TBPN

Why Midjourney Can Make Medical Machines

June 18, 2026

Happy Thursday.

We’re live on YouTube and 𝕏.

The current thing in tech and business is Midjourney’s pivot to medical machines.

Today’s Lineup

Writer Derek Thompson at 11:30 AM
Arm CEO Rene Haas at 12:00 PM
Reindustrialize Co-Founder Mike Slagh at 12:30 PM
Science Fiction Author Rob Reid at 12:40 PM
Taste Labs Founder & CEO Thais Castello Branco at 1:00 PM
Founders Podcast Host David Senra at 1:10 PM
Entrepreneur & Boxer Jake Paul at 1:30 PM

Run of ShowScreencap from Midjourney’s technical dive into Midjourney Scanner

Why Midjourney Can Make Medical Machines

What happened: Midjourney released not one but two videos yesterday that got more than 13 million combined views in something like 14 hours. The videos were the company’s introduction to Midjourney Medical and the Midjourney Scanner, a full-body ultrasound tomography machine that uses a water bath for 60-second full-body scans without any radiation.

The details of how this machine works are pretty technical, but basically it competes with traditional MRI, CT, and conventional ultrasound, and in principle can produce close to the same or better results faster, more cheaply, and more safely than its competitors. You can look at a gallery of the machine’s early scans here.

And here’s a table that compares ultrasound tomography to conventional medical scanning, courtesy of Grok:

The machines are meant to provide preventive scans, and (again in theory) would be used for detecting tumors, cardiovascular issues that lead to heart attacks, strokes, etc., organ diseases, and musculoskeletal issues (e.g. muscle tears in athletes, osteoporosis complications). From Grok, again:

In summary, it targets silent, progressive conditions where repeated safe imaging provides the biggest edge: cancers, heart/vascular disease, metabolic organ issues, and body composition risks. This shifts healthcare toward true prevention through accessible, data-rich monitoring.

Hank Green added some nuance to (and poured a little water on) the discourse that’s worth reading, as well.

It looks like Midjourney will roll the first of these machines out at its newly planned Midjourney Spa, the first of which will open in SF’s Union Square toward the end of next year. Midjourney wrote all of these announcements up in a blogpost yesterday, including its roadmap to mass-scale medical scanning, which you can read here. — Brandon

How did Midjourney get here? The story of David Holz is pretty fascinating and inspiring. He grew up in Florida, the son of a dentist. Like most kids, he played video games, but quickly graduated to hacking on them for fun.

Clearly, David is very intelligent. PhD in math track, worked at NASA, but eventually branched out into entrepreneurship. Leap Motion was a pretty traditional venture-backed startup, and by all accounts it was a grind. Its first product brought hand tracking to VR. Big hardware challenges, complex market dynamics with bigger companies (Meta bought Oculus, Apple investing heavily, etc). Eventually Apple actually tried to buy David’s company, but the deal fell through at the eleventh hour. It was really rough for the company, they had to do layoffs, but they kept building. The devices sold to hackers, but broader VR adoption just never took off enough to sustain the business.

After about a decade, David was able to take a step back and start hacking on the next thing: AI image generation. Midjourney was a huge viral success that broke a lot of the traditional rules. No venture funding, no web front-end or app for years. The entire experience was contained to a massive Discord. A very odd product decision to be so dependent on another platform, but it had a bunch of knock on benefits. First, Midjourney basically had an extremely polished desktop and mobile experience from day one. This let the team focus on the core innovation — opinionated image generation that actually looked tasteful (to fans, myself included). Second, Discord allowed for a multiplayer AI interaction experience. New users could take inspiration from other users, see what prompts they were having the most luck with, and easily remix images. By generating four candidate images for each prompt, Midjourney got a bunch of really valuable data on what a “correct” image looked like for a specific prompt. A really solid data / RL flywheel.

The company still hasn’t taken venture funding, and after the big Meta deal to power Vibes, there’s clearly enough cash flow to fund other projects. I was personally hoping for a VR headset, but this seems way more impactful and unique. Very excited to see where it goes next. — John

Clip Spotlight: Apple’s biggest product year ever coming in 2027

Yesterday on the show, Bloomberg’s Mark Gurman predicted that 2027 will be Apple’s biggest product year in its history.

What he says is coming out: