‑powered crop developer Avalo, Inc. to expand Avalo’s AI‑driven cotton innovation programme. The collaboration will deliver structured field‑level data capture, streamlined grower workflows, and real‑time visibility across US cotton operations. By standardizing...
Tether Launches QVAC SDK as the AI Universal Building Block that Runs, Trains, and Evolves Intelligence Across Any Device and...
Tether unveiled the QVAC SDK, an open‑source, cross‑platform toolkit that lets developers run, fine‑tune, and evolve AI models directly on any device—from smartphones to industrial servers. Built on a fork of llama.cpp and integrating Whisper, Parakeet, and Bergamot, the SDK...
Deep Dive: Why the US and China Are Leading the AI Race
A new comparative study finds that heavy government spending and deep military integration are the decisive factors behind U.S. and Chinese leadership in artificial intelligence, outweighing private‑sector dynamism or semiconductor independence. Using Qualitative Comparative Analysis, the researchers scored the two...
Meta's Muse Spark Sent Its AI App Soaring up the App Store Rankings
Meta’s new Muse Spark model, the first release from its Superintelligence Labs, sparked a dramatic surge in downloads of the company’s AI app. Sensor Tower reports an 87% day‑over‑day increase in the US, lifting the app to No. 6 among free iOS...
Agentic AI Assistants Set to Be Judge and Jury for Brands
The CX Horizons 2026 report identifies agentic AI assistants as a transformative force in customer experience, shifting brand evaluation from marketing metrics to operational performance. Over 340 global practitioners highlighted AI agents that can negotiate, rebook and purchase on behalf...
Microsoft Strips Extra Copilot Buttons From Windows 11 Apps
Microsoft has started removing unnecessary Copilot buttons from Windows 11 apps. The unnecessary Copilot buttons are disappearing from Notepad, Snipping Tool, Photos, and Widgets. Details 👇 https://t.co/BOnSfig5oC
Robot Labs May Threaten Biologists, Sparks Debate
Will self-driving ‘#Robot labs’ replace biologists? Paper sparks debate by Ewen Callaway @Nature Learn more: https://t.co/CfOHJ2PcV3 #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology https://t.co/36lvOyi0uZ
Silex, Edge Impulse Team on Edge AI Development
Silex Technology and Edge Impulse have teamed up to integrate Edge Impulse’s end‑to‑end AI development platform with Silex’s EP‑200Q system‑on‑module, which runs on Qualcomm’s Dragonwing QCS6490 processor. The joint solution bundles a Wi‑Fi 7 driver and AI acceleration to simplify the...
Efficient Optimization of Large Language Models via Parameter-Efficient Tuning and Adaptive Inference
The paper introduces a unified framework that merges data‑centric learning, parameter‑efficient fine‑tuning, and adaptive inference to streamline large language model (LLM) optimization. By targeting scalability and computational cost, the method trims memory footprints and accelerates inference without sacrificing accuracy. Experimental...
AI Forces SaaS Vendors to Rethink Pricing Models
Software vendors will adapt to AI, finding new revenue streams even as user-based pricing becomes obsolete. AI agents doing more work could decrease revenue if employee numbers fall. Expect pricing evolution. #AI #SaaS #BusinessStrategy https://t.co/7KugKaNFOZ
.NEXT 2026 - AI Cost Management and FinOps Top of Mind for Digital Leaders
At Nutanix NEXT 2026, digital leaders highlighted the rising cost pressures of enterprise AI, especially token‑driven models, and advocated Cloud FinOps as the preferred framework for managing those expenses. Executives cited examples such as a developer consuming 55,000 tokens and generating...
Digital Bank Revolut Introduces AI Tools to Streamline Money Management
Revolut has rolled out an AI-powered assistant called AI by Revolut (AIR) to UK customers, embedding spending insights, portfolio checks, subscription control and travel budgeting into its app. The feature, launched on April 9, 2026, uses natural‑language conversation to act as a...
OpenAI Is Building a Cybersecurity Product for a Select Group of Companies
OpenAI is developing a cybersecurity product that will be offered only to a select group of companies through its Trusted Access for Cyber pilot. The offering, tied to the GPT‑5.3‑Codex model, provides highly capable AI tools for defensive security tasks...
Alibaba Pivots AI Focus From Openness to Profit
Alibaba is shifting its AI strategy. The focus is moving from open-source momentum to revenue generation, a change that could impact developers relying on its Qwen models. It reflects a broader trend. As AI matures, business models are starting to take priority...
Can Universities Adopt AI and Remain in Control?
Universities confront a 92% student adoption rate of AI tools such as ChatGPT, Copilot and Gemini, raising alarms over plagiarism, hallucinations and data security. Edward Müller of Mint Group argues that institutions should move from restriction to building in‑house, governed AI...
Employee-Built AI Is How Companies Escape the Workslop Trap
Generative AI is spawning "workslop"—polished but context‑free content that employees must edit, costing up to $9 million a year for a 10,000‑person firm. While 92% of companies plan to increase AI spend, only 1% consider their deployments mature, leaving a gap...
What AI CEOs Still Don't Get About Washington
AI CEOs, led by OpenAI’s Sam Altman and Anthropic’s Dario Amodei, are championing sweeping governance proposals ranging from tax reform to model transparency. OpenAI’s new industrial‑policy paper likens AI’s impact to the Industrial Revolution, while Anthropic pushes internal audits, export...
HR AI Is Becoming a Change Management Story
AI is reshaping HR software by automating repetitive tasks such as survey analysis, training recommendations, and routine employee queries. The real breakthrough, however, is the shift from a pure technology story to a change‑management challenge once AI is embedded in...
DOL’s Workplace AI Strategy Follows Historical Approach to Technology
The U.S. Department of Labor is adopting a historically consistent, guidance‑first strategy for AI, emphasizing workforce readiness before regulation. In February 2026 it issued an AI Literacy Framework to teach workers how to generate and evaluate AI outputs. The agency...
Rethinking Professional Development in the Age of AI with Allison Rodman – Bonus Episode with Jotform
In a bonus Jotform podcast, educator Allison Rodman explains how artificial intelligence can transform professional development (PD) from a one‑size‑fits‑all model to a personalized experience. She argues that pre‑session surveys combined with AI‑driven data analysis enable facilitators to tailor content to...
AI Agents Aren’t Failing. The Coordination Layer Is Failing
Enterprises deploying multiple AI agents often see impressive isolated performance, but production systems quickly degrade as agents compete for resources. Direct point‑to‑point calls cause quadratic growth in connections, leading to race conditions, stale context, and cascading failures. The author proposes...