Welcome back and Happy Juneteenth! Today marks the emancipation of African American slaves in the US, and there’s something fitting about pausing on a day meant to commemorate freedom and community. So here’s my pitch: Close the laptop a little earlier today. Go find your people. Sit on a porch, walk somewhere with no destination, share a meal with someone. The AI news will still be here tomorrow (it always is).
Just a few quick things before you go…
Wait, Midjourney Just Built What?
Remember last week when David Holz teased some mystery hardware Midjourney was about to launch? Whatever you were guessing, I'm betting it wasn't this.
On Wednesday, the AI image-generation company revealed its first hardware product, and it's not a creative tool, a wearable, or anything image-related at all. It's a full-body ultrasound scanner.
Here's what they actually unveiled:
A new division called Midjourney Medical and a product called The Midjourney Scanner—branded as an "Ultrasonic CT" machine.
You step onto a platform that descends into a shallow pool of water (for the ultrasound coupling), and 40 Butterfly Network ultrasound-on-chip modules scan your whole body in about 60 seconds.
But how do you get one? You go to a Midjourney Spa. They're opening their flagship in San Francisco in late 2027 with 10 scanners alongside hot tubs, saunas, and cold plunges.
The long-term ambition is genuinely bonkers: 5,000 spa locations and 50,000 scanners globally by 2031, with a stated goal of a billion full-body scans every month.
My POV: Midjourney pivoting from generating images of dragons to generating images of your insides is one of the most unexpected company moves I've seen in a long time. But it also kind of tracks—they're a research lab whose whole identity has been about "amplifying human imagination," and now they're pointing those generative-imaging chops at the literal human body.
There's a long road before this is a real diagnostic tool. FDA approval is still pending, and first applications are limited to body composition mapping (muscle, fat, bone).
I said last week I'd probably end up buying whatever Midjourney shipped. Turns out the answer is: I'll book the appointment.
— Matt
Anthropic Makes Its First Climate Move
Via Quartz
Anthropic’s been everywhere lately, and here’s the latest: it just became the first AI startup to join Frontier, the carbon removal coalition.
Here’s the deal: Through Frontier, companies pool money and pre-purchase carbon credits, which gives early-stage removal projects the funding certainty to actually get built. It was started in 2022 by Stripe, Google, and Shopify, so the AI-adjacent crowd has been here a while. But Anthropic is the first pure AI company to walk in the door.
Anthropic’s piece of it: The company is contributing to a fresh $915 million round of pledges, nearly doubling Frontier’s total to $1.8 billion. This is Anthropic’s first climate-related deal, full stop—no sustainability report yet, and it’s openly favored an “all of the above” energy approach.
Why it matters: AI runs on electricity, and the data center buildout has the whole industry on an energy-buying spree that hasn’t always been clean. A move like this hints the frontier labs are starting to reckon with how much power they’re pulling—and looking for ways to answer for it.
AI render studio for design teams
Via Rendervi
Rendervi turns model previews, sketches, or clay renders into client-ready photoreal images and videos. It analyzes your visuals and builds prompts automatically, then lets you swap materials, tweak environments, and upscale with one click.
How you can use it
Transform rough sketches into polished renders instantly
Change out materials and textures without manual rework
Maintain visual consistency across multiple views
Generate short videos from static renders
Pricing: Free and paid
Jobs, announcements, and big ideas
OpenAI adds spend controls and usage analytics to ChatGPT Enterprise, giving admins tighter budget oversight.
Anthropic brings live, shareable artifacts to Claude Code, letting developers preview and pass around their work.
Perplexity unveils Brain, a self-improving memory system that lets AI agents learn from past tasks.
Google DeepMind publishes an AI Control roadmap laying out how to secure increasingly capable agents.
Anthropic adds centralized enterprise authorization for MCP connectors in Claude, streamlining access control.
PewDiePie built an open source AI workspace called Project Odysseus…and I put it through its paces so you don't have to.
That’s a wrap! See you next week for more.
