Midjourney Wants Full-Body Health Scans for Anyone, Anytime
New Database Lets AI Search Thousands of Local Laws
June 18, 2026
D.A.D. today covers 11 stories — about a 10-minute read. What's New, What's Innovative, What's Controversial, What's in the Lab, and What's in Academe.
The Daily AI Digest is a daily AI briefing automated by Alexander Panetta — a veteran political journalist tracking the field during a Master's in AI Management at Georgetown University.
D.A.D. Joke of the Day: My AI assistant asked if I had any feedback. I said it was too verbose. It replied with a 500-word apology.
What's New
AI developments from the last 24 hours
Midjourney, the AI Art Company, Pitches a 60-Second Full-Body Scan
Medical imaging is one of the biggest bottlenecks in stretched health systems: MRI machines cost millions, scans run hundreds to thousands of dollars, and wait lists stretch for months. Midjourney — the AI image generator — says it has a way around all of it. Its new healthcare division has unveiled a scanner, branded "Ultrasonic CT," that it claims can produce an MRI-grade, sub-millimeter map of the entire body in about 60 seconds using only sound waves and water — no radiation, no magnets, no multi-ton magnet room. The company wants to make scanning so cheap and routine it becomes a casual habit, deploying 50,000 machines for a billion scans a month by 2031, and claims that catching disease this early could eventually help avoid "30% of all deaths and 50% of all healthcare costs." The catch: the first one won't open — inside a spa in San Francisco — until late 2027.
Why it matters: The reaction from technologists was immediate and skeptical — on Hacker News, a physician-engineer said the scanner isn't "going to be displacing MRIs or remotely close," while others scoffed that Midjourney had "lost the plot, especially with the spa." That caution is warranted: this is a vision, not a product, and the mortality and cost figures are Midjourney's own hope, not evidence. But the fundamentals are real. The company is profitable with no venture money and $200M-plus in revenue, the underlying physics (ultrasound tomography) is established science, and as RDWorld reported, Midjourney licensed the chip technology from Butterfly Network last November. The hard part is the medicine, not the imaging. "Ultrasonic CT" is marketing — it uses neither X-rays (CT) nor magnets (MRI) — and Midjourney sidesteps the FDA at launch by offering only non-diagnostic "body composition maps," the same lane as full-body-MRI startups like Prenuvo, which doctors fault for flooding patients with false alarms and costly follow-ups. For institutions weighing what could actually relieve a strained system, cheaper scanning only helps if the results are accurate enough to act on.
Source: midjourney.com · Source: rdworldonline.com · Discuss on Hacker News
White House's Condition for Restoring Anthropic's Top Models May Be Impossible to Meet
A week after the Trump administration forced Anthropic to pull its two most advanced models — Fable 5 and Mythos 5 — offline, there's still no deal, and the bar Washington has set for restoring them may not be reachable. WIRED reports that officials want Anthropic to guarantee the models' guardrails can't be circumvented before any re-release — in effect, to block every possible jailbreak. Security researchers told the magazine that's something no AI model can promise, raising the question of whether the condition is meant to be unmeetable. Publicly, President Trump played it down: after meeting CEO Dario Amodei at the G7 summit in France on Wednesday — their first public encounter since the order — he told reporters the talks were "going fine," Reuters reported. Inside the company it looks rockier. Monday and Tuesday meetings produced no breakthrough, and The New York Times, citing internal chats, reports that many of Anthropic's roughly 3,000 employees feel singled out, one asking, "Are we being bullied based on bad vibes?" and another, "At what point does this just feel like they don't want us to exist?"
Why it matters: If the price of relisting a model is proving it can never be jailbroken — a guarantee no lab can make — then Washington can effectively keep a frontier model offline indefinitely, turning a one-time safety call into open-ended leverage over a single company. The substance is contested, too. The capability the government flagged (drawn from an Amazon paper, per the Times) is, according to a cybersecurity expert who reviewed it, exactly what makes the model useful to defenders — not a flaw to patch — which is why more than 150 security researchers have asked the Commerce Department to reverse course. For institutions building on frontier AI, it compounds last week's lesson: model access now hinges on a political relationship, and possibly on a technical standard that can't be satisfied.
Source: WIRED · Reuters · The New York Times
Amodei, in a Wide-Ranging Bloomberg Interview, Defends Military Work and Doubles Down on Job Warnings
In an extended interview for Bloomberg Originals' The Circuit, Anthropic CEO Dario Amodei addressed that very standoff — and the company's other flashpoints. The standout exchanges:
- The fight with Washington. Amodei rejected the idea of "winning" the standoff — "I don't think there's any winning this fight... this is more a debate about what the proper use of AI by the government is" — and described his White House dealings as a mix of officials "who we get along with very well" and "other parts that are harder to get along with."
- AI in war. Pressed on a Bloomberg report that Claude-assisted targeting, via Palantir's Maven system, was used in the Iran war — including a strike on a girls' school that reportedly killed more than 150 people, most of them children — Amodei said Anthropic doesn't "know exactly how these models were used," but defended its red lines: a human, not AI, must make the final call, and the company refuses to enable fully autonomous weapons or mass surveillance. He called himself "a patriot" who backs a militarily strong US as a deterrent.
- Jobs. He stood by his warning that AI could eliminate a large share of entry-level white-collar work, floated a future of "very fast GDP growth and high unemployment," and pushed back hard on critics (including Nvidia's Jensen Huang) who call it "doom marketing."
- The "SaaSpocalypse." After Claude Cowork's launch wiped a reported $285 billion in software market value overnight, he predicted the software industry "gets larger, not smaller" — but with "some big losers" among incumbents that don't adapt.
- A government takeover. Amodei said he shares concerns about the state seizing AI labs, arguing instead for "checks and balances" — noting Anthropic's Long-Term Benefit Trust already has the power to fire him.
Why it matters: It's the fullest picture yet of how the most safety-vocal frontier lab squares working with the military, courting a near-$1-trillion valuation, and warning about its own technology — all while fighting Washington over the very models in the story above.
Watch on Bloomberg Originals — The Circuit
750-Task Benchmark Tests Whether AI Can Handle Real Biotech Research
A consortium of 173 scientists has released LifeSciBench, a benchmark designed to test whether AI systems can handle actual life science research work—not just answer textbook biology questions. The 750 tasks span drug discovery, genomics, and other biotech workflows, with 79% requiring multi-step reasoning and over half demanding interpretation of real research artifacts like figures, sequence files, and chemical structures. Tasks were authored and reviewed by Ph.D.-trained scientists with industry experience, requiring 90%+ reviewer agreement for inclusion.
Why it matters: As pharma and biotech companies explore AI for research acceleration, this benchmark offers a rigorous test of whether current models can actually assist scientists—or just pass biology exams.
What's in the Lab
New announcements from major AI labs
OpenAI's AI Chemist Autonomously Improves Drug-Making Reactions
OpenAI connected GPT-5.4 to Molecule.one's robotic chemistry lab and let it autonomously research ways to improve Chan-Lam coupling—a reaction used in drug development that's notoriously finicky. The AI independently identified a difficult substrate class and proposed adding TEMPO, a mild oxidant, as a fix. Results: yields improved for 88% of boronic acids and 83% of sulfonamides tested, with average yields jumping from 16.6% to 25.2%. Human chemists validated the findings at bench scale, seeing more than twofold yield increases in most cases.
Why it matters: This is among the clearest demonstrations yet of AI functioning as a near-autonomous research partner—not just analyzing data but proposing hypotheses, running experiments through robotic labs, and producing results that hold up under human verification.
Cohere Details How It Prevents One Customer's Traffic From Slowing Others
Cohere published a technical deep-dive on how it prevents one customer's traffic spike from degrading service for others—the 'noisy neighbor' problem common in shared AI infrastructure. Their solution layers multiple scheduling techniques: rate limiting, performance tiers, a 'Deficit Round Robin' algorithm that ensures each tenant gets fair capacity over time, and priority-based queuing. The post is architecture documentation rather than a product announcement.
Why it matters: Enterprise buyers evaluating AI providers increasingly care about service guarantees—this signals Cohere positioning itself on reliability and transparency for multi-tenant deployments, though no benchmark data accompanies the claims.
What's in Academe
New papers on AI and its effects from researchers
Researchers Map Where Tech Falls Short for Dementia Caregivers
Researchers have developed a taxonomy mapping Alzheimer's and dementia caregivers' mental health needs to technology-based interventions. The framework, built from literature review and caregiver interviews, identifies gaps in current tech support—particularly for relational strain and compassion fatigue, two areas caregivers prioritize but existing tools largely ignore. The study notes that 11 million Americans provided 18 billion hours of unpaid dementia care in 2023, underscoring the scale of unmet need.
Why it matters: For healthcare organizations and AI developers building caregiver support tools, this taxonomy flags where current solutions fall short—and where new products could find an underserved market.
Paper Proposes Rebuilding the Web Around AI Agents, Not Humans
A new research paper argues the web's foundational assumption—that humans are the end consumers—is now broken. With AI agents increasingly acting as intermediaries between people and information, the authors propose redesigning web infrastructure to treat agents as 'first-class citizens.' Their framework includes agent identification in HTTP headers, token-based subscription models for AI access, and a new markup language with cryptographic provenance chains to address AI-generated content becoming detached from verifiable human sources.
Why it matters: If AI agents will browse, summarize, and transact on our behalf, the web's technical and economic architecture may need rethinking—raising questions about content monetization, trust verification, and who controls the agent layer.
Developers Need Different AI Assistant Styles Based on How They Think
A study of 27 developers using GitHub Copilot's chat feature found that problem-solving style and experience level significantly shape how people interact with AI coding assistants. Researchers identified five distinct 'interaction modes' and ten underlying needs—meaning two developers using the same tool may approach it in fundamentally different ways. Some developers want step-by-step guidance, others want quick answers, and the differences correlate with cognitive style, not just skill level.
Why it matters: As AI assistants become standard workplace tools, teams may need to account for individual cognitive styles when training employees or evaluating productivity gains—what works brilliantly for one person may frustrate another.
New Database Lets AI Search Thousands of Local Laws
Researchers have built LOCUS, the first large-scale, machine-readable database of U.S. local ordinances—the municipal and county codes that govern zoning, permits, business licenses, and daily life but have been locked inside vendor platforms designed for human browsing. The corpus covers 9,239 cities and counties in raw form, with a harmonized layer spanning 2,309 of 3,144 U.S. counties. Dataset and models are available on Hugging Face.
Why it matters: Local law has been a blind spot for legal AI—this dataset could enable tools that help businesses, real estate firms, and compliance teams navigate the patchwork of rules that vary block by block.
Benchmark Reveals AI Safety Guardrails Fail on Technical Science Queries
A research team has released SciRisk-Bench, a benchmark testing whether AI models can safely handle scientific queries that might touch on biosecurity, chemical hazards, or dual-use research. The benchmark spans 7 scientific disciplines, 31 subdisciplines, and 10 risk categories. Evaluating both general-purpose LLMs and science-focused models, researchers found gaps in safety guardrails when queries are framed in technical scientific language rather than obvious red-flag phrasing.
Why it matters: As AI tools become standard in research labs and R&D departments, this benchmark offers a systematic way to audit whether models can distinguish legitimate scientific inquiry from requests that could enable harm—a growing concern for institutions deploying AI in sensitive research contexts.
What's On The Pod
Some new podcast episodes
The Cognitive Revolution — Radically Better Reasoning: Elicit's Andreas Stuhlmüller & Jungwon Byun on World Models for Research
How I AI — How to design AI agent loops: schedules, goals, and subagents in Claude Code and Codex
AI in Business — Scaling Agentic AI in CX Without Losing the Customer - with Shri Nandan of Comcast
AI in Business — How Vertical AI Achieves Defensible Accuracy - with Steve Hasker of Thomson Reuters