👀 OpenAI’s secret device leaks

Good Morning! The first rumored details of OpenAI’s highly anticipated hardware debut are here. Plus, you’ll learn how to create a fully functional AI agent using a no-code platform that turns plain text instructions into deployable workflows.

In today’s AI newsletter:

• How to Build a Production-Ready AI Agent in Minutes

• OpenAI’s Rumored “Sweetpea” Audio Wearable

• Atlas Enters Mass Production

• Google Research has released MedGemma 1.5

• 4 new AI tools worth trying

AI NEWS

🎧 OpenAI’s Rumored “Sweetpea” Audio Wearable

Fresh reports suggest OpenAI’s first consumer device is an audio-first wearable, internally codenamed Sweetpea, now the top priority for Jony Ive’s hardware team.

• Designed as an AirPods replacement, featuring a metal “eggstone” core with two removable capsule-style pieces that rest behind the ear

• Powered by a 2nm smartphone-class chip (Samsung Exynos favored), plus a custom chip aimed at handling phone-like actions via voice

• Massive ambition: 40–50M units projected in year one, with a target launch rumored around September

• Built with near smartphone-level BOM, signaling far more capability than standard earbuds

If Sweetpea can replace everyday phone actions with voice, it becomes the front door to AI, not an accessory. Audio isn’t a feature anymore, it’s the platform.

ROBOTICS

🤖 Atlas Enters Mass Production

Boston Dynamics unveiled the production-ready Atlas humanoid at CES 2026 and has begun manufacturing immediately, backed by Hyundai’s plan to deploy robots at massive scale.

• Atlas offers 56 degrees of freedom, lifts up to 50 kg, and autonomously swaps its own batteries

• 2026 deployments are fully booked, starting with Hyundai factories and Google DeepMind

• Trained using new AI foundation models to handle diverse industrial tasks, starting in automotive

• Hyundai is building robotics factories targeting 30,000 robots per year, aiming for tens of thousands of Atlas units in operation

With real customers, real scale, and AI-trained skills that replicate across fleets, Atlas marks the shift from humanoids as research projects to humanoids as industrial infrastructure. The question is no longer if humanoid robots enter industry, it’s how fast they spread.

SURVEY WITH BEEHIIV

Quick question about newsletter ads

We’re running a super short survey to see if our newsletter ads are being noticed. It takes about 20 seconds and there's just a few easy questions.

Your feedback helps us make smarter, better ads.

Take the survey now.

AI MODELS

🩺 Google Research has released MedGemma 1.5

MedGemma 1.5 boosts performance across medical imaging, records, and speech, while staying fully open for research and commercial use.

• Up to +14% accuracy on CT/MRI/histopathology and +22% on EHR Q&A

• Supports high-dimensional imaging (CT, MRI, whole-slide pathology) plus longitudinal X-ray timelines

• New MedASR cuts medical dictation errors by up to 82% vs general ASR
  4B model runs offline, scales on Vertex AI, and pairs speech + vision + text end-to-end

By combining medical images, records, and speech into one adaptable stack, Google is lowering the barrier for hospitals, startups, and governments to build domain-specific healthcare tools without locking into closed models.

HOW TO AI

💻 How to Build a Production-Ready AI Agent in Minutes

In this tutorial, you’ll learn how to create a fully functional AI agent using Vellum, a no-code platform that turns plain text instructions into deployable workflows.

🧰 Who is This For

• Developers who want to automate workflows fast

• Product teams experimenting with AI-powered features

• Founders building MVPs without writing code

• Anyone who wants to turn text ideas into working automations

STEP 1: Access the Agent Builder

Head over to Vellum.ai and Sign Up for free. You’ll land on a clean, chat-style dashboard that looks familiar if you’ve used ChatGPT before.

Right in the center, there’s a text box that says: “Describe what you want your agent to do.”

This is your command center, everything starts here.

STEP 2: Describe Your Agent in Detail

Now, simply describe what you want your agent to do, just like you’d explain it to a teammate.

For example, you can type:

“Create an agent that reviews legal contracts against a checklist and generates risk assessments with lawyer-friendly summaries…”

Once done, click enter and that’s it.

Vellum will instantly break down your request into clear steps and build the workflow for you automatically.

STEP 3: Customize and Configure

Once the workflow appears, you can click on any part of it to edit or fine-tune.

Change model types, swap integrations, or adjust prompts, all from one screen.

When ready, click Test Run to see your agent in action.

It’ll perform each step in real time, pulling data, writing text, or creating summaries right in front of you.

If something feels off, just tweak your instructions and test again until it’s perfect.

STEP 4: Deploy

When you’re happy with the results, click Publish.

From now on, your agent runs automatically whenever the conditions you set are met, no extra setup needed.

Meta confirms it is laying off Reality Labs staff and will “reinvest the savings” in wearables; CTO Andrew Bosworth says 1,000+ staff will be notified Tuesday.

Z.ai releases GLM-Image, an open-source multimodal AI model trained on Huawei chips that it says is China's first to be fully trained using domestic chips

Bandcamp bans music and audio that is “generated wholly or in substantial part by AI”, hoping to build trust with fans that the music they find was human-made.

The US Senate passes the Defiance Act that would allow victims to sue over nonconsensual, sexually explicit AI-generated images, amid furor over Grok images

🤖 Claude Cowork: Anthropic’s feature that turns Claude into a fully autonomous virtual assistant

🌐 Datastripes: Build data workflows in your browser with 300+ drag-and-drop blocks

📱 Gridfy: Turn raw data into interactive website widgets

📫 Gmail: Google’s email now powered by Gemini 3

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