Good morning {{first_name| Artificial intelligence lovers}}. Two months ago, OpenAI’s Fidji Simo told employees that the company was in “code red” due to the rise of Anthropic. New spending data suggests the leaderboard has indeed flipped.
Ramp’s latest AI Index just showed leading human adoption among paid business users for the first time, a 4x rise in adoption since 2025 that may tell the story of why OpenAI will radically shift its priorities throughout 2026.
Note: Our next live workshop, “Finally Getting AI to Do Real Work,” will be held today at 2 PM EST. Join for details on how AI will actually work in 2026 and the contextual habits that feed users at the top to get reliable output from any model. RSVP here.
-
Anthropic takes over business AI leadership in OpenAI
-
Amazon is doubling down on Alexa+ for shopping
-
Create content with Claude Code and Higgsfield
-
Adaption automates AI training with AutoScientist
-
4 new tools for AI, community workflow, and more
Anthropic and Opinai
Rundown: Financial technology company Slope only published Its latest AI Index shows Anthropic taking the lead in paid work adoption of OpenAI for the first time, an enterprise boom whose usage has quadrupled over the past year while OpenAI has stabilized.
-
Ramp tracks corporate card payments and invoices from more than 50,000 US businesses, making this a spending signal, not full market share.
-
Anthropic adoption rose 3.8% in April to 34.4%, while OpenAI adoption fell 2.9% to 32.3% as overall AI adoption continued to rise to 50.6%.
-
Claude Code helped create many swings, as Anthropic expanded from technical teams into financial, legal, and research workflows.
-
Ramp highlighted several risks facing Anthropic despite this trend, including recent Claude service outages and increased costs compared to OAI and open source.
Why it matters: OpenAI wasn’t suddenly cooked, as Ramp didn’t track some of the larger enterprise deals and ChatGPT remained the larger consumer brand. The yearly trend speaks louder, and it’s likely the same type of chart that caused OAI’s Fidji Simo to lead Axiswhich has become clear through recent Codex and other institutional pushes.
TOGETHER WITH GOOGLE CLOUD
Rundown: Scale your AI operations with Gemini Enterprise Agent Platform, Google’s new evolution for creating, managing, and optimizing AI agents. Using the open source Agent Development Kit (ADK), this hands-on code lesson teaches you how to orchestrate specialized agents such as researchers and judges into self-correcting workflows that solve complex, enterprise-wide problems.
In this hands-on programming lesson, you will:
-
Build self-correcting AI feedback loops.
-
Coordinate task handovers between specialized agents.
-
Deploy production-ready workflows to Google Cloud Run.
Amazon
Rundown: Amazon Folded Its autonomous shopping chatbot Rufus is housed within Alexa for Shopping, a new agent that takes over Amazon search and follows shoppers across devices with a shared memory of past purchases, preferences and conversations.
-
Rufus attracted over 300 million users in 2025 while still in beta, thanks to its knowledge of products and shopping history that now feeds Alexa for shopping answers.
-
Amazon says the new assistant relies on catalog data, reviews, delivery timings, previous purchases and Alexa conversations for information.
-
Alexa can now ask questions in the search bar, perform side-by-side comparisons, track prices, and auto-purchase items when prices reach a target.
-
The new Buy for Me feature handles checkouts at non-Amazon stores, with the ability for scheduled actions to automatically restock products according to a cadence.
Why it matters: RIP to Rufus, but consolidating under one Alexa AI brand sounds like a better game, and the huge amount of customer history Amazon provides gives its agent the right moat to work with. But with a consumer base already moving to other AI platforms for their shopping needs, will Alexa play well and integrate with them?
Artificial intelligence training
Rundown: In this guide, you’ll learn how to connect Higgsfield to Claude Code using the Higgsfield CLI, and then use Claude Code to send a single image vector to multiple AI image models simultaneously.
-
Create a new project folder, install the Higgsfield CLI (npm install -g @higgsfield/cli), authenticate (higgsfield auth login), and add a Higgsfield skill (npx add-skills higgsfield-ai/skills) for Claude Code
-
Open Claude Code in the same folder and have him scan it higgsfield.ai/cliCheck the installation, and list the sample images available
-
Give Claude Code an image prompt and have it run it through six models, save the output in the higgsfield model test folder, and create a Compare.MD File with notes for each result
-
Choose the best direction, then ask Claude Code to refine the winning claim or perform another comparison
Pro Tip: If setup becomes confusing, ask Claude Code to check Node/npm and provide detailed instructions. Higgsfield also supports video, so you can try that with short clips as well.
Introduction of the tape
Rundown: AI companies are generating faster rates of revenue than ever before, but knowing how to monetize AI products is still one of the most difficult problems founders face. Stripe interviewed teams at Anthropic, Vercel, Clay, and more to build a five-step pricing framework for AI products.
Download the guide to learn best practices on how to:
-
Choose the appropriate pricing model
-
Prevent billing surprises
-
Improve your pricing over time
Adaptation
Rundown: Adaption, the AI startup from former Cohere Vice President of Research Sarah Hooker, just foot AutoScientist, a new system that automatically customizes AI models to specific jobs by fine-tuning what the model learns and how it learns.
-
AutoScientist tests different training data and settings, then iterates until the model meets the user’s goal.
-
In internal tests, AutoScientist outperformed its expert-tuned models by 35% on average, with success rates jumping from 48% to 64%.
-
Results were achieved across multiple AI models, a wide range of dataset sizes, and 8 diverse industries, including financial, legal, and medical.
-
Initial adaptive data for adaptation He releases In February, the goal was to increase the quality of datasets, with Autoscientist now moving toward model customization.
Why it matters: A few thousand people in the world know how to properly train and tune the frontier model, and almost all of them work in the same group of laboratories. If a tool like Autoscientist can begin to automate this expertise, creating custom models for individual companies and use cases could become more practical.
Nvidia become The first company to reach a market value of $5.5 trillion, it comes as CEO Jensen Huang arrives in China to join US President Donald Trump in meetings with Xi Jinping.
Sam Altman to attest In the legal battle between Elon Musk and OpenAI, Musk’s “specific plans on safety” made him concerned, and he suggested passing the company on to his children.
The indescribable intelligence of David Silver and Nvidia Announce Partnership to build training pipelines for RL agents, with early work targeting Nvidia’s Vera Rubin hardware.
Microsoft foot MDASH, an AI security tool that connects more than 100 specialized bug-hunting agents, detects 16 flaws across Windows.
AI Safety Institute in the UK He said The ability of AI to complete cyberattacks is doubling every few months, with Mythos Preview and GPT-5.5 ending simulated breaches.
In each newsletter we show how the reader is using AI to work smarter, save time or make life easier.
Today’s workflow comes from the reader Rod R. In Elizabethtown, Kentucky:
“I’m an avid cyclist. I turned 61 this year, but I still participate in multi-day cycling events. I needed advice on what kind of riding I should do to prepare, and I thought I could benefit from a coach, but I can’t afford one. Then it occurred to me to try ChatGPT.
This worked tremendously! We had a “conversation” about my goals, medical history, current fitness, event dates, etc. I have been following the training plan for about 6 weeks.
ChatGPT continues to coach me as I report on each trip, helps me restructure the plan when life requires me to tweak it, and even remembers things we’ve already discussed. I love how this has helped me prepare, and I have peace of mind knowing what I need to do each day. Today is a rest day! 😊“
How do you use artificial intelligence? Tell us here.
That’s all for today!
Before you go, we’d love to know what you thought of today’s newsletter to help us improve The Rundown experience for you.
Rowan, Joey, Zack, Shubham and Jennifer – the humans behind The Rundown
