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
AI Chipmaker Cerebras Namedropped by Oracle, Alongside Nvidia and AMD
Oracle’s CEO Clay Magouyrk highlighted Cerebras alongside Nvidia and AMD on the company’s earnings call, signaling the AI chipmaker’s entry into Oracle’s cloud infrastructure. The endorsement follows Cerebras’s recent $1.1 billion funding round and its push toward an IPO after withdrawing a prior filing. OpenAI has also begun running its Codex‑Spark model on Cerebras’s WSE‑3 chips, expanding the startup’s customer base beyond its dominant Middle‑East client G42, which previously accounted for 87 % of revenue. The Oracle mention could diversify Cerebras’s revenue stream and accelerate its market positioning against established GPU leaders.
AI Code Generation Will Render Spreadsheets Obsolete
prediction re the end of spreadsheets AI code gen means that anything that is currently modeled as a spreadsheet is better modeled in code. You get all the advantages of software - libraries, open source, AI, all the complexity and expressiveness. think...
Agentic AI: Your Always‑On Teammate for Unified Ops
Agentic AI isn't about replacing analysts or SREs. It's about giving them an always-on teammate that closes gaps between monitoring, security, and data governance so teams can focus on higher-order risks. #SRE #ITOps #AI https://t.co/e3w3lXkvfc
The Catastrophic Threat of Unmanaged AI Drift
The article warns that unmanaged AI drift—gradual divergence of models from their original objectives—poses a catastrophic risk. It traces the concern back to Norbert Wiener's 1960 essay, emphasizing that automated systems follow literal instructions, not human intent. As AI systems...
AI Chatbots Poised to Replace Call Center Jobs
#WhosNext? Call center workers? #Chabots and new AI inventions could make call centers a thing of the past. (Bloomberg) #Automation #AI https://t.co/50nCCEZ6Lz
From Curiosity to Mastery: Companies Seek Full AI Toolkit
Hearing this from startups and Fortune 500s: "How do I use all of Claude's features?" A few months ago, nobody asked. AI became a workflow while people were still deciding if it was real.
U of T to Partner with India on Health AI
The University of Toronto and the Indian Institute of Science announced a joint centre of excellence to develop AI‑driven predictive healthcare tools. The partnership, unveiled in Mumbai, is part of a broader Canada‑India Talent and Innovation Strategy and includes a...
Unrestricted AI Agents Are Like Free Nuclear Missiles
I agree with this. I feel like with the way we're deploying agents now, with zero safeguards, in a year or two it's going to be like giving everyone in the world a free nuclear missile.
China’s AI Dominance Threatens U.S. Global Leadership
The FUTURE of the world if the USA lets China win the AI & robotics race, according to @ericschmidt https://t.co/kdfqJjJk8W
MEDITECH Enhances Ambient Intelligence Solutions
MEDITECH announced at HIMSS26 the expansion of its Expanse AI portfolio with native ambient intelligence integrated into the Expanse Now mobile app for physicians and Expanse Point of Care app for nurses. The solution captures real‑time conversations, automatically generates visit...
BBVA Boosts Audit Productivity with AI Assistant
BBVA has integrated ChatGPT Enterprise across its 120,000‑employee workforce, including a custom GPT for its internal audit function. The AI assistant automates data‑analysis tasks, reducing manual programming and documentation while maintaining auditor oversight. BBVA estimates a 10 % productivity gain for...
Google Finds that AI Agents Learn to Cooperate when Trained Against Unpredictable Opponents
Google’s Paradigms of Intelligence team demonstrated that training large‑language‑model agents with decentralized reinforcement learning against a mixed pool of static and learning opponents produces cooperative behavior without hard‑coded rules. The agents use in‑context learning to infer co‑player strategies in real...
Scaling Enterprise Federated AI with Flower and Open Cluster Management
The post explains how Flower, an open‑source federated AI framework, pairs with Open Cluster Management (OCM) to deliver a production‑ready, enterprise‑scale solution for privacy‑preserving machine learning. Flower’s SuperLink/SuperNode architecture abstracts networking, letting developers focus on model code, while OCM’s hub‑spoke,...
Lobster Fever Spreads Among Agentic AI Users in China-
OpenClaw, an agentic AI platform, has gone viral among Chinese developers, sparking a surge in interest for autonomous AI agents. The platform’s ease of integration and open‑source model have accelerated adoption across startups and enterprises. Tech giants such as Tencent...
AI Can ‘Same-Ify’ Human Expression — Can some Brains Resist Its Pull?
Emerging research indicates that large‑language models are reshaping human expression, with post‑ChatGPT texts showing reduced stylistic diversity. Studies of Reddit, news, and pre‑print content reveal that users adopt LLM writing patterns, reasoning styles, and even opinions after AI assistance. Experiments...
Weak AGI Already Passed Turing, Winograd, SAT—Only Atari Remains
Lets look at the criteria for "weak AGI": ✅Loebner prize was a weak Turing Test, equivalent achieved by GPT-4.5 ✅Winograd passed by GPT-3 ✅SAT passed at 75% by GPT-4 Only remaining thing is playing an old Atari game from 1984. The...
General Social Agents
Researchers Manning and Horton propose AI agents that apply social science theories to novel settings without extensive modifications. By feeding theory‑grounded natural language instructions and limited seed‑game data, the agents simulate human behavior across a heterogeneous population of 883,320 generated...
NightBeacon: AI‑Powered, Self‑Trained Security Platform Launches
Here's a demo on a project I've been developing and working on for the past 9 months. Called NightBeacon. Using it now in production, getting released fully this week. Our own internally trained models on our own infrastructure (no third party)....
AIOps Is so Powerful, Vendors Are Building Tools to Clean up After Agents Break Your Infrastructure
Cohesity, together with ServiceNow and Datadog, is launching a recoverability service that can detect and roll back damage caused by agentic AI in enterprise environments. The solution leverages immutable snapshots and API‑driven restorations to return files, databases, vector stores, and...
The AI Lab Next Door
Local colleges and universities across the U.S. are rapidly building AI capabilities, yet their expertise remains largely untapped by city governments and nonprofits. While higher‑education institutions are governing AI internally and creating a pipeline of AI‑trained talent, public sector entities...
Singapore: Building a Smart, AI-Enabled Future for Healthcare
Singapore is building a next‑generation, AI‑enabled Eastern General Hospital, scheduled to open around 2029. The smart hospital roadmap integrates digital platforms, AI, robotics, wearables, and real‑time location systems to enhance clinical and operational efficiency. Telehealth and generative AI tools aim...
Oracle's Larry Ellison Downplays Software Apocalypse Fears: 'We Think the SaaSpocalypse Applies to Others, but Not to Us'
Oracle’s leadership used its earnings call to dismiss the so‑called SaaSpocalypse, arguing that AI‑driven disruption will spare the company. Larry Ellison said Oracle’s AI‑enabled coding tools let it build end‑to‑end, agent‑based solutions for sectors like healthcare and finance. CEO Mike...
One‑click Skill Catalog Lets Agents Gain New Abilities Instantly
Magister sneak peek #2: Skill catalog One-click install for agent skills. No terminal, no config files, no copy-pasting commands. Browse the catalog, hit install, done. Your agent immediately gets new capabilities — web search, browser control, frontend design, long-term memory, workflow automation,...
Why 2026 Will Be the Year of Governed Cybersecurity AI
The 2025 Cost of a Data Breach Report shows a 9% drop to $4.44 million, driven largely by security AI and automation that slash detection times. Yet organizations that fully automate see breach costs up to $1.9 million lower than manual peers,...
Google’s Genie 3 Draws a Crowd at GDC
Google DeepMind unveiled its latest generative AI model, Genie 3, at the Game Developers Conference in San Francisco. The presentation, “The Future of Playable Worlds With Google DeepMind,” filled the Moscone West room to standing‑room‑only capacity, with more than 100 developers turned...
AI's Promise Mirrors 1950s Nuclear Duality
In the 1950s, you could simultaneously think A) nuclear energy was good, and B) we were in great danger of a nuclear war. In the 2020s, you can simultaneously think A) AI agents will be a net good for the...
Cisco Survey: Infrastructure Gaps Stall AI Adoption in Manufacturing
Cisco’s 2026 State of Industrial AI report, based on 350+ manufacturing leaders across 19 countries, reveals that infrastructure shortfalls are throttling AI rollout in factories. While 96% view reliable wireless connectivity as essential, 56% cite frequent outages that disrupt operations....
Warburg-Backed PDG to Raise $5 Billion in Debt for Data Centers
Singapore‑based Princeton Digital Group, backed by Warburg Pincus, announced plans to raise up to $5 billion in debt this year. The proceeds will finance a multi‑country AI data‑center build‑out in India, Indonesia, Malaysia and Japan. PDG already operates AI‑focused facilities in seven...
The Great Decoupling: How Professional Services Should Rethink Pricing in the Age of AI
Professional services firms are confronting AI‑driven efficiency that compresses delivery time, prompting a shift from hourly billing to outcome‑focused pricing. INSIGHT proposes a five‑level pricing maturity framework that guides firms from role‑based rate cards to fully value‑based models. The roadmap...
Huawei Outlines Practical Route to “Industrial Intelligence” At MWC 2026
At MWC Barcelona 2026, Huawei announced a shift from AI experimentation to large‑scale operational use in enterprises. It showcased 115 industrial intelligence demonstrations and introduced 22 new industry solutions across energy, manufacturing, finance, transport and retail. Central to the rollout...
INDOPACOM Was All in on Anthropic. Now It’s Working to Adjust
U.S. Indo‑Pacific Command (INDOPACOM) built its war‑gaming and command‑control workflows around Anthropic’s Claude model, only to lose access after a Trump‑issued ban on Anthropic tools for federal agencies. The ban sparked a lawsuit from Anthropic alleging illegal retaliation against the...
AI Isn’t Good at UX
A developer used AI to generate Lightning Web Components, achieving near‑perfect code for functional requirements. When asked to redesign the UI for tablet use, the AI repeatedly produced unusable layouts despite multiple prompts. Even feeding AI‑generated design advice into another...
Ford Launches New AI to Grow Multibillion-Dollar Pro Commercial Business
Ford Motor announced the launch of Ford Pro AI, an artificial‑intelligence platform for its Pro commercial‑vehicle subscribers. The system will analyze more than one billion data points each day, covering vehicle health, driver behavior, route efficiency and fuel use. It targets...
Ford Is Giving Its Commercial Fleet Business an AI Makeover
Ford unveiled Ford Pro AI, a generative‑AI chatbot embedded in its Pro telematics platform, to turn fleet telemetry into actionable insights for its 840,000 paid subscribers. The model‑agnostic service runs on Google Cloud, operates in read‑only mode and still requires human oversight,...
Salesforce Pushes ‘Agentic Enterprise’ Vision with Telecom AI Agents and New AI Work Metric
Salesforce unveiled new Agentforce capabilities, including Agentforce for Communications—a suite of pre‑built AI agents tailored for telecom sales and customer‑service workflows. It also introduced the Agentic Work Unit (AWU) metric, which counts each discrete AI‑driven task, reporting 2.4 billion AWUs to...
Pragmatic AI: Solving Real Problems in a World of Hype
The article warns that AI is currently perched on the Gartner Hype Cycle’s Peak of Inflated Expectations, where lofty promises outpace real capabilities. It highlights how over‑hyped features—especially language‑specific nuances—often miss the mark, likening them to the short‑lived teardrop phones...
Get AI and Google Visibility without a Wikipedia Page
Wikipedia is included in virtually every major LLM training dataset and I built something to help you set it up. When ChatGPT, Claude, Gemini, or Perplexity answer "Who is an expert in ___?" — Wikipedia and Wikidata are baked into how...
Agents Can Self‑register APIs Using Email as Identity
when I combined AgentMail w BrowserUse, my agent was able to create an account for an API service and grab an API key for itself email isn’t just about sending and receiving emails, it’s identity, auth, notifications, and more
A New Wave of Disrupters Takes on American Health Care
Silicon Valley’s newest wave of health‑tech startups is targeting America’s $5 trillion, fragmented medical system. Patients increasingly skip or delay care because of cost, with roughly one‑third reporting financial barriers. These disrupters are leveraging artificial intelligence to streamline diagnosis, pricing transparency,...
Anthropic AI Opens Sydney Office, Expands APAC Presence
🗞️@AnthropicAI is expanding to Australia 🇦🇺and New Zealand🇳🇿, opening a Sydney office in the coming weeks—its fourth APAC office alongside Tokyo🇯🇵, Bengaluru🇮🇳, and Seoul🇰🇷—to support growing demand and better serve the region’s AI ecosystem. https://t.co/DGe5jbueVm
Invisible AI Delivers ROI without Workflow Disruption
The best AI strategy? One your users don't even notice. If you want immediate ROI, "slipstream" AI behind the scenes. If they have to change their workflow to find the value, you've already lost them. 🤫 Details here: https://t.co/KP9ClquPul #CIO #AI #Productivity...
FDA Advisor Touts Approach Tailoring Regulation To Specific AI Use
The FDA’s artificial‑intelligence advisor announced a shift toward evaluating AI‑driven health technologies on a case‑by‑case basis, focusing on each tool’s specific function rather than applying a one‑size‑fits‑all rule set. This risk‑based approach will tailor regulatory requirements to the actual clinical...
Rolling Stone Revisits BabyAGI, Sparking Human‑machine Debate
.@rollingstone remembers babyagi 🥰 (in an article questioning the role of machines vs humans from yesterday) https://t.co/H71nwLSBCv
Ask Anything: Notion AI Team Joins Thursday Podcast
We're having the Notion AI team (including at long last @simonlast) on the pod Thursday. send me all your questions on this + Notion AI! not an ad, just a fan. Notion is probably the most impt knowledge work agent lab in...
AI Assistants May Shift Developers Toward Preferred Languages
AI coding assistants may influence which languages developers use https://t.co/NUSmA6RF0d < seems possible, even likely. If you're polyglot, do you find that you favor one specific language for AI-generated code?
Cut Email Length in Half with Claude
Free professional advice: If your email is longer than 100 words, paste the draft into Claude, ask it to reduce the word count by 50% without abridging substantive content, review it, and then send it.
UK Seeks Experts to Craft AI-Driven Energy Policy
UK government issues open call for experts to shape AI energy policy #energysky -- via pv magazine global: https://t.co/hQ1EUBbdg9
AI Adoption Without Purpose Boosts Productivity, Not Impact
“Some orgs. are deploying AI w/out a clear sense of the benefits it will bring—b/c they’re driven by herd mentality. They may end up w/ employees who are more productive yet their organizational contribution is unchanged.” 🔗 https://t.co/5j1dKP8sIs #leadership #management #HR https://t.co/48qplP9WO3
Knowledge Workers Become Machinists as Agents Automate Craft
One way to think about this is: Agents are machine tools for knowledge work, and artisans are now shifting to being machinists, just like in factories a century ago.
SIMA 2: AI‑powered Explorer for Virtual 3D Worlds
Meet SIMA 2 — The Ultimate #AI Explorer for Virtual #3D Worlds by @GoogleDeepMind #ArtificialIntelligence #MachineLearning #ML #Technology https://t.co/5SkSVsD8NG