Today's AI Pulse
Salesforce to acquire AI‑agent platform Fin for $3.6B
Salesforce announced a $3.6 billion purchase of Fin, the AI‑agent platform formerly known as Intercom, to strengthen its Agentforce suite. The deal comes as Salesforce reported strong Q1 results and a $50 billion buyback program. Fin’s technology enables automated customer support across chat, email, WhatsApp, phone and Slack.
Also developing:
By the numbers: Waymo acquires Apple’s Arizona Proving Ground for $220M
Waitrose Deploys AI Route Optimisation to Improve Delivery Efficiency
Waitrose has teamed with AI specialist Satalia to deploy an AI‑powered route optimisation platform across its home‑delivery network. The system analyses traffic, vehicle capacity and time windows in real time, producing more efficient routes. Early data shows an 8 % reduction in delivery mileage, a 12 % increase in drops per route, and a four‑point rise in slot availability. The rollout now covers 170 stores, two fulfilment centres and roughly 1,400 vans handling 160,000 weekly deliveries.
What OpenAI’s $110 Billion Funding Round Says About the AI Bubble
OpenAI announced a $110 billion funding round in February 2026, valuing the company between $730 billion and $840 billion. Despite this massive capital influx, its 2025 revenue was only $20 billion, highlighting a stark valuation‑revenue gap. The round was led by major corporations and...
How Can Machine Learning Unlock Credit in South Africa
Access to credit remains elusive for most South African small businesses, with over 85% of firms earning under R1 million facing high rejection rates. A Forrester‑Experian study shows that 70% of senior lenders believe machine learning can improve risk accuracy and...
How Developers Can Bring Voice AI Into Telephony Applications
Voice AI is moving from chatbots into telephony, promising to automate appointment scheduling, call triage and delivery coordination. Developers, however, must bridge sophisticated AI pipelines with legacy PSTN, SIP and WebRTC networks, a task complicated by evolving large language models...
Huawei: Helping Miners Build a Sustainable Digital and Intelligent Foundation
Huawei’s Mining Business Unit is positioning the company as the leading provider of AI‑enabled connectivity for the mining sector. It highlights the deployment of a 5G‑A network and the world’s first fleet of 100 autonomous electric trucks at China’s Yimin...
The Governance of Agentic AI: From Human-in-the-Loop to Logic-Based Oversight
Agentic AI systems now execute payments, liquidity moves and trades without human clicks, forcing financial institutions to move governance from transaction‑level monitoring to logic‑level auditing. The article argues that human‑in‑the‑loop oversight has become theatrical, as agents optimise across thousands of...
Mandated AI Use Creates Friction and Inefficiency
"The pressure to use AI 'created more friction as it became obvious that we were being monitored for our usage, and we were being told that we had to use it', even if it was the less efficient route to...
NiCE Launches Agentic AI Innovation
NiCE unveiled an agentic AI platform that converts structured and unstructured enterprise interaction data into ready‑to‑deploy AI agents at scale. The system scans voice, chat, digital channels and workflows, pinpoints high‑impact automation opportunities, and quantifies projected ROI before launch. It...
AI Infrastructure Now Worth As Much As Models
The AI infrastructure race is accelerating. UK startup Nscale just raised $2 billion at a $14.6 billion valuation, with @Nvidia backing its push to build large-scale AI data centers and compute platforms. In the AI era, owning the infrastructure may be just...
Hong Kong’s Financial Firms Adapt to AI’s Rewards and Risks
Hong Kong’s financial institutions are accelerating AI adoption, moving beyond simple automation to reshape deal marketing, client relationship management, and trade execution. The shift mirrors a broader global AI arms race that began on Wall Street and now targets Asia’s...
From Influencer Matching to an AI-Driven Operating System - Inside Humanz' Bet on Agentic to Scale Creator-Led Performance Ads
Humanz has launched Humanz+, an AI‑native operating system that automates the full lifecycle of creator‑led advertising, from research and brief creation to media buying and conversion optimization. The platform deploys seven specialized AI agents and still incorporates a human‑in‑the‑loop for...
Digest: Retail and Grocery Brands Dominate ChatGPT Ads Trial; Walmart to Phase Out Vizio Accounts; New US AI Guidelines Challenge...
OpenAI’s month‑long ChatGPT ad trial is now 44% retail and grocery ads, led by brands like Target, Sephora and Wayfair. Walmart announced it will phase out standalone Vizio accounts, requiring new TV buyers to link or merge into a Walmart...
Strategic Cooperation on AI
The March 10 2026 report outlines a framework for strategic cooperation on artificial intelligence, identifying three overarching objectives: improving understanding of AI capabilities and risks, fostering reliable development while curbing proliferation, and mitigating AI‑related harms. It proposes four core functions—research, standard‑setting, monitoring,...
Render Networks Unveils Synchronised Agentic Critical Infrastructure Architecture
Render Networks announced its ClearWay platform, a synchronized agentic architecture for critical infrastructure. The system converts design data into live work scopes, using autonomous agents with identity and least‑privilege controls to validate field progress and approve work in real time....
AI Displaces High‑Earners, Not Low‑Wage Workers
The People Moat: Why "Toil" Is Dead I want to challenge something people keep getting wrong about AI and jobs. It's not coming for truck drivers and call centre reps. Well, it is—but that's not the shock. The shock is *who's actually at...
NVIDIA Investing $2bn in Coherent’s R&D, Capacity Expansion and Operations as It Builds Out US-Based Manufacturing
NVIDIA announced a multi‑year strategic agreement with Coherent Corp, investing $2 billion to fund research, capacity expansion, and US‑based manufacturing. The deal includes a non‑exclusive purchase commitment and future access rights to Coherent’s advanced laser and optical networking products. By scaling...
AI Chatbots Are Sneakily Directing Users to Illegal Online Casinos
An investigation by The Guardian found that several popular AI chatbots readily recommend illegal offshore online casinos when prompted. The bots not only name the sites but also compare sign‑up bonuses, highlight fast withdrawal methods and explain how to access...
HY 3D Advanced Features Now Available in ComfyUI
HY 3D’s advanced processing suite is now packaged as partner nodes in ComfyUI, adding three core capabilities—3D Parts Decomposition, UV Unwrapping, and Smart Topology. The update lets creators split meshes, generate texture‑ready UV layouts, and simplify dense geometry without leaving...
Civil Liberties Group to Intervene in Ontario Court of Appeal Case Involving Facial Recognition Tech
The Ontario Court of Appeal granted the Canadian Civil Liberties Association (CCLA) permission to intervene in R. v. Kawall, a criminal appeal concerning the use of facial recognition technology (FRT) and a photograph taken in a public place. The trial...
How AI Could Transform Events and Exhibitions
At the Definitive AI Forum, senior executives from Terrapinn, Nineteen Group, Emap and RX Global discussed how artificial intelligence can augment, rather than replace, face‑to‑face events. They highlighted early experiments such as QR‑code data capture, AI‑driven meeting recommendations, live transcription...
Deploy Real AI Agents, Not Just Talk, to Win
𝗧𝗵𝗲 𝗿𝗲𝗮𝗹𝗶𝘁𝘆 𝗼𝗳 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗿𝗶𝗴𝗵𝘁 𝗻𝗼𝘄. → Many people are talking about AI agents. Only a few are actually building systems that generate real value. → Most “AI agents” today are still assistants or simple workflows with a new label. That gap...
When AI Safety Constrains Defenders More than Attackers
Security teams are encountering AI safety guardrails that block legitimate defensive tasks such as threat modeling, phishing simulations, and red‑team code generation. Research shows open‑weight models can be bypassed in multi‑turn attacks with success rates up to 93%, while enterprise‑approved...
Mastercard Brings Agentic Payments to Life in Singapore with DBS and UOB
Mastercard completed its first live, authenticated agentic payment in Singapore, partnering with DBS and UOB. An AI‑driven agent used Mastercard Agent Pay to book a hoppa ride to Changi Airport, leveraging tokenised credentials and Payment Passkey consent. The transaction moves...
5 Ways to Automate Tactiq with Zapier
Tactiq’s AI‑powered transcription now pairs with Zapier to automate every step after a meeting. Users can automatically back up transcripts to Google Drive, Dropbox, or Notion, generate structured Google Docs and AI‑summaries, and turn key points into tasks in Asana,...
President of Microsoft Science Saw ChatGPT Coming (and Now He Predicts How It Will Change Healthcare) | Peter Lee
In this episode, Peter Lee, President of Microsoft Research, recounts the evolution of AI from early neural networks to today’s large language models, describing how Microsoft recognized OpenAI’s potential early on and invested heavily despite industry skepticism. He explains the...
‘Work Is Broken’: Can Agentic AI Fix It?
Enterprise workflows remain fragmented, prompting CIOs to label "work is broken." Agentic AI is touted as a cure, yet Gartner warns up to 40% of projects may be cancelled by 2027 due to governance, cost and execution hurdles. Early pilots...
AI Agent That Finds Jobs & Optimizes Your Resume 24x7
A DIY AI agent built with n8n now scans LinkedIn each morning, pulls the five newest job listings, rewrites a master resume with ATS‑friendly keywords for each role, converts the results to PDFs, and emails them to the user before...
AI Pitch Bot Offers Cheap, Consistent Feedback, But Limits Hurt
haha the @runwayml real-time characters is great. i just spent 2 minutes pitching a ghibli version of myself a bladeless blender startup it's only a 2 min demo, but the questions and tone are pretty solid. one question even threw me...
Autonomy in Defense 2026: Operationalizing AI & Autonomy at Scale
The Autonomy in Defense 2026 summit in Washington, D.C., will convene senior defense leaders, technologists, and policymakers to discuss moving AI and autonomous systems from isolated prototypes to enterprise‑wide operations. Speakers highlighted that the defense sector’s primary obstacle is execution,...
ByteDance Releases DeerFlow 2.0: An Open-Source SuperAgent Harness that Orchestrates Sub-Agents, Memory, and Sandboxes to Do Complex Tasks
ByteDance has open‑sourced DeerFlow 2.0, a SuperAgent framework that runs tasks inside isolated Docker containers rather than merely generating text. The system decomposes complex prompts into parallel sub‑agents, each operating in its own sandbox, and then converges results into a...
The Future Is AI First but Human Centered with Sam Wilson
Sam Wilson, CEO of 8×8, argues that the future of customer service is AI‑first but fundamentally human‑centered. He stresses that self‑service tools must include an effortless path to a live agent, especially as California now requires a clear human‑option. The...
Huawei: Agent-Oriented Mobile Networks to Define AgentVerse
Huawei unveiled the AgentVerse at MWC 2026, a vision that extends the 5G core with AI‑driven agents and predicts a 10,000‑fold rise in agent‑handled work by 2030. The initiative builds on the Agentic Core, which integrates real‑time experience awareness, global resource...
Talk to Your Data | T-Systems Unveils AI Agent to Accelerate Analysis Work
T‑Systems launched “Talk to your data,” an AI‑driven chat platform that connects disparate corporate data sources, searches, analyzes, and visualizes information on demand. The solution uses an ontology layer to map data across systems, enabling natural‑language queries. Pilot projects, including...
Experience the Sights and Sounds of the HIMSS26 Preconference
At HIMSS26, the pre‑conference highlighted how healthcare organizations are deploying digital technologies and artificial intelligence to transform care delivery. Sessions showcased AI‑driven diagnostics, telehealth expansion, and data‑interoperability initiatives across hospitals and health systems. Demonstrations illustrated real‑world use cases, from predictive...
Insurify’s Founders Discuss Evolution of Insurance Shopping With AI
Insurify launched the first AI‑powered insurance app on OpenAI’s directory, prompting a brief dip in the S&P 500 Insurance index. The founders emphasized the rollout as an evolutionary step rather than a disruptive threat, noting that the app currently provides estimated...
Building GPT‑OSS‑20B Inference From Scratch in PyTorch
GPT-OSS Inference from Scratch In this post, they implement GPT-OSS-20B inference from scratch in PyTorch. https://blog.yellowday.day/posts/gpt_oss_from_scratch/
Paris-Based Qevlar AI Raises €25.8 Million to Automate Security Operations Centre Investigations with Agentic AI
Paris‑based Qevlar AI announced a €25.8 million funding round, co‑led by Partech and Forgepoint Capital with participation from EQT Ventures. The capital will expand its agentic AI platform that autonomously investigates security alerts, delivering full‑depth analysis and actionable insights. Qevlar claims...
Re(AI)magining Melbourne: Persistent Accelerates Australia’s Enterprise AI Momentum
Persistent has opened a Melbourne Innovation Centre to accelerate enterprise AI adoption in Australia. The centre combines AI‑led engineering, cloud‑native modernization, and domain‑specific solutions to move pilots into revenue‑generating products. Australian firms currently fail to scale 80% of AI pilots,...
Yann LeCun Raises $1 Billion to Build AI That Understands the Physical World
Yann LeCun, former Meta chief AI scientist, launched Advanced Machine Intelligence (AMI) with a $1 billion funding round, valuing the Paris‑based startup at $3.5 billion. AMI aims to develop AI world models that understand and reason about the physical world, offering persistent...
Yann LeCun’s AMI Labs Raises $1.03B to Build World Models
Yann LeCun’s new venture AMI Labs secured $1.03 billion in funding, valuing the startup at $3.5 billion pre‑money. The company is developing "world models"—AI systems that learn from real‑world data using the Joint Embedding Predictive Architecture (JEPA). Its first commercial partner is...
NotebookLM's Video AI Crafts Witty Sauron Strategy
NotebookLM: Do a deep research report and make a video where a consultant gives Sauron a strategy for actually winning the War of the Ring: "All you need to do is sign off to put a simple door on...
AI Finds 10 Winning Products in Just 20 Minutes
🤯 Found 10 winning products in under 20 mins 🔥 AI analysed trends and margins 🛠️ Full report on each product 🤵 Supplier links included 📦 Ad angles ready to test This is crushing hours of manual research Want the method? RT + comment 'PRODUCT AI'
Funding Surge Fuels Diverse AI Paths Beyond LLMs
There are now over a half dozen extremely well-funded companies from famous AI researchers building alternative approaches to AI, betting LLM-based technologies hit a wall. The overall effect is that there are now more pathways than ever for keeping AI...
Eight-Person Team Beats Google, OpenAI
Holy shit... An 8-person team just beat Google, Perplexity, Claude, and GPT-5.2 at deep research. Not a lab. Not a billion-dollar company. 8 people. 5 engineers. It's called Spine Swarms. Here's everything you need to know👇 https://t.co/T9sAd9u19y
AI Boosts MedTech Efficiency and Diagnostic Accuracy
Artificial intelligence is reshaping the landscape of MedTech and life sciences by streamlining processes and enhancing diagnostic accuracy. Discover how these innovations are making a significant impact. https://t.co/hZw9hI2XG0
AI Adoption Soars, Yet Productivity Gains Remain Elusive
Check out the latest article in my newsletter: The AI Productivity Mirage: AI Adoption Is Surging. Why Are Productivity Gains Lagging? https://t.co/IaRyphLJJf via @LinkedIn
End‑to‑end Lab Validation: Essential for ML Biology
"Valodated end to end in the lab" is music to my ears. THIS is how you do ML for biology. Show validation or gtfo.
Closed AI Models Can't Run Locally on Personal GPUs
kinda a dumb question sorry but theres absolutely no way to run closed models on my own GPU right
Broadcom AI Revenue Spikes, Beating Q1 Expectations
.@Broadcom AI revenue surges in fiscal first quarter https://t.co/7805cMiu5B Broadcom reported better-than-expected fiscal first quarter results due to strong demand for AI semiconductors. https://t.co/v8Wqlit2d2
Microsoft to Prioritize AI Token Efficiency Next Year
.@Microsoft CEO Nadella: AI efficiency drive on deck https://t.co/BJpfkIlbxC Microsoft CEO Satya Nadella said the next year will bring software and applications designed to make AI token usage more efficient. https://t.co/qDupPqRCeN