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
Do You Really Need Claude Code If You Already Have Cowork?
The post examines Anthropic’s Claude Code and Claude Cowork, two AI agents that share the same underlying model but differ in interface and target user. Claude Code runs in terminals, VS Code and other developer environments, offering direct system access, while Cowork lives inside the Claude desktop app with a visual workflow and sandboxed permissions for non‑technical users. The author tests both tools on three real‑world tasks—data cleaning, slide‑deck creation, and competitor analysis—to compare speed, output quality, and ease of use. A decision framework is offered to help users choose the right tool for their workflow.
CopilotKit's $27M AG-UI Solves a Trivial Sync Issue
CopilotKit raised $27M for a protocol called AG-UI — keeps agent state and frontend in sync when the agent is streaming, calling tools, or waiting for input. My first reaction: that's not a hard problem. I've written this glue code.
IMF Warns of 'Inevitable' AI-Powered Threats to World
The International Monetary Fund issued a warning that AI‑driven cyber‑attacks pose a systemic threat to global financial stability. Its latest report says extreme AI‑enabled incidents could strain funding, jeopardize solvency and disrupt markets. The IMF highlights that advanced models dramatically...
New Reality Defender Ethics Committee Not Mere Theater, Says CEO
Reality Defender CEO Ben Coleman announced a new Ethics Committee composed of former Google privacy chief Keith Enright, Yale Digital Ethics Center director Luciano Floridi, and former Twitter trust‑and‑safety head Yoel Roth. The group will address the growing “verifier’s power”...
Build Unready Solutions, AI Advances Will Validate Them
anthropic alpha from yesterday: "you should be building things that don't even work yet." their idea is simple: claude gets meaningfully smarter every few months. and every time it does, ideas that used to fail suddenly work. but the people who capitalized on...
AI for Science and Autonomous Labs to Come Together at SciFM 26
The SciFM 26 conference, held May 27‑29 at the University of Chicago, will explore how foundational AI models can move from digital analysis to controlling physical laboratory workflows. Organizers highlight the Department of Energy’s $320 million investment in AI‑for‑science projects and...
Mobai Face Biometrics, Liveness Selected for Norway’s Public Sector Digital ID
Mobai, in partnership with Commfides Norge, won Norway’s Digdir tender to supply face‑biometrics and liveness detection for the national digital ID platform, MinID. The Digital ID‑kontroll solution will verify identity documents and capture selfie biometrics, moving MinID toward the higher‑assurance...
Wendy's Beat Taco Bell at Customer-Facing AI. The Decision that Decided It.
Wendy’s FreshAI now completes 86% of drive‑thru orders without human input, achieving roughly 99% accuracy once built‑in escalation is applied. By contrast, Taco Bell’s voice‑AI plateaued at 83% accuracy, with staff stepping in on 62% of AI errors, and McDonald’s...
Snowflake Startup Spotlight: TestGorilla
TestGorilla, a global talent discovery platform, leverages more than 2 million skills‑tested candidates and over 350 science‑backed assessments to enable AI‑driven, skills‑based hiring. The company centralizes all recruitment, finance and CRM data in Snowflake, reporting up to ten‑fold faster query performance...
Worries About AI’s Risks to Humanity Loom Over the Trial Pitting Musk Against OpenAI’s Leaders
Elon Musk has sued OpenAI, claiming the company violated its original nonprofit charter, while OpenAI’s leadership counters that Musk is trying to undermine the firm for his own AI ventures. The federal trial in Oakland has become a proxy battle...
GPT‑5.5 Instant Boosts M365 Copilot Speed and Accuracy
Great to bring GPT 5.5 Instant to M365 Copilot today. With quicker, clearer, and more accurate responses, you can get to useful answers with less back and forth. Also rolling out to Copilot Studio and Foundry. All part of our focus on...
Greece, Birthplace of Democracy, Seeks to Put Humanity Ahead of AI in Updated Constitution
Greece is preparing a sweeping constitutional revision that would legally require artificial intelligence to serve individual freedom and societal prosperity. Prime Minister Kyriakos Mitsotakis presented the amendment as a safeguard against AI‑related risks to democracy and future generations. The draft...
$700 Billion in Capex. $50 Billion in Revenue. AI’s Math Is Broken.
Anthropic has amassed $72 billion in funding and now claims $30 billion in annualized revenue, yet projects a $14 billion loss for 2026 and no free cash flow until at least 2028. OpenAI mirrors this pattern, posting a 48% gross margin on inference...
ContinuumGlobal Launches Campaign Optimizer to Improve Messaging Performance
ContinuumGlobal introduced Campaign Optimizer, an AI‑driven add‑on to its Smart Marketing Engine that promises to boost email engagement by over 15 %. The tool lets marketers refresh existing copy or draw inspiration from other brands, automatically generating brand‑compliant, production‑ready HTML in...
How AI Shapes the Way IBM's Legal Chief Approaches Almost Everything
IBM appointed Anne Robinson as general counsel in July 2024 and has since made artificial intelligence a cornerstone of the legal function. She is deploying IBM Watson and custom large‑language‑model tools to automate contract review, predict litigation outcomes, and streamline...
Hermes Agent Guide: You Will Never Need to Repeat Context
Hermes Agent is an open‑source wrapper that gives large language models persistent memory, tool access, and the ability to create reusable skills. Unlike typical chat apps, it can run scheduled tasks, remember past interactions, and improve over time without retraining...
Hallucinations, Inaccuracies, And The Erosion Of The Rule Of Law
Sullivan & Cromwell, a leading global law firm, recently filed a brief that contained 36 AI‑generated citation errors across three pages, including fabricated case passages. The firm blamed non‑compliance with its internal AI policies, showing that existing training and safeguards...
What Does It Actually Take to Get Your Hotel on ChatGPT? - By Paige Lopez
Hotels are now racing to secure visibility inside AI‑driven travel assistants like ChatGPT, where conversational search can drive bookings. Lighthouse’s Connect AI platform creates an “Anchor” that automatically pulls a property’s rates, photos and descriptions from websites, OTAs and reviews, then...
Disney Bets on AI -- and the Backlash Against It
Disney announced a $1 billion investment in OpenAI to license its characters for the AI video app Sora, positioning generative AI as a long‑term strategic opportunity. However, OpenAI abruptly shut down Sora in March, prompting Disney to suspend the planned investment....
AI Summit London Unveils 10th Anniversary Speaker Line-Up Featuring Global Leaders in Enterprise AI
The AI Summit London announced its 10th‑anniversary speaker lineup for the June 10‑11, 2026 event, featuring senior leaders from NVIDIA, AWS, JPMorgan Chase, AstraZeneca, Virgin Atlantic and more. The two‑day program will cover AI deployment, governance, infrastructure and industry‑specific use...
From AI Investment to Innovation: What It Takes to Deliver Real Business Impact
Organizations are pouring money into AI, yet many CIOs still see limited ROI. Jeff Baker of PwC argues that moving AI from isolated experiments to enterprise‑wide, business‑outcome‑driven initiatives is essential. He emphasizes pairing AI engineers with business units, securing clean...
AI Governance: The Complete 2026 Guide for Leaders | ClearPoint Strategy Blog
The ClearPoint Strategy guide reveals that only 16.9% of 327,582 tracked measures have a designated owner, creating a phantom‑owner risk for AI governance. In 2026, three major AI statutes—Texas RAIGA, Colorado AI Act, and the EU AI Act—take effect, while...
Mastermind Cuts 50+ Hours, Saves $25K for Leaders
When someone like Amy Porterfield—who’s built one of the most recognized brands in online marketing—says a mastermind changed her business, it’s worth paying attention. The AI Business Lab® Mastermind saved her 50+ hours through strategic prompts and workflows. Her CEO compressed...
Learn About Vertex Smart Categorization
Vertex Inc. released a new white paper titled “Creating a Retail Experience that Optimizes Accuracy, Efficiency, and Consistency,” highlighting how real‑time tax automation can eliminate pricing friction across channels. The company also announced a live webcast on June 9, 2026, covering AI‑enabled...
Samsung Says Galaxy Watch Can Predict Fainting up to 5 Minutes Early
Samsung announced that its Galaxy Watch 6 can forecast vasovagal syncope up to five minutes before it occurs, based on a clinical study with Chung‑Ang University Gwangmyeong Hospital. The AI‑driven algorithm achieved 84.6% overall accuracy, with 90% sensitivity and 64%...
WittKieffer Execs on How the Chief AI Officer Role Is Evolving
Health systems are rapidly adding chief AI officer (CAIO) positions, with the number of incumbents having more than quadrupled in the past few years. The role is shifting from back‑office AI pilots to enterprise‑wide programs that cover clinical, imaging, and...
Amazon Builds AI Agent Payments With Coinbase and Stripe
Amazon Web Services unveiled Bedrock AgentCore Payments, a managed feature that lets AI agents authenticate wallets, hold funds, and complete transactions within their execution loop. The preview relies on Coinbase’s x402 stablecoin protocol for instant micropayments and Stripe’s Privy wallet‑as‑a‑service...
Gluware’s Titan Rises to Meet Mythos Network Vulnerability Challenge
Gluware unveiled Titan Exposure Management, an AI‑driven, closed‑loop solution that identifies which devices across an enterprise are truly exposed to a given vulnerability and automates remediation. Built on the company’s DIAL (Device Interface and Automation Layer) framework, the tool continuously...
AI Image Generator Cheat Sheet: Nano Banana, ChatGPT, and More
The article provides a cheat sheet comparing leading AI image generators, highlighting strengths, weaknesses, and ideal use cases. Google Gemini’s Nano Banana Pro is crowned best overall with 4K resolution and reliable text‑in‑image, while OpenAI’s GPT Image 2 offers the strongest...
AGI Expected Within Five Years, Transforming Economy
Demis Hassabis, CEO of DeepMind on AI: - AGI will come in around 5 years, probably - London has been a great base (lots of talent & low competition; fewer distractions than SV) - Potential to be 10x the size of Industrial Revolution,...
AI Model Training on Unlicensed Works: Fair Use?
“Is researching, developing, and improving AI models on unlicensed works to advance the state of the art in the United States, a transformative fair use purpose—or not?” - this is the question we all need an answer to https://t.co/PjM6ZNgPU2
Build an AI Tool, Land a Job: Data Center Operator Launches Hiring Contest
DataVita, a Scottish data‑center operator, has launched the OpenClaw Challenge, an AI‑building competition that replaces traditional resumes. Participants must create a functional AI tool using Anthropic’s open‑source OpenClaw agent, judged on originality, technical merit, business value, security, and communication. The...
GKE Pod Snapshots Eliminate Cold Starts for Massive Models
Facing a cold start when your XX billion parameter model is loading up for inference on Kubernetes? You might love the new @googlecloud GKE pod snapshots. All your app state, along with most file system and networking state, get saved and...
AI Supply Chain Bottlenecked, Regulation Likely Biggest Constraint
$AMD and $ARM further validated this week that the entire AI supply chain is constrained. Compute, memory, storage, networking, energy, and talent. All the while regulatory may be the biggest bottleneck of all. https://t.co/mezQjhNsfi
4 Steps to Integrate AI Into Your Operations
Artificial intelligence has moved from novelty to a core strategy for accounting firms, with 81% reporting productivity gains and 86% noting reduced mental load, according to Intuit’s 2025 Accountant Technology Survey. Adoption varies, as many firms remain in experimental phases,...
Exploring Payment Models for AI-Driven Care Delivery
I've been canvassing the market to figure out all the possible pathways for getting paid for using AI in care delivery. Am I missing anything? https://t.co/vINromdwUT
AI‑Curated Viral Pipelines Are the Future of Social Media
Love it or hate it, this is where social media is heading. Why force originality when you can curate viral trends, run them through Claude agents, strip the AI smell, and auto-publish? Robin just open-sourced this pipeline (Claude Code, Postiz, Apify). Fascinating to...
Returning to Classic: Clippy Beats AI Assistant Overload
Too many choices for AI assistants - I am just going to go back to where it all began - Microsoft Clippy https://t.co/cqQlnzkE1c
AI Matchmaking: Are Bots Finding Real Relationships?
Dating app companies seem poised to embrace features like AI assistants or bots that date for you -- I wonder, have you or anyone you know actually met a partner through some sort of AI-based matchmaking and would you be...
Memory, Not Math, Drives AI's 1000× Energy Gains
To improve AI energy efficiency by 1000x, it's all about memory and data movement, not the efficiency of the arithmetic. Biomimicry of the brain, with analog in-memory compute, affords a 100x. Getting to 1000x might be unconventional: a continuous-time physical dynamical...
Robots Learn Real-World Cleaning and Folding Skills
From Clicks to Chores:Teaching #Robots to Clean, Fold, and Learn in the Real World by @jason_chen998 #Robotics #Engineering #ArtificialIntelligence #Technology https://t.co/Ycc9cpKNP8
GPT‑Realtime‑2 Launches, Boosting Voice‑Driven AI Interaction
people are really starting to use voice to interact with AI, especially when they have a lot of context to dump. GPT-Realtime-2 comes to the API today; it is a pretty big step forward. (we are working on improvements to voice in...
Prefer REST APIs; Reserve MCP/A2A for Complex Reasoning
Choosing between APIs, MCP, and Agent-to-Agent architectures https://t.co/4TnEq1Dbq6 < @lak_luster says to default to function tools that use REST APIs, and only bring in MCP or A2A when there's more reasoning involved, or dynamic use cases. https://t.co/Am9DfWv1Qy
AI Speeds Work, but Scaling Needs More Hires
a IC can build 100x faster with ai but when you have 20 sessions window open working furiously, there still a limit to how much more you can do. the solution? hire more people.
AI Could Either Boost or Undermine Agency Profitability
Which do you believe? 3 years from now, will marketing agencies be more profitable and more valuable, or less profitable and less valuable? Two trains of thought: 1. More valuable. AI expands margins, lean teams deliver far more. 2. Less valuable. AI replaces what...
AI Complexity Favors Established Firms over Pure‑play SaaS
Many of the big cos that will make it are likely the ones that benefit from the complexity that AI creates vs just betting on new agent success (also Datadog is more infra and not pure play SaaS - harder...
OpenAI Staff Resign over Safety Concerns, Notable Exits
Many, many OpenAI employees quit over safety concerns, including @DKokotajlo, William Saunders, @sjgadler, etc as well @Janleike. The founders of Anthropic such as @DarioAmodei and @jackclarkSF may have left for similar reasons. Does anybody have a full list?
Open‑source Real‑time World Models Now Under 50 Ms Latency
Open source real-time world models legit feel like puppeteering reality itself. Now @reactorworld is putting it in everyone’s hands at sub 50 ms latency -- type a prompt, change the scene, and control what unfolds as it generates. https://t.co/99PObbdqqb
China Rolls Out Thousands of Service Robots Nationwide
China Deploys Thousands of Service #Robots Across Its Cities by @ShangguanJiewen #Robotics #Engineering #ArtificialIntelligence #Innovation #Technology https://t.co/wMKdyeN7ay
AI Progress, Not Cash Flow, Drives Hyperscaler Value
The bears pointing to fall in hyperscaler free cash flow is a "dumb argument." "All that means is CapEx is high. - @WarrenPies The real "north star for this market": "is AI still improving," not 1's CapEx is 2's OpEx "Compute Demand...