Good morning, { AI enthusiasts }. While OAI was launching a video app, constructing an online browser, designing hardware, and adding shopping to ChatGPT, Anthropic quietly took control of the enterprise market.
Now, Applications CEO Fidji Simo is reining the corporate back in — telling staff they “cannot miss this moment because we’re distracted by side quests” and refocusing on two key pillars: coding tools and business customers.
In today’s AI rundown:
-
OpenAI scrapping ‘side quests’ to catch Anthropic
-
Mistral opens its model-training playbook
-
Generate a cohesive e-commerce product shoot
-
Microsoft redraws its AI org chart
-
4 recent AI tools, community workflows, and more
LATEST DEVELOPMENTS
OPENAI
Image source: Lovart / The Rundown
The Rundown: OpenAI is overhauling its product technique to deal with coding tools and businesses after CEO of Applications, Fidji Simo, called Anthropic’s enterprise dominance a “wake-up call” in a company-wide meeting, based on the WSJ.
The main points:
-
Powerful Claude Code and Cowork releases grabbed the lead with business customers, with Simo telling staff OAI is treating the gap as a “code red.”
-
Simo said OAI “can’t miss the moment because we’re distracted by side quests”, coming amid efforts including hardware, adult mode, ads, and more.
-
OAI’s 2025 launches included Sora, the Atlas browser, e-commerce features, and more, which insiders said led to confusion and constant compute shuffling.
-
It did claw back in coding, with Codex quadrupling its weekly users to 2M+ since January — alongside a brand new GPT 5.4 model targeting business workflows.
Why it matters: The Pentagon drama should still be fresh in consumer minds within the battle between OpenAI and Anthropic, but where the actual war is being fought is on the enterprise side. OAI is pulling in 1,000,000 different directions, and Simo saying so out loud to the entire company tells you ways real the Anthropic gap has gotten.
TOGETHER WITH HUBSPOT
The Rundown: HubSpot’s free, comprehensive “The right way to Use ChatGPT at Work” guide provides 100+ ready-to-use prompts to assist professionals boost efficiency and adopt AI-driven workflows.
Inside, you’ll find:
-
A fast crash course to master ChatGPT in under half-hour
-
Practical industry use cases to spark real-world inspiration
-
100+ prompts to streamline tasks and speed up productivity
-
Expert tricks to tackle common AI roadblocks with confidence
Get your free copy and join 10,000+ professionals leveling up with AI.
MISTRAL
Image source: Mistral
The Rundown: Mistral launched Forge, a platform that hands enterprises the identical training recipes and infrastructure the French AI lab uses internally — allowing corporations to construct custom models on proprietary data without ever sharing it.
The main points:
-
Moderately than basic fine-tuning, Forge offers full pre-training, post-training, and RL pipelines that mirror how Mistral builds its own flagship models.
-
Training can run entirely on an organization’s own servers with zero data exposure to Mistral, a tough requirement for defense, finance, and government buyers.
-
Early partners include ASML, Ericsson, and the European Space Agency, with use cases from legacy code migration to ancient manuscript restoration.
Why it matters: Most major enterprise AI boils all the way down to the identical thing: take a general model and hope it’s close enough. Mistral is making a unique bet — that corporations sitting on tons of proprietary data, compliance rules, and internal codebases need models trained on that knowledge, not only prompted with it.
AI TRAINING
The Rundown: On this guide, you’ll learn how you can generate a full set of product photos using Grok’s Aurora model, constructing each image off the last one so that they all appear to be they got here from the identical session.
Step-by-step:
-
Go to grok.com, upload a product photo, and prompt: “Product photo of [your product] on a matte black surface, soft box lighting from the upper left, shot on 85mm lens, f/2.8, shallow depth of field, business product photography”
-
Construct more shots from the primary output by keeping the identical studio description and changing only the angle or crop — a 3/4 angle, a close-up, or a white BG
-
Add a way of life shot. Describe the product in use: being worn, held, or placed in an actual environment. Include the setting, lighting, and mood
-
Lay out all of your images. They need to appear to be they got here from the identical shoot
Pro tip: Try picking a themed studio shoot, like basketball-themed, nature-themed, etc. Grok might surprise you here.
PRESENTED BY UNWRAP
The Rundown: Unwrap’s customer intelligence platform brings all of your customer feedback (surveys, reviews, support tickets, social comments, etc.) right into a single view, then uses AI + NLP to surface probably the most actionable insights and deliver them straight to your inbox.
With Unwrap, you get:
-
A transparent, single view of customer sentiment
-
The power to ask questions on your feedback using natural language with Assistant
-
Real-time alerts out of your feedback as they arise
-
Tools trusted at scale by Perplexity, Oura, Stripe, Clay, DoorDash, lululemon, Southwest Airlines, and others
Connect with Unwrap to get a free trial of the tools, exclusive to The Rundown AI readers.
MICROSOFT
Image source: Microsoft AI
The Rundown: Microsoft just overhauled its AI org chart, announcing the merge of its fragmented Copilot teams and shifting Microsoft AI CEO Mustafa Suleyman’s focus squarely towards its five-year mission to construct superintelligence in-house.
The main points:
-
Former Snap exec and recent Microsoft AI EVP Jacob Andreou will run the combined Copilot org, which can span across design, product, and engineering.
-
Suleyman said the move will “enable me to focus all my energy on our Superintelligence efforts”, with a deal with enterprise systems.
-
A reworked OAI partnership cleared the best way for Microsoft to construct toward AGI by itself, lifting a ban on solo development that ran through 2030.
-
Copilot continues to be struggling for traction, with 6M every day users in February vs. ChatGPT’s 440M — and its enterprise add-on reaching just 3% of Office subs.
Why it matters: Microsoft stock is down this 12 months, the legacy software corporations are under pressure to prove AI ROI, and Copilot adoption is a fraction of the large players. This reorg is Nadella betting that the fix starts on the model layer just as much because the product one — and that the corporate needs its own frontier systems to compete.
QUICK HITS
-
🧠 Adapt – Your organization’s AI computer. Learns your corporation and autonomously reasons and acts across your stack in Slack*
-
🤏 GPT-5.4 Mini & Nano – OAI’s fast, low cost small models for coding agents
-
🦞 NemoClaw – Nvidia’s open-source security layer for OpenClaw agents
-
🧠 Mistral Small 4 – Mistral’s recent AI combining reasoning, coding, and vision
OpenAI launched GPT-5.4 mini and nano, two smaller, faster versions of its flagship model built for coding assistants and multi-agent systems.
Mistral released Small 4, an open-source model that merges its reasoning, coding, and vision capabilities into one system.
Anthropic unveiled Dispatch, a Claude Desktop feature that lets users message the assistant from a phone as it really works on a PC, running code, browsing, and managing files.
Sam Altman’s proof-of-personhood company World launched AgentKit, a tool that lets web sites confirm an actual human is behind an AI shopping agent’s purchases.
Google announced that its Personal Intelligence feature is now rolling out to free-tier users across its AI Mode in Search, the Gemini app, and Chrome within the U.S.
Gamma introduced Imagine, an AI design tool baked into its presentation platform that generates logos, infographics, and social graphics with automatic brand styling.
COMMUNITY
Every newsletter, we showcase how a reader is using AI to work smarter, save time, or make life easier.
Today’s workflow comes from reader Dmitry K. in Livonia, MI:
“I actually have two vehicles and located it difficult to maintain meticulous maintenance records over time. Recently, I arrange a Claude project to create a Vehicle Maintenance Logbook. Each time I actually have a service appointment, I take an image of the receipt, including the work performed, and upload it to Claude.
It then enters all relevant information into my logbook, while keeping me informed of manufacturer-recommended maintenance and giving me an overall health rating for every vehicle. Now I actually have higher maintenance clarity and a totally digital record, which can turn out to be useful when selling either automotive!”
How do you utilize AI? Tell us here.
-
Read our last AI newsletter: Nvidia’s big AI day at GTC
-
Read our last Tech newsletter: China greenlights business brain implant
-
Read our last Robotics newsletter: Travis Kalanick’s stealth robot play
-
Today’s AI tool guide: Generate a cohesive e-commerce product shoot
-
RSVP to our next workshop on Thursday: AI coding bootcamp pt. 2
That is it for today!
- ⭐️⭐️⭐️⭐️⭐️ Nailed it
- ⭐️⭐️⭐️ Average
- ⭐️ Fail
Login or Subscribe to participate
See you soon,
