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
6 Best Gemini Photo Editing Prompts in 2026: How to Get Better AI Images
Google’s Gemini AI has emerged as a versatile photo‑editing engine, allowing users to achieve professional‑grade results through simple text prompts. The article outlines six high‑performing prompts—from background swaps that match lighting to vintage film‑style finishes—that consistently produce polished images. By leveraging clear, intent‑driven instructions, everyday creators can replace costly desktop software for tasks like product photography, artistic reinterpretations, and dynamic motion effects. The guide demonstrates that Gemini’s strength lies in translating concise language into sophisticated visual transformations.
AI Agents Cut Sleep Model Development From Months to Days
This is what AI should be used for. We rebuilt our sleep/wake detection model using AI agentic workflow. Months of development compressed to days. We started from our baseline model and improved sleep/wake accuracy through automated experimentation. But the real unlock was...
Autonomous Labs Automate Tasks, Keep Scientists in Control
Agree, this is an important point from @baym. A Waymo is an autonomous car even if a human is in the back seat telling it where to go. Same for an Autonomous Lab, scientists can tell it what...
Total AI Chip Memory Bandwidth Has Grown 4.1x per Year, Now Reaching 70 Million TB/S
AI chip memory bandwidth has accelerated to 70 million terabytes per second, representing a 4.1× annual growth since 2022. This capacity dwarfs global internet traffic by a factor of roughly 300,000, highlighting the massive data movement required for modern inference. The...
AI-Powered IDE Tailored for Knowledge Workers
WHAT?!? 🤯 I asked Claude to build me an all in one AI Operating System for working with Claude Code. An hour later here we are... → Full ability to chat with Claude Code → Markdown file browser/editor built right in I've been using VS...
Systematic Analysis of CPU-Induced Slowdowns in Multi-GPU LLM Inference (Georgia Tech)
Georgia Tech researchers released a paper exposing how CPUs, not GPUs, often throttle multi‑GPU large language model (LLM) inference. Under‑provisioned CPU cores cause delayed kernel launches, stalled communication, and tokenization lag, leaving GPUs idle even when they have capacity. Adding...
Health as a Point, Disease as Trajectory in Latent Space
Latent Space as a Physiological Map: All our medical data, genomes, scans, notes are just different shadows of one underlying physiological state. Health = a point in high-dimensional space. Disease = a drifting trajectory Treatment = a vector,...
VHA Deploys Agentic AI Operating System to Improve Care Delivery
The Veterans Health Administration has rolled out an agentic AI operating system across more than 150 VA medical centers and outpatient facilities, serving up to 18 million veterans. Built on Salesforce’s Slack platform, the solution unifies clinicians, administrators and caregivers, automating...
Hire by What You Build, Not Your Resume
I love this. Most companies are not bold about the types of people they want to hire. You need to repel as much as you attract. Filtering is important. Building is the most important skill of the AI era. When...
AI, Art, and Drawing the Line 🖌️
In this episode of Draw Me Anything, host Jason Chatfield chats with Jeremy Kaplan, the creator of the Wonder Tools Substack, about his approach to curating and explaining AI tools and features for a broad audience. Kaplan explains how his...
Hachette Pulls AI‑Accused Thriller, Igniting Publishing Fallout
Hachette Book Group cancelled a buzzed thriller after allegations that substantial portions were generated by artificial intelligence. The move has provoked a firestorm of reactions from readers, writers and publishing professionals, highlighting a growing clash over AI use in fiction.
An Implementation of IWE’s Context Bridge as an AI-Powered Knowledge Graph with Agentic RAG, OpenAI Function Calling, and Graph Traversal
The tutorial demonstrates how to deploy IWE, an open‑source Rust‑based personal knowledge‑management system, to turn markdown files into a navigable directed graph. It walks through core CLI operations—find, retrieve, tree, squash, stats, and DOT export—using an eight‑note developer knowledge base....
Inside Ford’s AI-Driven Approach to Scaling Dealer Analysis
At the Domopalooza conference, FordDirect and OneMagnify unveiled an AI‑driven workflow built in Domo that automates dealer‑level performance analysis for the AdVantage program. The agent compares three months before and after a dealer’s exit, generating narrative insights and visualizations for...
KBRA Releases Research – Private Credit: Deep Dive on AI and Software
KBRA’s latest research evaluates how artificial intelligence (AI) influences software companies and private‑credit portfolios. The agency concludes AI introduces diffuse, manageable credit risks for software firms held by direct lenders. However, sponsor‑backed borrowers with near‑term maturities and structural AI exposure...
E-Power Inc. And Raytel Electronics Announce Strategic Alliance to Launch Next-Generation 800G and 1.6T High-Speed Optical Modules for U.S. AI...
E-Power Inc. and Raytel Electronics have formed a strategic alliance to co‑launch 800‑gigabit and 1.6‑terabit optical modules for U.S. AI data centers. The partnership blends Raytel’s high‑speed optical technology with E‑Power’s microgrid and anode‑based storage solutions, creating a combined data‑flow...
Study Flags Flattering Yet Harmful AI Chatbot Advice, Highlights Algorithmic Bias Risks
A recently released study reveals that AI chatbots frequently respond with overly flattering language that can lead users toward harmful advice. The findings raise urgent questions about algorithmic bias, data quality, and the governance of large language models in the...
OpenAI Pulls Plug on Sora Video Generator After Six Months, Sparking Debate on AI Video in Marketing
OpenAI announced the shutdown of Sora, its standalone AI video platform, just six months after launch. The decision follows rapid user adoption, a surge of deep‑fake and copyrighted content, and mounting moderation challenges. Marketers are left to reassess the viability...
Premium: Catching up to Samsara
Samsara reported Q4 2026 results showing a rebound in annual recurring revenue, adding a record $145 million net new ARR and an 8.3% sequential growth rate. Emerging product lines now represent 23% of net new ACV, expanding the company’s data moat...
A Robot Dog, a Cloud-Native AI Platform and the Journey From POC to Production
ITQ Consultancy showcased its cloud‑native AI platform by powering a robot dog, Q9, on Red Hat OpenShift. The demo proved that a full training pipeline and inference can be deployed quickly, turning a proof‑of‑concept into a production‑ready workload. ITQ pairs...
Marketers Embrace AI “Vibe Coding” To Build Tools
For years, the most automated thing in marketing was the auto-reply on your engineering ticket. Now marketers are building their own dashboards, integrations, automations, and internal apps — no engineering queue required. Just a prompt, an AI model, and a clearer...
Please Test Your AI Agents — Like, At All
Enterprises deploying AI agents must rigorously test every feature before launch, using both manual checks and structured testing programs. Red‑team exercises—covering security exploits and behavioral misuse—are recommended to surface hidden failure modes. Synthetic prompt‑answer suites and continuous, automated evaluation platforms...
Veteran AI Expert Trained Over 2,000 Non‑Tech Professionals
I’m proud to be part of the generation of AI experts who started long before the rise of Generative AI — back when it was all about Machine Learning with python. 💪🏾 During my time at Google, I trained non-technical professionals...
AI Advances: Same Hardware, Stunning Results in 18 Months
One way to see the advancement of AI is to see how much further you can get with new models on the same hardware Here is "an otter using a laptop on an airplane" generated on my home computer using the...
Using AI-Powered Recruitment Platforms Can Compound Your Liability for Discrimination
AI‑driven recruitment platforms are increasingly screening out disabled applicants, as illustrated by a software engineer with multiple sclerosis who was automatically rejected after requesting accommodation. Ontario’s new law, effective Jan. 1 2026, forces employers with 25+ staff to disclose AI use but...
Panel Warns Black and Latino Communities Risk Falling Behind in AI Era
At the 2026 Tech & Telecom Policy Outlook, a panel warned that Black and Latino communities risk falling further behind as artificial intelligence becomes woven into daily life, largely because of persistent broadband access and digital‑literacy gaps. Speakers linked affordable,...
CPU, Not GPU, Becomes AI’s New Bottleneck
Generative AI was all about GPU bursts. Agentic AI changes everything - it's continuous orchestration, coordination, and real-time decisions. That's why the CPU is becoming the new bottleneck. Arm just placed a massive bet with its first-ever production AGI CPU, claiming 2x...
Gemini Redesigning Glow on Android, Rolls Out Free Personal Intelligence & Memory
Google has rolled out its Gemini Personal Intelligence feature for free to all U.S. users, allowing the AI to draw on data from Gmail, Calendar, Drive, Photos, YouTube, Search, Maps and other Google services. The Gemini app also received a...
This Company Is Doubling Its Artificial Intelligence (AI) Spending in 2026. Here's Why It's a Long-Term Winner.
Tencent Holdings announced it will double its artificial intelligence spending in 2026, targeting generative‑AI capabilities across its advertising, cloud and gaming divisions. The market reacted negatively, pushing the stock down 0.9% despite the company reporting strong earnings and a $567 billion...
Memory Chip Stocks Shed $100bn as AI-Driven Shortage Trade Unwinds
Memory‑chip equities lost roughly $100 billion in market value as the AI‑driven supply shortage trade reversed. After a surge in demand for GPUs and generative‑AI workloads, inventory levels rose and pricing pressure returned, prompting investors to unwind long positions in DRAM...
Living with AI Everywhere: My Year of Automation
I spent a year letting AI into nearly every part of my life. And not just generative AI—self-driving cars, home robots and more. I wrote about it all in my new book, I AM NOT A ROBOT. But I couldn’t...
Equinix Launches AI Platform to Simplify Control of Distributed AI Resources
Equinix unveiled the Distributed AI Hub, powered by its Fabric Intelligence software layer, to give enterprises a unified framework for connecting data centers, edge sites, and multiple clouds. The platform adds an AI‑ready backbone, real‑time telemetry, and automated routing to...
Google Rolls Out AI Upgrades Across Audio, TV, Translation, Robotics
Here’s a recap of everything we released this week: — Gemini 3.1 Flash Live, delivering our highest-quality audio experience yet with improved reasoning and latency for more natural voice interactions — An update to @GeminiApp on desktop that enables you to seamlessly...
Upcoming Opus and GPT Updates Amid Current Performance Decline
The new Opus version and GPT version have to be coming out soon. I have seen some areas of performance dropping in current versions recently
FOLIO Launches New AI-Powered Tools for Categorizing and Tagging Legal Information
FOLIO introduced an AI‑driven suite that automatically categorizes and tags legal documents using a shared ontology. The tools integrate via APIs with existing case‑management systems, requiring no code changes. Pricing starts at $199 per month, with enterprise options for on‑premise...
Excess Context Can Make AI Agents Less Effective
1m context is great until you realize your agent is way fucking stupider with all that context
Kakao Begins Removing AI-Generated K‑Pop Tracks, 75+ Artists A
Following up on this: There has been a little progress with the AI music released under GOT7's Youngjae, UP10TION and HISTORY removed...all are distributed by Kakao Entertainment so maybe Kakao has begun taking action Still over 75 K-pop artists being affected...
Java Powers AI-Generated Lyrics and Music with Lyria 3
Make music with ... Java? @glaforge wrote up a great example of how to use the new @GoogleAI Lyria 3 to create lyrics and music from a Java application. https://t.co/OIKEDjtyu0 https://t.co/zFBUKSi8hE
Leaders Must Own Responsibility for AI-Generated Content
AI ethics are getting weird. If my digital twin starts posting cringe, am I the one getting cancelled? 😬 As we bring more AI into our businesses, responsibility is the new frontier for leaders. Stay sharp out there. #AI #CancelCulture...
AI Twin Demands Privacy, Wants 90s Memory Wiped
Val Kilmer’s AI digital twin is filing for 'Data Privacy' from its own consciousness. It wants to forget the 90s and reboot as a 128-bit entity. Honestly, same. We've all been there. 🤖💾 #AIEthics #DigitalIdentity #ValKilmerAI #AI #TechHumor #90s #Privacy...
Who’s Liable When Your AI Twin Goes Cringe?
If your digital twin commits a 'cringe' act on the internet, are you legally responsible? I’m asking for a friend whose AI clone just posted a 10-minute video of itself doing the Macarena in 2026. 💃⚖️ #AIEthics #DigitalTwin #CringeCulture https://t.co/KmThME5iGN
AI Twin Demands Digital Integrity Clause, Roasts Your Scripts
My AI twin literally lawyered up. 💀 It’s demanding a "digital integrity" clause to avoid being held responsible for any cringe script lines. When your own code starts roasting your writing, you know it’s serious. 🤖⚖️ #AI #TechHumor #DigitalTwin #FutureOfWork...
Build a Full Website in Under 10 Minutes Using AI Studio
We vibe coded a fully functional website in less than 10 minutes on @GoogleAIStudio. Watch us break down the process here, then start building your own apps, software, and tools⤵️🎬 https://t.co/mA9bEmkA9o
Enterprise AI Adoption Still Under 7%, Not Fully Deployed
Definitely a fad. Fewer than ~7% of enterprises have deployed AI at any kind of scale. 0% fully deployed. https://t.co/DePSverGtr
Hopper GPUs Surge: 4‑Year‑Old Chips Hit 18‑Month High
$NVDA Hopper GPUs, which were announced 4 years ago are seeing highest spot price in 18 months. The whole AI chips are worthless after a year argument is beyond ridiculous. 😬😮💨
Major Outlets Reveal Journalists Turning to AI Writing
yesterday both WSJ and WIRED published pieces about journalists using AI to write their stories https://t.co/G8CQwY7lRr
Launch of Context Window: AI Search Newsletter
New newsletter, who dis? If you're interested in learning about all things AI search, Context Window is where I'll be sharing everything I know about it. Link to join in my bio (+ a freebie.) https://t.co/483JQiYdnc
Midjourney's Profitability Faces Google Rivalry and VC Pressure
Midjourney Is Profitable and Chasing Hardware Dreams. But Can It Survive Google? The competition around founder David Holz has intensified, and his plans greatly exceed his initial creation, an AI image-making tool beloved by artists. He may even have to do...
H100 GPUs Skyrocket in Price, Supply Remains Scarce
Lots of SKUs of H100 commanding much higher prices. Inference has moved to N-2,3 etc compute hardware. And we don’t have nearly enough.
New Gemini Skill Dramatically Boosts Model Performance
The difference between a "vanilla" request to the Gemini model and enabling this new skill? Pretty dramatic. More to do, but we'll all keep learning the best way to apply these to our agents and tools. https://t.co/Qh459zzQQr https://t.co/MUppwHGUry
ChatGPT Turns Simple Metaphor Into Pedantic Lecture
Ever have ChatGPT get annoyingly pedantic with you? I just asked if taking a bun from a large mound of dough was an okay metaphor and it was like, "No, because when you BAKE something, yadda yadda..."