Today's AI Pulse
Salesforce to acquire AI customer‑service platform Fin for $3.6B
Salesforce announced a $3.6 billion acquisition of Fin, an AI‑driven customer‑service platform, slated to close in fiscal Q4 2027. Fin’s AI agents handle chat, email, WhatsApp, text, phone and Slack, resolving roughly 76% of tickets without human intervention. The deal adds over 30,000 enterprise customers and $400 million in annual recurring revenue.
Also developing:
By the numbers: Waymo acquires Apple’s Arizona Proving Ground for $220M
Alibaba's Chief AI Developer Quits, Taking Key Team Members with Him
Alibaba’s lead AI researcher Junyang Lin, the architect of the Qwen model series, announced his unexpected resignation. Several core engineers—including Qwen coder Binyuan Hui and post‑training specialist Bowen Yu—left the company on the same day. In response, CEO Eddie Wu created a Foundation Model Task Force, co‑led with CTO Zeming Wu and Jingren Zhou, and reaffirmed Alibaba’s commitment to open‑source AI. The firm also signaled a boost in AI investment to sustain its competitive edge.
February 2026: Month in Review
February 2026 highlighted how AI is no longer a theoretical add‑on for the built environment but a catalyst reshaping careers, business models, and industry structures. The month’s posts argued that AI is dismantling traditional gatekeeping—credentials, PE licenses, and SaaS lock‑in—while...
When No Two Spines Are Alike: Inside the First AI-Designed Cervical Implant
UC San Diego Health performed the world’s first fully personalized anterior cervical spine implant, combining high‑resolution imaging, AI‑driven design, and titanium 3D printing. The AI algorithm generated a patient‑specific geometry that matches the vertebral endplates, restores natural lordosis, and optimizes...
Craig Matthews, LEAP CEO: LegalEx London Shows How AI, Used Properly, Can Unlock Real Efficiency for Law Firms
At LegalEx London, LEAP CEO Craig Matthews emphasized that AI must be integrated, governed, and purpose‑built to deliver real efficiency for law firms. He warned that firms face daily pressures around productivity, risk, cybersecurity, and burnout, and that AI can...
Treasury Taps Tech Firms to Advise on Public Sector AI Adoption
HM Treasury has set up an advisory panel that includes IBM, Faculty AI and the Tony Blair Institute to accelerate artificial intelligence adoption across government departments. The group will evaluate existing AI pilots, recommend scaling strategies, and feed insights into...
‘Anthropic Unveils Claude Legal Plugin and Causes Market Meltdown’ – Our Analysis
Anthropic unveiled a Claude legal plugin on February 3, 2026, enabling its large‑language model to perform document review and other legal tasks. The announcement triggered a sharp sell‑off in publicly listed legal‑software firms such as Pearson, RELX, Thomson Reuters, Wolters Kluwer and...
Loop Capital Thinks This Chip Manufacturer Could More than Double on an AI ‘Super Cycle’
Loop Capital has initiated Astera Labs (ALAB) with a buy rating and a $250 price target, implying roughly 120% upside from its recent close. The analyst positions Astera as the most diversified AI‑silicon pure‑play outside Nvidia, poised to benefit from...
BirdsEyeView Launches AI Data Scrubbing to Streamline Hazard Modelling
BirdsEyeView, an ESA‑backed insurtech, unveiled AI Data Scrubbing, an automated tool that cleans, standardises and geolocates Statement of Values (SOV) files. The solution transforms raw exposure data into modelling‑ready inputs within minutes, handling up to 10,000 locations per run and...
Körber’s Michael Becher to Explore AI-Driven Intralogistics at IntraLogisteX 2026
Körber’s Vice President of Digitalisation & AI, Michael Becher, will lead a session at IntraLogisteX 2026 on AI‑driven intralogistics, focusing on digital twins, agentic AI and predictive maintenance. The presentation, scheduled for March 19 at NEC Birmingham, will demonstrate how...
Perplexity AI Introduces Voice Mode in Perplexity Computer
Perplexity AI has launched Voice Mode within its Perplexity Computer platform, enabling users to ask questions aloud and receive near‑instant answers. The feature supports continuous, multi‑turn conversations, allowing follow‑up queries without re‑activating the assistant. By integrating speech input directly into...
OpenAI Shifts ChatGPT Purchases to Partner Apps
Wait, nobody is actually buying things in ChatGPT? SHOCKING :) Big advantage Google -> OpenAI is scaling back shopping directly inside ChatGPT via Instant Checkout by having checkouts instead take place in the specific apps that plug into ChatGPT "OpenAI staff...
Lack of Regulatory Action on Hyperscaler Dominance Prompts Inquiry Chair to Quit
Kip Meek, chair of the UK Competition and Markets Authority's cloud inquiry, resigned in protest over the regulator's sluggish response to hyperscaler dominance. The CMA’s report highlighted that Microsoft Azure and Amazon Web Services control roughly 70‑90% of the UK...
Tata Elxsi Launches DevStudio.ai, a Multi-Agent, ASPICE-Aligned GenAI Platform to Accelerate Automotive Software Engineering
Global design firm Tata Elxsi unveiled DevStudio.ai, a generative‑AI platform built for the automotive software development lifecycle. The solution uses a multi‑agent architecture aligned with ASPICE standards and can run on cloud or air‑gapped on‑premise environments. Early deployments with OEMs...
The Sequence Opinion #819: How AI Chips Are Made?
The post explains that AI performance in 2026 hinges more on hardware than algorithms, with GPUs—originally built for graphics—serving as the foundation for neural‑network training. It outlines the engineering journey from high‑level RTL and Verilog code through physical design to...
Active Investors Spent More On Fewer Deals In February
Active venture investors concentrated capital on a handful of mega‑deals in February, with generative‑AI heavyweights dominating the landscape. OpenAI secured a record‑breaking $110 billion round, the largest startup financing ever, while overall deal count remained flat. Y Combinator, Andreessen Horowitz and...
The Great Migration: Why Every AI Platform Is Converging on Kubernetes
Kubernetes has become the de‑facto operating system for AI, with 82% of container users and two‑thirds of generative‑AI teams running production workloads on the platform. The shift from stateless microservices to data processing, distributed training, LLM inference, and autonomous agents...
Hyperlocal AI with a Million Subscribers
Patch, operating under the PatchAM brand, has launched AI‑generated newsletters that now serve 14,000 hyperlocal communities and have amassed nearly one million subscribers. The service pulls content from aggregation sources, event calendars and platforms like Nextdoor, delivering daily or twice‑weekly...
Q&A: AWS Launches Amazon Connect Health to Streamline Healthcare Workflows
Amazon Web Services unveiled Amazon Connect Health, an agentic AI platform that plugs directly into electronic health records to automate patient verification, appointment scheduling, ambient documentation, and medical coding. The service, built on FHIR standards and Amazon HealthLake, offers a...
Your AI Agents Can Talk to Each Other - but Are They Saying Anything Useful? Confluent Intelligence Aims for Insight
Confluent announced that its Streaming Agents now support Anthropic’s Model Context Protocol (MCP) and the Agent2Agent (A2A) protocol, creating a dual‑layer for real‑time data ingestion and inter‑agent coordination. The platform also introduced multivariate anomaly detection powered by Apache Flink, offering...
LLM Agents Interview Questions #11 - The Lost-in-the-Middle Trap
In a senior AI engineer interview at Stripe, candidates are asked why a text‑to‑SQL agent that packs 50 grammar rules into an 8k prompt loses constraints and hallucinates joins. The trap reveals a misunderstanding of attention density versus raw context...
AI Agents Need Logins Too: Identity, Security, and the Future of AI | Greg Keller, CTO, JumpCloud
In this episode, JumpCloud CTO Greg Keller explains the evolving role of a CTO and how JumpCloud reimagines identity and access management (IAM) for modern, heterogeneous IT environments, contrasting it with legacy solutions like Microsoft AD and Okta's SSO focus....
UpKeep Launches Studio, an App Platform That Lets Maintenance Teams Build Their Own Software
UpKeep announced Studio, an app platform that lets maintenance and operations teams build, install, and run custom software inside their CMMS without writing code. The platform features an AI‑driven chat builder, a curated marketplace of 30+ ready‑made apps, and a...
LiveU Marks First Large-Scale Global Deployment of AI-Driven LIQ at Winter Games
LiveU deployed its AI‑driven LiveU IQ (LIQ) technology at the Winter Games, marking the first large‑scale global rollout for a multi‑venue sporting event. Broadcasters used LIQ for roughly 60% of live sessions, achieving a 36% increase in average bitrate and...
Particular Audience Just Smashed Open the Blackbox of AI Search
Particular Audience unveiled Search Model A/B Testing, enabling retailers to directly compare AI relevance models that drive organic and sponsored search results. Building on its Adaptive Transformer Search (ATS) technology, which has cut zero-result queries to under 0.5%—far below the...
ŌURA Acquires Doublepoint to Expand AI-Driven Interaction Capabilities
Oura announced the acquisition of Doublepoint, a Helsinki‑based AI gesture‑recognition startup, to embed natural, biometric interactions into its wearable platform. The deal brings Doublepoint’s four founders and AI architects into Oura’s R&D, reinforcing a roadmap that blends voice and gesture...
AI Is Changing How People Use AT&T Home Internet—And Helping Stabilize the Network
AT&T reports that consumer upload traffic is growing twice as fast as download traffic, driven by AI interactions such as voice, video, and code submissions. By 2025, this shift reflects a broader change in home internet usage patterns, with users...
More Raw Data ≠ Higher Intelligence, Says Visual Compression
I keep reading this take (below) every few months, presented as if extremely profound, and it is just offensively dumb. It confuses data and information, it ignores the fact that not all information is equally valuable, and it ignores the...
How AI Is 'Terraforming' A New Grindr - the Internal and External Impact of a Tech Revolution in Action
Grindr CEO George Arison says the dating app is rapidly becoming an AI‑native organization, with roughly 70% of its code now produced by AI agents. The AI‑driven workflow has boosted engineer productivity by about 1.5 times and enables features like Edge’s...
AgentPass Rebrands as Paz.ai, the Agentic Commerce Platform for Retailers
AgentPass has rebranded to Paz.ai, positioning itself as the infrastructure layer that links retail product catalogs to conversational AI shopping agents such as ChatGPT, Google AI Mode, and Perplexity. The platform ingests feeds from Shopify, Adobe Commerce and Salesforce Commerce...
Financial Times – Ex-Police Officer Wrongly Flagged as a Shoplifter in Shop Using Facial Recognition
Former police officer Lavinia McIntyre was mistakenly identified as a shoplifter by a Budgens store using Facewatch facial‑recognition software. The technology, now employed by several UK retailers, flagged her without any proof, prompting staff to ask her to leave. The incident...
Orange 142 Launches Ignition+, an Integrated, AI-Enabled Programmatic Media Solution Focused on Transparency and Efficiency
Orange 142, a Direct Digital Holdings division, unveiled Ignition+, an AI‑powered programmatic platform designed for Fortune 1,000 brands and large agencies. The solution consolidates activation, supply access, optimization and measurement into a single stack, promising greater transparency and a flat‑fee pricing...
Simmons & Simmons Launches AI + Legal Privilege Guide
International law firm Simmons & Simmons has released an “AI and Legal Privilege Guide and Policy Framework” to help organisations manage confidentiality risks when using artificial intelligence. The guide responds to a recent UK court ruling that uploading client documents...
Inside the Modern Underwriting Strategy
Underwriting is transitioning from retrospective portfolio reviews to AI‑driven, real‑time decision support. A 2025 Accenture study shows 71% of underwriting executives view AI and automation as critical, and 75% expect meaningful AI impact within three years. Modern workbenches now surface...
AI Legal Research Startup Descrybe Launches ‘Legal Reasoning’ Tool; Says It Outperforms ChatGPT, Claude, and Gemini on Bar Exam Benchmark
Descrybe, an AI legal research startup, unveiled DescrybeLM, a legal reasoning engine that it claims surpasses leading general‑purpose models such as ChatGPT, Claude, and Gemini on a standardized bar‑exam benchmark. The company published the benchmark methodology and scoring data, inviting...
British Bosses Now Ask AI Whether They Should Use AI
Confluent’s Quick Thinking 2026 survey of 200 UK CEOs reveals that 62% now rely on AI for the majority of their decisions, and 70% admit to second‑guessing their own judgment when AI recommendations differ. The report shows a shift toward...
EP. 06: Making The Most of Technologies And Trends Impacting Adjusters
Rob Evans, director of claim process technology at Broadspire, emphasized that AI adoption should intentionally support adjusters rather than replace them. In a recent Claims Journal podcast, he highlighted how litigation prediction models and AI-driven notifications can improve claim outcomes...
Mosaic Clinical Technologies Announces FDA Breakthrough Device Designation for Cognita’s Generative AI Model for Radiology
Mosaic Clinical Technologies announced that its AI unit Cognita received FDA Breakthrough Device Designation for Cognita Chest X‑Ray (CXR), the first generative vision‑language model in radiology to earn this status. The designation grants prioritized FDA interaction, potentially accelerating clearance. Internal...
Meta’s New AI Unit Takes Flat Management Structures to the Extreme
Meta has launched a new applied AI unit staffed with a flat 50:1 manager‑to‑engineer ratio. The group, headed by Maher Saba and reporting to CTO Andrew Bosworth, will collaborate with the Superintelligence Lab to build a data engine that accelerates model improvement....
Bookipi Survey Shows Small Businesses Embracing AI BUT, Complexity, Not Cost, Slows Workflow Adoption
The 2026 Bookipi Small Business AI Adoption Report, based on 2,121 owners across four regions, shows nearly half of small firms (48%) will raise AI budgets while only 2% plan cuts. Companies are comfortable using AI for customer‑facing tasks such...
Guiding Nano Assembly for Drug Delivery with Machine Learning
Researchers repurposed the FAP inhibitor SP‑13786 as a co‑assembly excipient to create SP co‑assembled nanoparticles (SCAN) that encapsulate hydrophobic drugs. Using molecular dynamics and a random‑forest machine‑learning model, they identified 228 physicochemical descriptors that predict successful nano‑co‑assembly, highlighting aromaticity and...
The Guardian Updates Its AI Policies Around Training, Trust and In-House Tools
The Guardian has revised its generative AI policy, mandating AI training for all staff and developing proprietary AI tools for editorial tasks. The updated policy adds a transparency requirement, with footnotes indicating significant AI use in stories. Revised editorial code...
AI + Buildings 2026: Redefining Autonomy From People to Portfolios
The third annual AI and Autonomous Buildings panel highlighted a shift from reactive HVAC controls to agent‑based intelligence that can negotiate power, predict maintenance, and balance cost, carbon, and comfort. Attendees saw digital twins scaling from single sites to entire...
Opinion: AI Could Revolutionize Antibiotics — but the Market Is Standing in the Way
Antibiotic resistance now kills roughly 4.9 million people each year, and the last truly novel class of antibiotics was introduced over three decades ago. Artificial intelligence promises to accelerate the discovery of new antimicrobial agents by scanning vast chemical spaces and...
Capitec’s New AI Tool Knows Your Problem Before You Explain It
Capitec has introduced Pulse, an AI‑powered contact‑centre tool that delivers real‑time context about a client’s issue before the call starts. By aggregating payment data, app diagnostics and risk signals, Pulse enables agents to resolve queries faster, cutting handling times by...
Future‑proof AI: Learning Ability Outweighs Launch Accuracy
Why the most valuable AI systems are not the most accurate ones today, but the ones designed to learn tomorrow In the early days of enterprise AI, success was measured in a single moment: the model launch. A team would...
Google's AI Recipe Widget Serves Frankenstein Dishes
Google testings new AI Mode recipe widget/card that sucks searchers into AI-Frankenstein recipes https://t.co/6X6gqAk77S via @inspiredtaste https://t.co/manqzqxFIa
Gemini Projects Rebranded as Notebooks, Integrated with NotebookLM
Gemini "Projects" may get a new name and a powerful NotebookLM tie-in Leaked strings reveal "Notebooks" will let users group chats by topic and share sources between Gemini and NotebookLM ✅ More details - https://t.co/48rQwIdXWM https://t.co/yLmaIQWa2i
Ultrathink Returns: Deep Reasoning Boosts Bug Fixing
ULTRATHINK IS BACK. Claude Code v2.1.68 brings back the fan-favorite deep reasoning mode to your terminal. 🔥 → Purpose: Solving impossible bugs + "tough spots" → How: Just type "ultrathink" in your prompt Massive efficiency gain 👊 https://t.co/8E1wqu8RAw
Autonomous Networks Evolve Into KPI‑Driven Economic Strategy
At #MWC26 one theme stood out at the @Ericsson pavilion: Autonomous networks are no longer just about automation. They are becoming an economic strategy, where intent-driven architectures align network behavior directly with business KPIs. My latest analysis. https://t.co/eC7D8S6bFP #AI @MWCHub @GSMA
Humachineology: Harnessing Human‑Machine Synergy for Prosperity
Humachineology and humachinekind provide a framework to develop prosperity through studying the synergy between humans and machines.