🚀 Sam Altman to buy rocket company

Good morning, AI enthusiast. New reports reveal Sam Altman is exploring plans to build a rocket company to rival Elon Musk’s SpaceX.

In today’s AI newsletter:

• Sam Altman considered buying a rocket startup

• US turns its attention to robotics

• Scientists find a way to “Recharge” aging cells

• How to generate stunning 4K photorealistic AI images

• AI tools & more

AI NEWS

🚀 Sam Altman Considered Buying a Rocket Startup

OpenAI held advanced talks to acquire or partner with Stoke Space, a fast-growing aerospace startup building fully reusable rockets. The plan: create orbital infrastructure for future AI data centers and reduce reliance on Earth-bound energy and cooling systems.

• Discussions included Altman raising billions for a controlling stake

• Stoke Space builds fully reusable rockets and is viewed as an emerging SpaceX challenge

• Partnership would have given OpenAI a direct entry into the launch industry

• Talks collapsed as OpenAI shifted focus to improving ChatGPT and its next-gen model “Garlic”

AI compute demand is exploding, and Altman sees orbit as the next frontier for energy-efficient, planet-scale data centers. While the rocket deal didn’t happen, OpenAI’s continued push into infrastructure, including custom chips and the upcoming Garlic model, shows how far it’s willing to go to stay ahead of Google and Anthropic.

ROBOTICS

🤖 US Turns Its Attention to Robotics

Five months after announcing its national AI plan, the administration is now moving to accelerate America’s robotics industry. Commerce Secretary Howard Lutnick has been privately meeting with robotics CEOs and is “all in” on boosting development, with an executive order being considered for next year.

• Commerce says robotics and advanced manufacturing are essential for reshoring critical production

• Department of Transportation preparing a new federal robotics working group

• Congress is exploring a national robotics commission, with more legislation coming

• Robotics seen as the next front in the US–China tech race, following AI

• China currently leads with 1.8M industrial robots, 4× the US, and already has a national robotics plan

With federal backing and military procurement, the U.S. could jumpstart a robotics boom, bringing humanoid workers into factories, logistics, and defense years earlier than projected, while shaping global leadership in next-gen automation.

SCIENCE

🧬 Scientists Find a Way to “Recharge” Aging Cells

Researchers at Texas A&M have discovered a technique that boosts young stem cells using tiny “nanoflowers,” triggering them to produce extra mitochondria, the cell’s power stations, and then transfer that energy to old, weak cells.

• Nanoflowers soak up harmful oxygen molecules, activating genes that ramp up mitochondrial production

• Supercharged stem cells donate 2× more mitochondria to aging or damaged cells than normal

• Smooth muscle cells (like those in the heart) grew 3–4× stronger

• In heart cells weakened by chemotherapy, survival improved significantly

• The process requires no drugs or genetic modification

This could become a major breakthrough in cellular rejuvenation. If scientists can safely scale this in humans, boosting this natural “battery-sharing” system might one day help slow, or even reverse, certain aspects of aging.

HOW TO AI

🗂️ How to Use Seedream 4.5 to Generate Stunning 4K Photorealistic AI Images

In this tutorial, you’ll learn how to generate perfectly consistent images using Seedream 4.5, a multimodal model designed for sharp details, stable characters, and clean artistic logic across every variation you create.

🧰 Who is This For

• Graphic designers producing clean commercial visuals

• Brand creators and advertisers needing perfect text & layouts

• Product photographers and e-commerce teams

• Anyone who wants fast, consistent, studio-quality images

STEP 1: Access Seedream 4.5

Head over to Imagen Art, since you can use it for free for some initial creations. Once you’re on the site, sign up and you’ll land inside the main Imagen interface.

At the bottom-center of the screen, you’ll see the main Prompt box. Directly beneath it is the Model Selector. Open the dropdown and choose Seedream 4.5 to make sure your generations use the newest version.

STEP 2: Generate Your Image

Now in the prompt box, describe your image in detail. For this tutorial, type:

"A portrait of an edgy woman with black hair and bangs, wearing yellow striped halter top dress, dark lipstick, shot on film, flash photography, moody lighting, low contrast"

Below the text box, under Aspect Ratio, select your preferred dimensions.

Once everything is set, click Generate next to the prompt box.

STEP 3: Refine and Enhance

Now click any generated thumbnail to open it fully. Below the image, you’ll see options:

Variation: creates matching alternates in the same style.
Expand” uses Magic Expand to widen the composition while keeping details intact.
Edit: opens the style and lighting controls for fine-tuning your look.

You can tweak tone, contrast, saturation, lighting style, or apply light stylistic adjustments. Every change updates instantly so you can shape your final output exactly how you want it.

STEP 4: Export Your Final Image

Once you’re satisfied with your result, simply download it from the image viewer panel. the tool maintains extremely clean edges, perfect text, and consistent detailing, making the export ready for presentations, campaigns, posters, brand kits, or product showcases.

You now have a polished, professional image created entirely from a prompt, powered by one of the strongest image models released this year.

Apple confirms interface design chief Alan Dye is leaving the company to join Meta; Tim Cook says Stephen Lemay, Apple's longtime designer, will replace Dye.

Mark Zuckerberg announces a new creative studio in Meta's Reality Labs led by Alan Dye, who will be joined by Apple design lead Billy Sorrentino

Google launches Workspace Studio, a no-code tool to create, design, manage, and share AI agents, for users on Business and Enterprise Workspace plans.

Dario Amodei suggests some AI companies are taking on too much risk by “YOLO-ing” and committing to spend hundreds of billions on data centers

🐳 DeepSeek V3.2: DeepSeek’s new powerful open-source model.

🎬 Kling O1: A video model that can understand and edit videos.

🍌 Nano Banana 2: Google’s Gemini 3 image model with perfect text and character consistency

🤖 Claude Opus 4.5: Anthropic’s new flagship model for coding, agents, and real computer use

THAT’S IT FOR TODAY

Thanks for making it to the end! I put my heart into every email I send, I hope you are enjoying it. Let me know your thoughts so I can make the next one even better!

See you tomorrow :)

- Dr. Alvaro Cintas