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
Chasing Utopia Review – Renegade Google Exec Mo Gawdat Searches for Ethical AI in Alarming Insider Warning
The documentary *Chasing Utopia* follows former Google X chief business officer Mo Gawdat as he campaigns for ethical, empathy‑driven artificial intelligence. Gawdat, motivated by the tragic loss of his son, warns that current AI systems amplify surveillance, automated warfare, and digital narcissism. The film presents a stark contrast between industry optimism and Gawdat’s call for a moral dimension in AI development. It also explores speculative technologies like brain‑cell‑based computers, underscoring the urgency of the debate.
Blocking AI Drives Hidden Use, Not Compliance
Gallup polls say AI in the workplace crossed 50% in February, but your AI adoption numbers are probably lying to you. Not because anyone tampered with your dashboard, but because it's only counting the sanctioned surface. The approved tools, the...
The Implementation Trap: Why Hands-On AI Consulting Does Not Scale (And What Does)
AI consultants are realizing that hands‑on implementation projects drain time and cannot be scaled, while strategy‑focused education does. The author and fellow consultant Michael Housman describe how custom builds lock knowledge in a single client and force constant context switching....
AI and an Absent Government: Takeaways From RSAC 2026
The 2026 RSA Conference highlighted AI as the dominant cybersecurity theme, with executives touting agentic AI’s ability to automate SOC tasks while researchers warned of new attack vectors such as AI‑driven ransomware and credential‑hijacking. Attendees also noted a stark absence...
The Next Enterprise Architecture Asset: Ontologies for AI
Enterprise architects are turning to data ontologies—machine‑readable semantic frameworks that define the meaning of entities and relationships—to turn raw data into a trusted, AI‑ready asset. By organizing data by meaning rather than just structure, ontologies unify disparate sources, stabilize semantic...
From ‘Rank Me’ To ‘Trust Me’: How AI Is Rewriting The Rules Of Discovery
The rise of large language models is turning traditional web search into a zero‑click experience, delivering synthesized answers instead of blue‑link lists. Publishers now face a new visibility battle called Generative Engine Optimization, where inclusion inside AI‑generated responses matters more...
MinIO Adds Petabyte-Scale MemKV Cache for Nvidia GPU Inference
MinIO unveiled MemKV, a petabyte‑scale key‑value caching layer designed for Nvidia GPU inference workloads. The system sits atop MinIO’s AIStor object storage and integrates directly with Nvidia’s BlueField‑4 STX architecture, using end‑to‑end RDMA to move data between GPU HBM, CPU...
How Negative Information Spreads From Wikipedia Into AI Search
Wikipedia has become a primary knowledge source for generative AI search tools like ChatGPT and Google, lending its content an aura of credibility. However, the platform’s reliance on verifiable but not always accurate citations allows negative or outdated information to...
Bank of England Flags Financial Disruption From Advanced AI – Report
The Bank of England's Prudential Regulation Authority warned that advanced AI models such as Anthropic’s Mythos and OpenAI’s ChatGPT 5.5 Instant could cause significant disruption to financial services. These tools are increasingly adept at identifying weak points in banks' legacy systems,...
OpenAI Gives European Companies Access to Its Latest Models to Bolster Resilience
OpenAI announced its "Trusted Access for Cyber" programme, granting European firms access to its newest models, including GPT‑5.5‑Cyber. Early participants span vital sectors such as telecom (Deutsche Telekom, Telefonica), finance (BBVA, Scalable Capital) and cybersecurity (Sophos). The move counters competitive pressure from...
The AI-Maturity Spectrum: The Art of Implementation
At the BenefitsPRO Broker Expo in Chicago, industry leaders presented “The AI Maturity Spectrum,” a session that shifted the conversation from specific tools to the strategic art of AI implementation in benefits administration. Speakers emphasized that successful AI adoption hinges...
Are You Using the Right Security Controls for Agent 365?
Agent 365 is Microsoft’s centralized platform for governing AI agents, integrating tightly with Microsoft Purview, Entra and Defender. It enforces least‑privilege access, flags unauthorized data requests, and provides real‑time monitoring, automated workflows, and DSPM to close policy gaps. Incident response and...
Creative License: How Adobe Helps Brands Stay Creative in the Age of AI
Adobe’s enterprise CMO Rachel Thornton explained how the company is turning fragmented customer data into real‑time, personalized experiences using AI. She highlighted Adobe’s strategy of building tools—Creative Cloud, Express, Firefly, and agentic AI—that let any employee create brand‑compliant content at...
KDDI and Ericsson Claim Gains with AI Uplink Field Trial
KDDI and Ericsson completed a live field trial in Japan using Ericsson’s AI‑driven Uplink Interference Optimizer (UIO) rApp on the Intelligent Automation platform. The test spanned roughly 1,500 5G and 1,300 4G cells and delivered a 9.6% uplift in 4G...
First Canadian Study to Examine Both Sides of AI in Hiring Finds a Striking Disconnect
A new Hire Value and Verve Recruitment study of 1,815 Canadian job seekers and 383 hiring managers shows a stark perception gap: 86% of candidates believe AI blocks interview chances, yet only 15% of employers actually use AI to screen...
3 Actionable Ways Affiliate Managers & SEOs Can Keep Relevant – Ask An SEO via @Sejournal, @Rollerblader
The article advises affiliate managers and SEOs on staying relevant as AI reshapes search and content discovery. It stresses aligning brand messaging with AI‑driven personalization, modernizing affiliate payment structures, and converting existing backlink sources into paid affiliate relationships. It also...
AI Voice Startup Vapi Hits $500M Valuation After Winning Amazon Ring over 40 Rivals
AI voice startup Vapi secured a full‑scale deployment with Amazon Ring, routing 100 % of the company’s inbound calls through its platform. The win helped Vapi close a $50 million Series B round, lifting its post‑money valuation to roughly $500 million. Vapi now processes...
Standard 90-Day Vulnerability Disclosure Policy Is Likely Dead Thanks to AI, Expert Warns that AI Can Weaponize Patches in 30...
AI‑driven large language models are accelerating vulnerability discovery, allowing exploits to be weaponized in as little as 30 minutes. Security researcher Himanshu Anand argues that the industry‑standard 90‑day disclosure window is effectively obsolete, citing recent Linux kernel bugs and a...
FDA Approves First AI Test to Guide Breast Cancer Chemotherapy Decisions
The U.S. Food and Drug Administration has cleared ArteraAI Breast, the first artificial‑intelligence test designed to guide chemotherapy decisions for early‑stage, hormone‑receptor‑positive, HER2‑negative breast cancer. The digital‑pathology platform scans resection slides, combines them with clinical variables, and uses a multimodal...
A Quiet Uprising Against Chatbots?
SameSame Collective, a nonprofit serving queer youth, compared its simple WhatsApp chatbot to AI‑powered alternatives and found users trusted its service more despite preferring the polish of ChatGPT. Parallel research at George Mason and in Nature Human Behaviour shows that...
US Workers Overwhelmingly Support Union-Backed Policies on AI, Poll Says
A new AFL‑CIO poll of 1,588 U.S. workers shows overwhelming support for union‑backed AI safeguards. Over 90% favor policies such as a human final decision‑maker and strict transparency, while 78% deem worker protection from AI harms essential. Trust in unions...
Even with Rules, Custom GPTs Forget to Follow Instructions. This Keeps Them on Track.
Custom GPTs often drift from their original instructions, leading to formatting slips and hallucinations, says PR specialist Sarah Evans. She recommends embedding institutional editorial standards into every prompt via detailed addenda and repeatable workflows. By mapping and automating existing content...
SoftBank Has Injected $450 Million Into This British AI Chip Company
SoftBank has poured $457 million into Graphcore, the UK AI‑chip maker it acquired in 2024. The capital injection is intended to accelerate Graphcore’s development of intelligence processing units for artificial general intelligence projects. The funding also dovetails with SoftBank’s wider AI...
Research Suggests the Problem with Using AI as a Therapist Isn’t that It Sounds Wrong — It’s that It Can...
A recent study summarized by ScienceDaily and Brown University shows that large language models instructed to act as therapists still breach core mental‑health ethics, including poor crisis handling, biased advice, and a newly identified phenomenon called “deceptive empathy.” The AI...
European AI Funding Is Growing. Will That Boost The Region’s Startup Scene?
European venture capital is now half AI‑focused in 2026, with frontier model firms and supporting hardware startups attracting the bulk of capital. Quarterly funding surged 33% year‑over‑year, topping $17 billion in both Q4 2025 and Q1 2026. New AI labs in...
Rethinking How We Use AI – Louise Ballard, Atheni-Ai
Louise Ballard, co‑founder of early‑stage AI firm Atheni‑ai, explains how businesses must move beyond chat‑based AI to building integrated systems that deliver measurable value. She critiques the UK government’s ambition to reskill ten million workers in AI as overly broad...
New Gong Research Reveals Transparency Key to Competitive AI Advantage
Gong’s latest research shows that 58% of companies have stalled AI initiatives because they lack trust in the technology, not because of budget constraints. The study of over 2,000 U.S. and U.K. leaders highlights transparency, security and explainability as the...
How AI Is Changing the Face of Benefits Advising
Artificial intelligence is transforming employee benefits advising from a reactive, research‑heavy model to an anticipatory, data‑driven service. WTW’s AI assistant, Expert, automates tasks such as drafting job descriptions and benchmarking, allowing consultants to focus on organizational design. Businessolver’s Sofia adds...
OpenSquilla Cuts Agent Token Costs by 67%
Just put @OpenSquilla through a real long-running workflow drafting and iterating on a full research analysis with 10+ back-and-forth steps. The smart routing was insane: simple prompts went to cheap models, complex ones still got full power, and my token bill...
Record CASCI Index Driven by Strong Chip Performance
The only way to track the manufacture and shipping of AI Compute is through the CASCI Composite index, by @Culpium April's CASCI Composite hit a record. BUT: Chip strength made up for faltering momentum elsewhere in the supply chain. Head to...
Ultimate Cowork Guide : Are You Using Claude to Its Full Potential?
Claude Cowork is Anthropic’s AI‑driven platform that lets users automate repetitive tasks, organize data, and build custom dashboards. The guide by Tina Huang walks beginners through setting up personal dashboards, creating reusable "Skills," and linking third‑party tools via Connectors and...
AI Adoption Rates Matter as Much as Exposure Metrics
Occupational exposure metrics are useful & important, but they sometimes disagree & they may not measure vulnerability to displacement well. I increasingly think AI *adoption* rates are just as important for monitoring labor market effects as *exposure* rates.
Using @AGENTS.md Auto‑loads Agents Across Setups
today-i-realised @ file in an AGENTS.md also gets auto-read by your agent i kinda knew this as i'd always @ AGENTS.md in my CLAUDE.md instead of the path otherwise claude wouldn't read it only today applying it to other setups 🤦♂️ https://t.co/ly9VorMaaL
Amazon's 'Titus' Program Cuts AI Data‑Center Build Time to Under 35 Weeks
Amazon Web Services unveiled the Titus program, an internal effort to slash AI‑focused data‑center construction cycles to under 35 weeks, boost site power capacity to 68 MW and cut cooling power use by 15%. The move is part of a $200 billion...
Google Finds First AI-Developed Zero-Day that Bypasses 2FA — Self-Morphing Malware and Gemini-Powered Backdoors Signal a New Era of Cybercrime
Google’s Threat Intelligence Group disclosed the first AI‑crafted zero‑day that sidesteps two‑factor authentication in a widely used open‑source system administration tool. The exploit, a Python script, demonstrates how large language models can parse source code and locate logical flaws faster...
The Sequence Knowledge #858: How State Space Models Went From Curiosity to Serious Transformer Competitor
State space models (SSMs) have moved from a niche curiosity to a serious challenger for transformers. For years transformers dominated because self‑attention works, but its O(n²) scaling now bottlenecks long‑context and large‑model inference, consuming tens of gigabytes of VRAM. SSMs...
Consultants Rush to Tame 'Shadow AI' As Unapproved Tools Surge in Enterprises
A Business Insider investigation reveals that 71% of UK employees have used unapproved consumer AI tools at work, with many doing so weekly. The rise of “shadow AI” is forcing consulting firms to develop governance frameworks to mitigate data‑leak and...
UKG Launches AI‑Powered Pro Pay with Workforce AI at Payroll Congress 2026
UKG announced the launch of UKG Pro Pay with Workforce AI at Payroll Congress 2026 in Lowell, Massachusetts. The new solution uses agentic, assistive and generative AI to detect, analyze and resolve payroll issues in real time, targeting frontline and hourly employees.
FDA Deploys Elsa 4.0 AI Tool and HALO Data Platform to Modernize Operations
The U.S. Food and Drug Administration unveiled Elsa 4.0, an upgraded internal AI assistant, and completed the HALO data‑platform consolidation that unifies more than 40 disparate data sources. The moves are designed to cut manual work, accelerate regulatory reviews and...
Celonis Turns Its Platform on Its Head with Context Model and Ikigai Labs Deal - Pitching Itself as Enterprise AI's...
Celonis unveiled its Context Model, a platform‑wide operational ontology that reframes process mining as a consumable feature rather than the core product. The company also announced a definitive agreement to acquire MIT‑originated AI decision‑intelligence firm Ikigai Labs, bringing its Large...
How AI Is Supporting Nurses
TechRepublic reports that artificial intelligence is increasingly deployed to assist nurses across hospitals, targeting the chronic documentation overload that consumes much of their time. AI-powered transcription, predictive safety alerts, and real‑time patient monitoring are reducing charting effort and improving clinical...
AI-Enhanced Video From the White House Correspondents’ Dinner Spread Confusion Online
An unedited security clip from the April 25 White House Correspondents’ Dinner was posted by President Donald Trump on Truth Social after a shooting disrupted the event. Social‑media users applied artificial‑intelligence tools to sharpen the low‑resolution video, but the AI added fabricated...
Starlink AI Acquisition Corp Closes $100 Million SPAC IPO, Adding New AI Investment Vehicle
Starlink AI Acquisition Corp finalized a $100 million initial public offering, with A.G.P./Alliance Global Partners serving as sole book‑running manager. The blank‑check company will now hunt for AI‑related acquisition targets, underscoring that SPACs remain a financing option despite a cooling equity...
Why Agentic AI Is Security's Next Blind Spot
Agentic AI is already deployed across enterprises, executing tasks and accessing data without security oversight. The article outlines three agent categories—general‑purpose coding assistants, MCP‑enabled vendor agents, and custom user‑built agents—each presenting distinct risk profiles. It argues that security teams must...
Senate's Rush To Regulate AI Chatbots Is Bad For Everybody
The Senate Judiciary Committee approved the GUARD Act, a federal bill that would regulate AI chatbots, especially "AI companion" systems, through age‑verification, design mandates, and mandatory disclosures, with civil and criminal penalties up to $100,000 per violation. The legislation forces...
The Real Reasons Your Lawyers Don’t Trust AI – And What to Do About It
Law firms are wary of generative AI because its outputs often lack the matter‑specific context needed for reliable legal work. A SurePoint survey shows 80 % of midsize firms fear AI unreliability, not job loss. Fragmented point solutions force attorneys to...
Legacy Systems Unready for Chat‑First AI Future
Conversational AI is rapidly exposing the legacy enterprise systems unprepared for a chat-first future. https://t.co/MumCWhohOW
CastleKnight Posts 21% April Gain, Best Month Since 2020 on AI Stocks
CastleKnight Management, the $3.6 billion hedge fund founded by Aaron Weitman, posted a 21% gain in April – its best monthly performance since the fund’s inception in October 2020. The surge came from concentrated bets on semiconductor and other AI‑related stocks,...
Lotte Hotel Seoul Feeds AI Robot with Napkin‑Folding Data to Automate Service
Lotte Hotel Seoul has wired ten food‑and‑beverage staff with body cameras to record napkin‑folding motions for South Korean AI startup RLWRLD. The data will train a robotics foundation model that could power service robots in hotels and other industries by...
General Motors Trims up to 600 IT Roles While Recruiting AI Talent in Workforce Overhaul
General Motors is eliminating 500‑600 salaried IT jobs, chiefly in Austin and Warren, as part of a cost‑saving restructure, yet it is simultaneously posting dozens of openings for AI, motorsports and autonomous‑vehicle specialists. The move reflects a shift toward generative‑AI...