Today's AI Pulse
OpenAI unveils ChatGPT Images 2.0 with multilingual text and ‘Thinking’ mode
OpenAI rolled out ChatGPT Images 2.0, a next‑generation image model that adds multilingual typography, real‑time web research, and agentic reasoning. The new ‘Thinking’ mode lets paid users generate up to eight coherent images per prompt, including floor plans, maps and UI mock‑ups. The feature expands the model’s ability to work with uploaded content.
Also developing:
How AI Is Changing Food Supply Chains
Food manufacturers are turning to AI to overcome pandemic‑induced supply chain volatility, especially for perishable goods. CookUnity, a chef‑to‑consumer meal subscription service, now uses AI to lift its sales‑forecast accuracy from roughly 55% to 80‑90%, enabling precise, temperature‑controlled deliveries. The AI platform also provides real‑time disruption alerts, shared‑truck optimization, and automated consumer‑feedback analysis. Industry experts see AI becoming a logistics “copilot,” helping firms react faster and cut deadheading.
YouTube Bringing AI Feature to Connected TV App
YouTube is extending its AI‑driven Ask feature from the mobile app to the connected TV app, with a rollout slated for the coming months. The new Ask button lets viewers pose questions about the video they’re watching using their remote’s...
Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed
Bloomberg reports that roughly half of U.S. data centers slated for opening in 2026 are facing delays or cancellations. Of the 12 GW of capacity announced for this year, only about a third is actually under construction, with similar shortfalls projected...
ScaleOps Closes $130M Series C to Advance Autonomous AI Infrastructure Management
ScaleOps announced a $130 million Series C round led by Insight Partners, lifting its valuation to roughly $800 million. The funding backs its autonomous AI‑driven platform that continuously balances compute, memory, storage and networking resources. Founded in 2022, the company targets chronic GPU...
Real‑World AI Transforming Communications Through Tested Processes
Two people doing AI in comms the right way: @KyleArteaga at @bulleitgroup and @DandotLewis at The Comms Stack. Not theory. Not vibes. Actual process, actual testing, actual answers. Kyle has been doing the work at Bulleit, testing wire services for...
Bitly Introduces AI-Powered Features to Simplify and Accelerate Marketing Analytics
Bitly has launched two AI‑powered features—Bitly Assist, a conversational chat assistant, and Weekly Insights, an automated performance‑highlighting report—aimed at speeding up marketers' data‑to‑action workflow. Assist lets users query link and QR‑code metrics and create assets through natural language, delivering answers...
Commure Launches AI-Powered Speech-to-Cursor Dictation Tool for Clinical Workflows
Commure has unveiled Commure Dictation, an AI‑powered speech‑to‑cursor extension that turns the company’s Ambient mobile app into a wireless microphone for clinicians. The tool eliminates the need for costly, dedicated dictation hardware and enables physicians to dictate notes, messages, and referrals...
M-Files and Microsoft Deepen Strategic AI Partnership
Document management firm M-Files announced new integrations with Microsoft 365 Copilot and the Copilot Agent Builder, embedding its context‑first repository directly into the Microsoft 365 experience. The partnership adds permission‑aware answers, metadata‑rich context, and seamless access to M-Files content alongside...
Found From Claude Code: Chapter 1: The Harness Paradigm
An AI harness is an infrastructure layer that sits between large language models and external systems, directing model outputs into safe, structured actions. It tackles five core challenges: constraining action space, managing conversation state, enforcing permissions, handling failures, and optimizing...
Review Intelligence as a GTM Signal Layer
Slang AI, a voice‑AI reservation platform, is using restaurant reviews as a direct go‑to‑market (GTM) signal layer to pinpoint eateries that lose bookings due to phone‑answering gaps. The company built a five‑tool pipeline—Serper.dev for Google Maps discovery, Clay as a...
Humanoids: Real Reasoning or Just Pattern Matching?
Humanoids look impressive, but are they reasoning or just pattern matching? Gill Pratt’s take might change how you see current demos. https://spectrum.ieee.org/humanoid-robots-gill-pratt-darpa?share_id=9330907
ChartHop AI People Platform Now Available on SAP Store
ChartHop AI Pro, the AI‑powered people‑operations platform, is now listed on the SAP Store after its debut at Transform 2026. The solution integrates directly with SAP SuccessFactors Employee Central, delivering unified workforce intelligence, headcount planning, and real‑time analytics. By linking HR...
Crafting Reliable AI Systems with the Right Data Engineering
The DBTA webinar highlighted that AI projects fail more often due to fragile data pipelines than model flaws. Speakers Kevin Hu and Jerod Johnson outlined how data engineering must evolve to support continuous, real‑time data, lineage, and repeatable outputs for...
Nigerian Founder Launches ADT, a New AI Model for Cyber Defense
Glemad unveiled Autonomous Defence Transformers (ADT), the first frontier‑scale AI models built from the ground up for security reasoning and autonomous cyber defence. The PulseADT service now safeguards over 680,000 assets, handling 1.8 million security events per second, and delivers a...
Polygraf AI Announces Core AI Patent and Sweeps Major Cybersecurity Awards at RSAC 2026
Polygraf AI announced a core USPTO patent for its Content Source Detection AI model and swept several top cybersecurity honors at RSA Conference 2026, including the Global InfoSec Award for Most Innovative AI Usage Control, a gold win at the...
Relyance AI Releases Lyo, Aims to Set a New Enterprise Data Security Standard
Relyance AI unveiled Lyo, an autonomous data defense engineer that continuously monitors AI agents' interactions with enterprise data. Leveraging the company’s Data Exposure Graph, Lyo maps AI‑to‑data relationships, flags over‑privileged access, and provides real‑time contextual alerts. The platform includes a...
Generative AI Shifts From Market Boom to Disruption Risk
Prometeia’s latest event‑study shows investor sentiment on generative AI turning sharply negative, especially in technology, financial and real‑estate sectors. Over a ten‑day window after AI‑related news, S&P 500 firms now post cumulative abnormal returns below expectations, with software firms lagging 0.75%...
GPU Power Draw Beats Scheduling as Utilization Metric
Without getting all the way down to performance counters, GPU power from nvidia-smi is a better indicator of true utilization than job scheduling or “gpu busy”. I would love to see animated “heat maps” of the big data centers, with...
AI Tools Simplify Complex Care Data, but Risks Persist
AI-driven platforms are increasingly helping patients decode complex medical records, offering simplified visualizations and personalized insights. However, experts like Leigh Burchell of Altera Digital Health warn that many of these tools lack HIPAA compliance, exposing sensitive health data to breaches....
5 Best Practices to Secure AI Systems
As AI becomes integral to critical operations, its unique attack surface outpaces traditional security measures. Experts recommend a multi‑layered strategy that starts with strict access controls and encryption, followed by AI‑specific defenses such as firewalls and red‑team testing. Continuous visibility...
AI Revolutionizes Communications Strategy for Greater Impact
How AI Is Transforming the Communications Industry — Lessons From FleishmanHillard AI is reshaping communications strategy and execution, and the insights from FleishmanHillard show how leaders can harness it for more impactful engagement. Read more 👉 https://lnkd.in/eHyXPmS4 #AI #Communications #Innovation...
AI Models Fail at Robot Control without Human-Designed Building Blocks but Agentic Scaffolding Closes the Gap
Researchers from Nvidia, UC Berkeley, Stanford and CMU introduced CaP‑X, an open‑access framework that evaluates how large language models control robots via self‑written code. Testing twelve frontier models—including Gemini‑3‑Pro, GPT‑5.2 and Claude Opus 4.5—across seven manipulation tasks revealed that without high‑level...
AI OS Replaces Social Media Manager, Handles Everything
You no longer need a social media manager. Neon is an AI OS that handles your entire account from DMs to daily content, end-to-end. → DMs managed completely on your behalf → Fresh niche-based inspiration every single day → Replies...
Managing Drift in AI Models and Data
AI models inevitably experience drift as real‑world data diverges from their original training sets, eroding accuracy and business value. The article distinguishes data drift—shifts in input distributions—and functional (concept) drift—changes in underlying relationships. It outlines detection methods such as direct...
The Hype Cycle Meets Malpractice Law: Why the Jobs Persist
Anthropic CEO Dario Amodei warned that half of entry‑level lawyers, consultants and finance professionals could disappear within five years as AI matures. The article counters that while large language models can draft memos and build models, liability and professional judgment...
HubSpot Moves to Outcome-Based Pricing for some Breeze AI Agents
HubSpot announced that, starting April 14, 2026, its Breeze AI Customer Agent and Prospecting Agent will shift to outcome‑based pricing. The Customer Agent will be charged $0.50 only for conversations it resolves, down from $1 per conversation, reflecting its 65%...
OpenClaw Vs. Hermes Agent: The Race to Build AI Assistants that Never Forget
Developers are frustrated by AI coding assistants that lose context after each session, prompting a split in the AI agent market between short‑lived tools and always‑on agents. OpenClaw and Hermes Agent represent the latter, offering persistent runtimes that remember past...
Navan Launches AI-Powered Audit Engine for Expense Risk Screening
Navan unveiled an AI‑driven Audit Engine that adds over 45 configurable checks to its travel‑and‑expense platform, enabling real‑time fraud, compliance and policy violation detection. The system automatically approves roughly 75% of transactions, directing finance teams to the remaining 25% that...
Mattoboard Introduces Design Stream: The First AI Concept and Curation Engine for Interior Projects Using Real Materials
Mattoboard launched Design Stream, an AI‑driven engine that converts interior briefs into fully visualized concepts using real, sourceable materials in seconds. The solution blends large language models with a proprietary material intelligence layer, delivering photorealistic renders and product specifications aligned...
AI Agents Will Soon Need Crypto Wallets via X402
Every AI agent deserves a crypto wallet. In fact, there will be more AI agents transacting online than humans very soon. x402 is the internet payments layer (which has been missing for the last 30 years), and will enable this. The new...
AI Generates Full Landing Page and Theme in 20 Minutes
I gave an AI design tool a rough brief and a screenshot of a website this week. What came back was a polished, interactive landing page. Not a rough wireframe. An actual design with relevant copy, hover states, and responsive layouts....
Strawberry Farming Leverages AI and Robotics
MISUMI Group has partnered with U.S. vertical‑farming specialist Oishii Farm to provide AI‑enabled robotics and mechanical components for its Ametalas indoor strawberry operation. The collaboration uses MISUMI’s extensive catalog and its Fictiv subsidiary to supply precision parts that automate climate...
OpenAI CEO Informs Disney CEOs About Sora Shutdown
OpenAI CEO Sam Altman called both former Disney CEO Bob Iger and new Disney CEO Josh D'Amaro to tell them about the decision to shutter Sora. "I get it," D'Amaro told Altman, the tech exec said in a new interview.
LLMs Erode Book Culture, Threatening Their Own Host
The value of LLMs "depends on our continued esteem for bookiness. But as they take over the tasks both of reading and (...) writing, they necessarily eat away at that esteem. They are a parasite that threatens to kill the...
Accelerating Drug Discovery with “Paradigm Shifting” AI Model
A multi‑institution team led by Michigan State University unveiled GPS, a machine‑learning platform that predicts how a compound will alter gene expression from its chemical structure. Trained on millions of transcriptomic measurements across more than 70 cell lines, GPS screened...
Local File-Based Memory Boosts Agent Accuracy Post-Leak
This is exactly what the ecosystem needed after the claude code leak. local, file-based memory for agents with top market accuracy, proven in real production. Crazy how fast open source moves right now.
US AI Training Risks Illegality, Undermining Global Dominance
if training AI isn’t fair use ↓ all US base models are now infringing & unlawful ...
Cyera Achieves FedRAMP High “In Process” Designation to Securely Accelerate AI Adoption
Cyera, an AI security platform, has earned a FedRAMP High “In Process” designation, moving it toward full federal authorization. The status reflects rigorous security reviews for handling Controlled Unclassified Information, positioning Cyera for government AI deployments. Its platform offers automated...
Altman and Ive Team Up on Cognitive Hardware
So, whatever happens to this "big" idea? Are Sam Altman and Jony Ive Creating Cognitive Hardware? https://t.co/G9DgJPeVg0
Havelock AI Parses Trump's Speech, Mostly Accurate with Minor Mislabels
Running Trump's speech last night through @Havelock_AI. Overall does very well. Not all span identifications are perfect. It flagged "Tonight, Iran's navy is gone" as a proverb, for example. But note the Self Correction in sentence 14, which is not...
How Schneider Electric Is Streamlining Onboarding with WalkMe
Schneider Electric has expanded its use of WalkMe, a digital adoption platform, from a pilot in North America to a global solution that embeds on‑screen guidance across its SaaS tools such as SAP S/4HANA and Salesforce. The rollout, backed by...
Fix Real-Time AI Voice Lag with Right APIs
Got a hitch in your real-time AI voice app? That little lag throws off the experience. @anniewangtech figured out what was going on, and did a deep dive into the right APIs and patterns to apply. GREAT post. https://t.co/Xz0J6M0tcL https://t.co/Rmp0xlERUq
Claude Code: A Revolutionary Breakthrough Rivaling Bitcoin
I’m not gonna lie Claude code has to be one of the most revolutionary things to happen to society since the creation of bitcoin
Wunderkind Launches New Braze Integration to Turn Identity Into Revenue-Driving, Orchestrated Customer Journeys
Wunderkind has launched a new integration with Braze that embeds its AI‑driven identity resolution and high‑intent behavioral signals directly into Braze Canvas. The connection lets marketers turn anonymous web visitors into addressable Braze profiles, grow email lists in real time,...
Established Firms Win by AI‑enhanced Products and New Ventures
Some of the biggest beneficiaries of AI will be established companies with a profitable business model that manage to leverage AI to make their existing products more compelling and even start new ones (like Adobe Podcast which is both new...
MAI Suite Launches on Foundry: Best Transcription, Voice, Image Models
We’re bringing our growing MAI model family to every developer in Foundry, including … · MAI-Transcribe-1, most accurate transcription model in world across 25 languages · MAI-Voice-1, natural, expressive speech generation · MAI-Image-2, our most capable image model yet Start building: https://t.co/Mls2y7nRQT
AI-Powered Cohorting Is Quietly Reshaping How Real-World Evidence Gets Built
AI‑powered cohorting is redefining how real‑world evidence is generated by replacing manual SQL‑driven processes with modular, workflow‑based automation. The new approach decomposes cohort construction into discrete steps—intent interpretation, clinical concept mapping, temporal reasoning, execution, validation, and explanation—each handled by specialized...
Submit AI Industry Tips Confidentially via Email or Signal
Got a tip or documents about the AI industry? From a non-work device, contact me confidentially (off the record) at rebecca.bellan@techcrunch.com or Signal: rebeccabellan.491.
Stop Waiting: AI Must Drive Immediate Sales
Rethinking the Sales Process: Why “Wait and See” AI is Failing Your Revenue Goals by @Timothy_Hughes https://t.co/7UsnSpKjGN @DLAIgnite #SocialSelling #DigitalSelling #Sales #SalesTips #SalesLeader #Salesforce #SalesEnablement #Marketing #Leadership #MarketingStrategy #ArtificialIntelligence
AI as Second Reader Boosts Breast Screening Accuracy
Impact of using artificial intelligence as a second reader in breast screening including arbitration https://t.co/15eScV9DfT