344
Productivity & Workflow355
Automation & Workflow224
Software Development250
Marketing & Growth192
AI Infrastructure & MLOps174
Writing & Content Creation203
Data & Analytics141
Design & Creative169
Photography & Imaging156
Customer Support131
Sales & Outreach125
Voice & Speech135
Education & Learning131
Operations & Admin87
Midjourney published a behind-the-scenes video of its dunk-tank ultrasound scanner, but key questions about image quality and proof of performance remain.
In short: Midjourney released a behind-the-scenes video of its dunk-tank ultrasound scanner, but it still has not provided clear evidence that it delivers the detailed scans it has promised.
Midjourney, a company best known for an AI image generator, posted a nearly 20-minute video showing its work on a full body ultrasound scanner. Ultrasound is the same scanning method often used in pregnancy checkups, and it uses sound waves, not radiation.
The video is hosted by YouTuber Marcin Plaza, who is also an engineer at Midjourney. He describes the machine as many ultrasound probes taken apart and attached to something like a “glorified hot tub with an elevator in it,” plus regular computers and small Raspberry Pi boards (tiny, low-cost computers often used in DIY projects).
Midjourney says it wants to place the scanner in spas and sell it as a wellness product. In the video, the company’s head of medical, Tom Calloway, says the first focus will be “body composition,” meaning things like estimating fat and muscle. That approach matters because a wellness product usually faces fewer rules than a medical device that diagnoses disease, which would typically require FDA clearance and clinical trials.
Experts previously told The Verge that Midjourney had not shown enough proof it can overcome known limits of ultrasound. The new video shows more hardware and building work, but it does not fully answer those physics and image-quality questions.
Full body scans can sound reassuring, like getting a detailed “check engine” report for your body. But without strong evidence and careful oversight, they can also mislead people, miss problems, or cause unnecessary worry and follow-up tests.
Source: The Verge AI