Today's AI Pulse
Macron bets on AI with €75 bn SoftBank pledge to cement France’s tech future
French President Emmanuel Macron is securing a SoftBank pledge of up to €75 bn (≈$86 bn) for AI projects, branding the effort “Project Marengo.” The plan leverages France’s nuclear power to build low‑cost data centres and position the country as Europe’s AI hub.
Also developing:
By the numbers: Waymo acquires Apple’s Arizona Proving Ground for $220M
Cresta Launches Conductor, the Agent for AI Agent Development
Cresta unveiled Conductor, a developer‑first engine that lets technical teams build, test, and optimize production‑grade conversational AI agents using natural language. The platform generates a full blueprint—including prompts, sub‑agent orchestration, configurations, and custom code—before any code is written, accelerating development by up to 2 times. Integrated with Cresta’s Testing Suite and Synthetic Customers, Conductor supports iterative pre‑launch testing and continuous post‑launch diagnostics. The launch targets enterprises struggling with the complexity of scaling reliable AI agents across disparate systems.
Nvidia Launches AI Factory Manager Blueprint for Autonomous Manufacturing
Nvidia unveiled the Factory Operations Blueprint (FOX) at GTC Taipei, offering a reference design for an AI‑driven factory manager that unifies machine data, quality systems, robot fleets and alerts. Built on the NemoClaw framework and Nemotron models, FOX integrates with...
AI Is Scaling Healthcare Costs Because the System Was Built That Way
AI is not creating new healthcare costs; it is accelerating the existing reimbursement model that rewards billing intensity. Revenue‑cycle management tools powered by AI now represent a $20 billion market and are projected to triple by 2030, driving higher claim values...
Stop Trying New AI Tools
The post argues that the real value of generative AI lies not in the underlying model but in the personalized context built over time. By consistently using a single tool, users accumulate a “moat” of voice, history, and preferences that...
National Data Center Backlash SITREP
The national backlash against AI‑driven hyperscale data centers is moving from local zoning disputes to state utility commissions, governors’ offices, and courts. Tier‑1 hotspots include Nashville Zoo (over 330,000 petition signatures), Coweta County, Georgia (9,000 MW load and eminent‑domain pressure), Hill...
AI Search Runs On Two Memory Systems. The Platforms Don’t Use Them The Same Way via @Sejournal, @DuaneForrester
Duane Forrester explains that AI‑driven search relies on two distinct memory systems—parametric memory baked into the model during training and live retrieval of fresh content. Search engines adopt different "memory postures": some (e.g., Perplexity, Google AI Overviews) retrieve on almost...
PharmaShots Magazine-June-2026 Edition
PharmaShots’ June 2026 edition spotlights how artificial intelligence, real‑world data and digital listening are reshaping post‑market safety monitoring into a proactive, predictive discipline. The issue examines AI‑driven pharmacovigilance tools—from social‑media signal detection to real‑world evidence generation—and the regulatory governance needed...
KFSH Advances Practical AI Applications Driving Transformation Across Specialized Healthcare
King Faisal Specialist Hospital & Research Centre (KFSH) will showcase its applied artificial‑intelligence portfolio at HLTH Europe 2026, highlighting a shift from pilot studies to production‑grade tools that accelerate diagnosis and streamline operations. Since 2019 its Centre for Healthcare Intelligence...
Deeply's Industrial Acoustic AI Solution Solves Connector Engagement Defects in Auto Assembly Lines, Expanding Application in Global Production Lines
Deeply Inc. announced that its Listen AI industrial acoustic AI solution reached 99.87% accuracy in detecting connector‑engagement defects on global auto‑assembly lines, including Korea and Mexico plants of a major automaker. The technology isolates subtle lock sounds amid 85 dB factory noise...
Microsoft Just Gave Investors 3 Dates They Can't Afford to Ignore
Microsoft’s Build 2026 unveiled a multi‑layered, agentic Copilot platform, but the announcement coincided with a near‑10% stock decline. While paid Microsoft 365 Copilot seats rose to 20 million, investors remain uneasy about whether the $190 billion FY 2026 capex will translate into sustainable revenue. The...
Apple Just Delivered Bad News for OpenAI and Anthropic, but Alphabet Could Be a Winner
Apple unveiled Siri AI at WWDC, a conversational chatbot built on Alphabet’s Gemini large‑language model. The service is embedded across iOS, iPadOS and macOS, giving Apple’s 2.5 billion active devices free, always‑on AI assistance without subscription fees. By providing a native...
Consortium Launches AI-Powered Retrofit Platform for Shipping
European consortium FIT‑HORIZONS, backed by €4 million (≈$4.4 million) EU funding, brings together 19 maritime firms, universities and technology providers to create an AI‑powered retrofit design environment. The platform will integrate simulation, machine‑learning and operational data to evaluate combinations of low‑ and...
Beyond Black Box Scores: How Musubi Trains Custom AI for Trust and Safety Teams
Musubi offers an AI‑native trust‑and‑safety toolkit that builds custom‑trained machine‑learning and LLM models for each content platform. The company showed how its prototype caught moderation issues that human reviewers missed and later packaged a policy‑optimizer that lets teams tweak rules...
Diagrid Brings Cryptographic Proof to AI Agent and Workflow Execution
Diagrid Inc. announced Dapr 1.18, an open‑source runtime update that introduces verifiable execution for AI agents and workflows. The release adds three capabilities—Workflow History Signing, Workflow History Propagation, and Workflow Attestation—allowing cryptographic proof of execution, custody, and tamper‑evidence. Additional enhancements include...
Exclusive: Tilebox Launches AI-Focused Update
Tilebox unveiled its 2.0 update to make AI agents transparent geospatial-data analysts, adding provenance tracking to AI‑generated insights. The middleware sits between large language models and Tilebox’s satellite‑image APIs, logging data sources and transformation steps. It supports Anthropic Claude Code, OpenAI...
4 Bets Accor Is Making on AI and the Future of Hospitality
Accor is betting on artificial intelligence across four pillars—brand discoverability, distribution orchestration, predictive loyalty, and self‑piloted operations. It has integrated its ALL Accor booking and loyalty app into ChatGPT and is partnering with OpenAI and Google Gemini to boost AI...
AWS’ Former CEO Launches New AI Infrastructure Startup Helix Digital, Backed By Nvidia And $10B In Capital
Former AWS chief Adam Selipsky has been named CEO of Helix Digital Infrastructure, a new AI‑focused startup created by KKR. Helix launches with over $10 billion in long‑duration capital commitments, backed by Nvidia, the Kuwait Investment Authority and Vistra. The firm...
Eve Builds on AI Workforce Launch with EveOS, An AI-Native Operational Platform for Plaintiff Firms
Eve, a legal‑AI specialist, unveiled EveOS, an AI‑native operating system that expands its platform with four new products—Atlas, Analyst, Communication Agents, and Research—while enhancing Auditor and Intake. Atlas creates a self‑updating case data layer that pulls information from case‑management tools,...
Nazaré Launches AI Academy
Nazaré, part of Inizio Engage, has launched AI Academy in partnership with employee‑experience firm Forty1 to help companies embed generative AI into daily work. The program, built on behavioral science and two decades of learning expertise, targets the gap between...
Microsoft Open Sources AI Evaluation Framework for Enterprise Agents
Microsoft has open‑sourced ASSERT, an Adaptive Spec‑driven Scoring framework that turns natural‑language requirements into executable AI evaluation tests. The MIT‑licensed tool auto‑generates scenarios, datasets, metrics and scorecards, aiming to plug the governance gap as enterprises scale agent deployments. Internal trials...
Chrome Warns WebMCP Can Hijack AI Agents
There will be growing pains. Serious growing pains -> WebMCP Can Be Used To Hijack AI Agents, Chrome Warns "According to Chrome’s agent security guidance, AI agents using WebMCP must defend against two primary attack vectors: malicious manifests and contaminated outputs." "A...
AI Should Empower, Not Replace, Healthcare Clinicians
Workforce survival in healthcare "The coming decade demands that we stop asking whether AI can replace clinicians and start asking how it can help us keep them." https://t.co/EkYGoRfxXj https://t.co/8Zs0G1Ssl7
White House, Senate Revive Push to Block State AI Laws Through Kids Safety Deal
The White House and Senate Republicans are reviving a push to preempt certain state AI regulations by bundling them with children‑online‑safety legislation, including the Kids Online Safety Act, the App Store Accountability Act, and the NO FAKES Act. The proposal,...
Leaked Gemini 3.5 Pro Details Reveal Why Google Is Falling Behind AI Rivals
A leaked review of Google’s Gemini 3.5 Pro shows the model falling short in advanced reasoning, code infilling, and long‑term task execution, lagging behind Anthropic’s Fable 5 and OpenAI’s GPT‑5.6. The upgrade focuses on better vision, multimodal understanding and stricter safety filters, which...
Production AI Agents: Closing the Gaps Between Idea and Reality by João Freitas
João Freitas details PagerDuty’s year‑long effort to move AI agents from prototype to production, exposing architectural hurdles like non‑determinism, context fatigue, and security exposure. He proposes a five‑pillar framework—Reliability, Control, Visibility, Integration, Economics—to guide robust, scalable agent design. The article...
India's SatSure Bags $2.6 Million Grant to Build AI-Powered Earth Observation Models
SatSure Analytics secured a 246 million‑rupee ($2.57 million) grant from India’s space regulator to build AI‑driven Earth‑observation models. The funding is part of a broader $104 million government initiative to accelerate private participation in the nation’s space sector. SatSure will develop large‑scale models...
Claude Fable 5 Trashed in Under 24 Hours
Anthropic unveiled Claude Fable 5 as a routine upgrade, but users quickly discovered restrictive access tiers, opaque data policies, and aggressive safety controls. Within 24 hours the release ignited a wave of criticism, with the tech community dubbing it an "AI...
‘Cisco Now Is Delivering that Critical Infrastructure for the AI Era’: Cisco's Infrastructure Unification Push Aims to Simplify Management for...
Cisco unveiled its Cloud Control platform at Cisco Live, promising a single pane of glass to manage disparate networking, security and cloud tools. The solution bundles Catalyst Center, Meraki Dashboard, Nexus One, Collab Command Center and Cisco Security Cloud Control,...
Easy Data Access Introduces DocAI: AI-Powered Document Management Add-On for Docsvault
Easy Data Access LLC has launched DocAI, an artificial‑intelligence add‑on for the Docsvault document‑management platform. The solution automatically extracts data and indexes metadata without requiring templates, training data, or a separate application. Existing Docsvault users can enable DocAI instantly, preserving...
AI May Boost Productivity but It's Also Causing Anxiety Among Canadians: Report
Canadian employers are rapidly adopting generative AI to close skills gaps and lift productivity, with 67% expecting efficiency gains and 63% viewing AI as essential for new hires. The technology is already delivering results: 24% cite higher productivity, 23% note...
Pinecone Nexus Now Integrates with Microsoft OneLake, Bringing AI Agents Directly to Enterprise Data
Pinecone announced that its Nexus knowledge engine now integrates directly with Microsoft OneLake, the unified data lake behind Fabric. The integration lets AI agents query OneLake through Nexus's KnowQL language, which builds task‑specific, permission‑scoped artifacts instead of pulling raw data...
HighPoint Unveils High-Performance Gen5 AI Compute Platform for Ultra-High Inference Throughput Empowered by Hailo
HighPoint Technologies and AI‑chip maker Hailo announced the Enterprise Edge AI Compute Platform, built around HighPoint's Rocket 1604L Gen5 retimer and up to four Hailo‑8 or Hailo‑10H M.2 acceleration modules. The solution packs 160 TOPS (INT4) of inference power into a single...
This AI “Brain” Wants To Get Rid Of The Grunt Work In Creative Campaigns
Innovid, now part of Mediaocean, launched NIVO AI, a conversational brain that automates the grunt work of creative campaign setup. The tool replaces spreadsheet‑driven asset tagging with natural‑language prompts, instantly generating and powering ads across channels. By linking predictive scoring,...
TestSprite Launches an Open-Source Command-Line Tool to Help AI Agents Check Their Own Work
TestSprite Inc. has open‑sourced its command‑line interface that lets autonomous AI coding agents verify their own output by running live browser or API tests and returning concrete failure data. The tool creates a continuous quality‑assurance loop, automatically expanding test coverage...
Clinch Launches Operational GEO Solution for the AI Discovery Era
Clinch introduced an operational Generative Engine Optimisation (GEO) solution, powered by Whitebox, that measures brand visibility across generative AI platforms such as ChatGPT, Gemini, Claude and Perplexity. Integrated into Clinch’s Flight Control, GEO feeds AI‑derived insights—sentiment, rankings, citation sources—directly into...
AI You Can Trust, Audit and Keep with Russell Moore, Co-Founder & CEO of Amotivv | Episode 494
In this episode, Russell Moore, co‑founder and CEO of Emotive, walks host Greg Myers through his career from network engineering to leading AI initiatives at Global Payments, culminating in the launch of Emotive. He explains Emotive’s three‑layer platform—memory, a governed...
Enterprise Software Is About to Get Personal
In 2025 a new wave of AI models and agentic infrastructure let developers orchestrate autonomous coding teams, moving beyond simple code‑suggestion tools. The article argues that retail, still shackled by fragmented SaaS stacks and legacy ERP systems, is poised for...
The Real Power Struggle in Agentic Commerce Isn’t Building the Smartest AI Agents; It’s Governing Them
Financial services firms are rapidly deploying AI agents that can browse websites, compare products, and complete purchases. While the technology race focuses on expanding capabilities, J.P. Morgan Payments’ Global Head of Merchant Services, Michael Lozanoff, warns that governance, trust, and...
Ricoh Launches AI Orchestration Co-Creation Initiative with Thread AI
Ricoh Company has partnered with Thread AI to launch an internal pilot that blends Thread’s AI orchestration infrastructure with Ricoh’s digital‑twin technology. The joint execution platform will ingest sensor, camera and operational data to automate facility‑management tasks such as inspections...
Why NVIDIA’s New ASR Model Is Beating Whisper in Live Transcription
Nvidia unveiled NeMoTron 3.5, a 600‑million‑parameter automatic speech‑recognition model that can be self‑hosted and transcribes live audio in 40 languages. The system adds streaming transcription, speaker diarization and a word‑boosting feature that lets users prioritize domain‑specific terms without retraining. Technical tricks...
The Sequence Opinion: Systems of Record Vs. Systems of Action
The post argues that agentic AI will not destroy SaaS but will redefine its purpose. Historically, enterprise software acted as a system of record, storing canonical data while humans performed actions. Now AI agents can safely act on that data,...
Agentic AI: What Leaders Wish They Knew Sooner
At the 2026 MIT Sloan CIO Symposium, senior technology and business leaders examined how humans and agentic AI collaborate in real workflows. Experts warned that human‑in‑the‑loop oversight is becoming a perfunctory checkbox, while many so‑called AI agents are over‑hyped and...
AI Robot Cleaners Leave the Lab for China's Living Rooms
Chinese household‑help platform 58.com has teamed with robotics firm X Square to offer an AI‑powered cleaning service using the Quanta X1 Pro robot. Priced at 149 yuan (≈ $22) for a three‑hour session, the service is available in Beijing and Shenzhen and has attracted about...
AI Vendor FDEs: Key Considerations and Concerns
The article examines the surge of forward‑deployed engineers (FDEs) from AI vendors, highlighting their ability to embed deep model expertise directly into enterprise teams. While FDEs can speed up AI rollouts, they also introduce vendor lock‑in, hidden long‑term costs, and...
Shutterstock, Prudential Financial Share AI Cost Strategies
CIOs are confronting soaring AI expenses as token consumption outpaces budgets, with global usage projected to reach 120 quadrillion tokens per month by 2030. Companies such as Uber and Shutterstock have already hit limits, prompting caps and deeper cost analysis. Shutterstock’s...
Nvidia Accelerates Google DeepMind’s DiffusionGemma
Google DeepMind unveiled DiffusionGemma, an experimental open‑source text generation model optimized for Nvidia GPUs. Built on the 26‑billion‑parameter Gemma 4 mixture‑of‑experts architecture, it can generate up to 256 tokens in parallel and fits within 18 GB of VRAM when quantized. Nvidia tailored...
What AI-Herding Scientists Can Learn From Watching ‘Sheepdog YouTube’
Scientists analyzed YouTube footage of sheepdog trials and identified a two‑step tactic—waiting for all sheep to align, then chasing—to steer tiny, noisy flocks. The insight inspired an "Indecisive Swarm Algorithm" that lets robots intermittently follow a central controller or neighboring...
ChatGPT Can Now Buy Things for You After Deal with Payments Giant Visa
Visa has embedded its payment network into ChatGPT, enabling the chatbot to shop and complete purchases on behalf of users. Customers link a Visa card, set spending limits and approval steps, and the AI can locate items—such as wireless headphones...
AI Board Decision-Making Is Becoming A Governance Test For Financial Leaders
UK boardrooms are rapidly confronting AI governance as 84% of directors say their boards have debated AI’s role in decision‑making, and 49% are already moving from discussion to implementation. The shift moves AI from back‑office efficiency tools to advisory "board...
B2B Publishers Call for Peers to Join AI Licensing ‘NATO for News’
The SPUR Coalition, dubbed a “NATO for news,” is expanding its AI licensing standards network, adding 30 new members—including its first B2B publishers such as Citywire, Times Higher Education, and AML Intelligence. Founded by major UK outlets like the FT,...