Midjourney Medical Scanner: Unanswered Questions Behind the Wellness Spa Vision
Midjourney, the AI art company, has just unveiled its first physical product: a full-body ultrasonic scanner. This ambitious device, which we'll refer to as the Midjourney medical scanner, is touted by founder David Holz as "in many ways superior to even MRI machines," radiation-free, and capable of preventing 30% of deaths. However, radiologists on platforms like Hacker News and Reddit are already expressing profound skepticism, labeling these claims as unsubstantiated. Far from a medical breakthrough, the Midjourney medical scanner appears to be a speculative data grab, eerily reminiscent of Theranos's controversial strategy of bypassing rigorous medical validation for the sake of data collection.
Unpacking the Midjourney Medical Scanner's Ambitious Claims
The initial pitch for the Midjourney medical scanner promises a full-body scan in under 60 seconds, generating a sub-millimeter 3D map of the human body. This sounds revolutionary. Yet, the current prototype tells a different story: it reportedly takes about 20 minutes per scan and has only been tested on a mere dozen individuals. This significant latency gap between the grand promise and the current prototype's reality raises immediate and serious questions about scalability, real-world utility, and the company's readiness to deliver on its bold assertions. The promise of the Midjourney medical scanner is alluring, but the practicalities present a stark contrast.
Midjourney Medical, a dedicated nine-person division led by Ahmad Abbas (formerly of Apple Vision Pro hardware), is aggressively pushing this initiative. Their strategic plan conspicuously avoids immediate FDA diagnostic approval. Instead, they are opting for a launch of "Midjourney Spas" in high-profile locations such as Union Square, San Francisco, with initial plans targeting late 2027. This approach, centered around the Midjourney medical scanner, is not merely a business decision; it's a calculated regulatory sidestep.
These projected 25,000-square-foot facilities are envisioned to feature amenities like hot tubs, saunas, and ten scanners operating 24/7. The services offered will be marketed as "body composition maps" under the FDA's "general wellness" lane. This regulatory loophole allows companies to collect vast amounts of data under the guise of 'wellness' services, effectively bypassing the stringent medical validation required for diagnostic devices. It's a pathway that prioritizes data acquisition over proven clinical efficacy, a concerning parallel to previous ventures that have faced severe scrutiny.
<img src="
Midjourney's Technical Claims and Inherent Physical Limitations of the Medical Scanner
The core technology underpinning the Midjourney medical scanner is Ultrasound Computed Tomography (USCT). This sophisticated system employs a dense array of approximately 500,000 sub-millimeter sensors, leveraging Butterfly Network's semiconductor-based "Ultrasound-on-Chip" technology. The process involves emitting and recording sound waves while the subject is submerged in water, effectively performing a whole-body acoustic scan. As the platform lowers the individual at 5 centimeters per second, these sensors fire sound waves from every conceivable angle and meticulously record the echoes.
The raw acoustic data streams off the device at an astonishing 17 gigabytes per second, processed by over two petaflops of on-device compute. Crucially, this isn't generative AI creating anatomical structures; it's physics-based signal processing. The system solves complex partial differential equations (Full Waveform Inversion) to reconstruct the 3D volume. Deep learning algorithms then take over, handling anatomical segmentation and labeling to interpret the vast dataset.
Midjourney's Technical Claims
The fundamental challenge facing the Midjourney medical scanner isn't a lack of ambition or computational power, but rather the inherent limitations imposed by ultrasound physics. Ultrasound, by its very nature, struggles precisely where MRI excels. Acoustic waves attenuate significantly at tissue interfaces, particularly bone (where attenuation can be up to 50 times more than in soft tissue) and air (which causes almost complete reflection). This physical reality creates unavoidable "blind spots" for the technology.
For instance, deep brain imaging through the skull is simply not feasible with this ultrasound-based technology. Similarly, detailed imaging of organs obscured by bowel gas or the ribcage—such as the lungs or significant portions of the abdomen—will be exceptionally difficult, if not impossible, to achieve with MRI-level resolution. Early iterations of the Midjourney medical scanner will likely show good detail in peripheral areas like limbs, joints, and superficial muscles. However, the deep abdomen, thorax, and pelvic cavities are almost guaranteed to appear as a blurry mess when compared to the clarity offered by an MRI. This represents a fundamental barrier to the device's utility for comprehensive whole-body diagnostics, far from being a minor limitation.
<img src="
The Regulatory Sidestep: 'General Wellness' and the Midjourney Medical Scanner's Data Collection
The decision by Midjourney Medical to launch "wellness spas" and market "body composition maps" under the FDA's "general wellness" category is a critical aspect of their strategy. This regulatory pathway is designed for products that promote a healthy lifestyle, not for those making diagnostic claims or intended to treat or diagnose disease. By operating within this lane, Midjourney can avoid the rigorous and lengthy clinical trials, safety assessments, and efficacy studies typically required for medical devices. This allows for rapid deployment and, crucially, the immediate commencement of data collection on a massive scale. The implications of the Midjourney medical scanner's approach extend beyond technical feasibility into ethical considerations.
However, this sidestep carries significant ethical and public health implications. Presenting a device with such advanced imaging capabilities, even under a "wellness" label, can create a false sense of security or, conversely, undue alarm among the public. Without proper medical validation, the data collected by the Midjourney medical scanner, while vast, lacks the clinical context and reliability necessary for true medical insight. This strategy, while expedient for business, places the burden of interpretation and potential misdiagnosis squarely on the individual, bypassing established medical safeguards and potentially overwhelming the existing healthcare infrastructure with unverified findings.
The Abstraction Cost: Flawed Data, AI Training Risks, and the Midjourney Medical Scanner
Midjourney's real, long-term goal is unequivocally data collection. The company projects an ambitious deployment of 50,000 scanners worldwide by 2031, aiming for an astounding billion scans per month. The ultimate purpose of this colossal dataset is to train future AI models, presumably to overcome the very physical limitations inherent in the Midjourney medical scanner's ultrasound technology. However, if the raw data itself is fundamentally limited by physics—with inherent blind spots and areas of poor resolution—the resulting AI will inevitably be trained on flawed and incomplete information. This incurs a significant 'abstraction cost,' meaning the AI's interpretations and diagnostic capabilities will be built upon an unreliable foundation, leading to potentially inaccurate and unreliable outcomes. The skepticism surrounding the Midjourney medical scanner is well-founded given these challenges.
Mass asymptomatic screening with a device that has known blind spots and unproven diagnostic capabilities will inevitably lead to a flood of incidental findings. These findings, often benign, will nonetheless require expensive, anxiety-inducing follow-up tests, placing an unnecessary burden on individuals and the healthcare system. Instead of genuinely improving healthcare outcomes, this strategy appears primarily focused on generating a massive, potentially flawed dataset under the guise of "wellness." This creates a high abstraction cost that could lead to significant practical, financial, and emotional burdens for patients and medical professionals alike.
In conclusion, the Midjourney medical scanner is not a "superior" alternative to MRI. It represents a high-stakes gamble on data collection, deliberately bypassing established medical validation processes, and hoping to train an AI on inherently incomplete information. The critical question remains: can Midjourney's data-driven gamble truly overcome these inherent physical limitations, or will the abstraction cost of flawed data ultimately render its AI unreliable and its wellness spas a source of more questions than answers? The medical community, and patients, are still waiting for definitive answers.
