In partnership with
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
.
TODAY IN AI
3 things that happened while you were busy
1. Meta storms back to the frontier with a model priced to undercut everyone. Zuck and company are back in the game with Muse Spark 1.3, a new model that puts Meta on par with the frontier offerings from Anthropic and OpenAI. It is currently free on OpenCode, with API pricing far below comparable models. Zuckerberg's pitch: "frontier performance almost too cheap to meter."
2. Google ships a Workspace image editor and its third Flash model in six weeks. Google Pics generates Nano-Banana-quality images with precise editing, so you can adjust the text or a single element without regenerating the whole picture. It is generally available across Workspace, AI Pro, and Ultra plans. Alphabet also debuted Gemini 3.8 Flash, a cost-efficient reasoning model and its third Flash release in six weeks.
3. Claude learns to run your computer while you do something else. Claude Cowork and Claude Code can now work in the background to finish tasks on your machine while you get on with your day. The feature is in beta on Pro and Max plans on macOS, and it only touches apps you have explicitly granted access to. Anthropic also published a prompting guide for Fable 5.1 with instructions for making Claude write more clearly.
FROM THE FRONTIER
The AI boom runs on Nvidia. The attacks now come from every direction.
The setup. Three years ago Nvidia booked roughly $7B in quarterly revenue. Now it books $96B and is on track to command 75% of the AI accelerator market in 2026, with its GPUs still training and running almost every large model. That kind of position paints a target on your back.
The challengers. Chip startups had raised $8.3B as of April, and their pitch is simple: GPUs were never purpose-built for AI, so leaner designs can cut both cost and energy. It is the same argument GPUs once made against CPUs.
Friends turned rivals. OpenAI says its first chip Jalapeño posts industry-leading numbers, Anthropic is hiring hardware executives, and chipmaker Cerebras went public in May to fund the fight. Google, Amazon, and Microsoft are all shipping in-house silicon too, on top of AMD and Broadcom.
The China cliff. Nvidia controlled 66% of China's accelerator market in 2024. Analysts now have it headed for 8% in 2026 as Cambricon and Huawei push the country toward self-sufficiency. That is a major market effectively walking away.
The moat. Nvidia's real defense is CUDA, the software layer that over 4 million developers build on, and the overall market is growing faster than any single challenger can absorb. So the thing to watch is not who ships a faster chip. It is who makes CUDA optional.
IN THE KNOW
What people are actually watching and sharing
Speedier gadgets. Hobbyists are using AI to upgrade the tech they already own instead of replacing it. Two viral posts claim Claude improved their Wi-Fi speeds and de-bloated a sluggish TV (1M views).
AI for seniors. A new smart speaker acts as a subtle caretaker for elderly relatives. No screen, no setup: plug it in and a voice agent offers daily companionship while quietly keeping tabs on their health (1M views).
Older than you think. AI dates back 70 years and has already survived two industry winters, in the 1970s and 80s. Quartz rounded up 18 more facts most people get wrong about the industry.
Prompt audit. AI educator Peter Yang suggests having Fable 5.1 run a prompt audit to strip outdated rules and redundancies from your workflows, ideally every time a strong new model drops.
Going small. Shopify CEO Tobi Lütke says his strategy for beating frontier models is training small models on niche tasks with a self-improving flywheel, and he posted the internal platform Shopify built to do it.
PROMPT STATION
Turn any photo into embroidery art
This one runs in ChatGPT Image 2.0. Upload a photo, paste the prompt, and get the same scene rebuilt entirely from stitches, yarn, and thread loops, down to embroidered clouds. It keeps faces and poses intact, which makes it a sneaky-good personalized gift generator. Try it on a family photo first.
COPY AND PASTE THIS PROMPT ChatGPT Image 2.0: Use the uploaded photo as the exact reference for the composition, pose, facial expressions, framing, camera angle, and overall arrangement. Preserve the identity, clothing, hairstyle, and background placement exactly, but transform the entire image into a luxurious handcrafted embroidery artwork. Every visible element—including skin, hair, clothing, smartphone, sky, clouds, trees, and background—must be recreated entirely from thick yarn, woven fibers, dense stitches, punch-needle embroidery, tufted textile, and layered thread loops. Ultra-detailed, handcrafted textile illustration, premium embroidery art, photorealistic thread detail, highly tactile surface, masterpiece quality, 8K. |
Make it yours: The uploaded photo is the only variable, so the prompt works as is. To change the craft, swap "embroidery" and "punch-needle" for "cross-stitch," "needlepoint," or "tufted rug," and drop the final "8K" line if you want a softer handmade look instead of hyper-detail.