Welcome, humans.

So, Midjourney (the image model company) apparently woke up one day recently and said y’know what we should make? A better MRI machine. Also: “what if this were faster, cheaper, wetter, and somehow a spa that looks like the carbon freezing chamber in Return of the Jedi?”

The company just introduced Midjourney Medical and the Midjourney Scanner, a full-body ultrasound system that sends millions of sound waves through water to create a detailed internal scan in about 60 seconds. The first “Midjourney Spa” is planned for San Francisco in 2027, and the long-term goal is 50,000 scanners worldwide producing roughly 1B scans per month by 2031.

So yes, the AI image company now wants to turn medical imaging into a futuristic bathhouse membership. Y’know what, I’m with it. Solve more hard problems ppl!

Here’s what happened in AI today:

• 🙀 Trump advisers reportedly weighed government equity stakes in major AI companies.

• 📰 Dario Amodei and Demis Hassabis pushed for a U.S.-led AI coalition at the G7.

• 📰 Amazon backed Odyssey's $310M round to build AI world models.

• 🍪 Snap launched $2,195 AI-powered AR glasses called SPECS.

• 🎓 An AI Skill for Reviewing the right code: by risks, not by size.

Hey: Want to reach 700,000+ AI-hungry readers? Advertise with us! 

P.S: Love robots? We have a new robotics newsletter. Sign up here.

😺 Washington Wants a Piece of the AI Boom

As Chubby from X just shared, we’ve got some Manhattan Project levels of cringe going on at the G7 summit this week as the AI CEOs meet with Trump and other world leaders. By the looks of it, current AI industry Main Character Anthropic CEO Dario Amodei is not exactly feeling like the the Belle of the Ball atm:

What’s going on here? Well, AI companies have spent the last few years asking Washington for rules.

Now, Washington may be asking for equity.

Semafor reported that senior Trump administration officials discussed possible government equity stakes in major AI companies before the recent export-control fight with Anthropic. One idea would seed "Trump Accounts" with AI equity. Another would route stakes into a sovereign wealth fund, which is basically a government investment fund.

Here's what happened:

• Treasury Secretary Scott Bessent reportedly favored using AI equity to fund Trump Accounts.

• Commerce Secretary Howard Lutnick reportedly preferred a sovereign wealth fund structure. Bernie Sanders just proposed a law that would do just that: giving Americans direct ownership stakes in the country’s largest AI companies.

• No decision has been made.

• The talks came before the U.S. forced Anthropic to restrict access to Fable 5 and Mythos 5.

• At the G7, Anthropic's Dario Amodei and Google DeepMind's Demis Hassabis reportedly pushed for a U.S.-led coalition on AI rules, chips, model access, and safety risks.

Why this matters: Washington is moving from regulating AI companies to negotiating leverage over them.

That is a big shift. Export controls decide who can use the models. Standards coalitions decide whose rules travel globally. Equity stakes would raise a stranger question: if AI becomes a strategic national asset, should the public get a financial piece of the upside?

The industry will hate parts of this. Microsoft and Meta have reportedly shown little interest in government equity ideas, and Semafor noted the plan remains an extremely tough sell beyond OpenAI, which pitched a version last year.

Our take: frontier AI is being treated less like software and more like strategic infrastructure. The government wants safety, leverage, and economic upside. The labs want global markets, talent, and freedom to move fast.

Users want one simpler thing: the AI tools they built around yesterday to still work tomorrow. The recent Fable debacle made that a very real concern.

TBH, all we’re seeing on social media, and also feeling deeply in our own hearts, is that we kinda miss Fable (the Claude model the gov just banned). It was low key really good, so hopefully the US Gov and the frontier AI lads can get this whole mess about national sovereignty over top tier AI models and who can use what sorted out.

• Best case scenario? Fable 5 is put back online, without any net new power grabs or strict regulations limiting its use (but maybe we do put some limits on future model releases and slow down the cadence of the industry from what one OpenAI team member called out as new Mythos-class models “every month.” as that would just be unsustainable bro).

  • Most importantly, open source AI development remains untouched, as this little episode showed it is more important than ever that we have access to open research and models you can run on your own servers so the whole industry (and in my opinion, all of humanity) can flourish with available intelligence on demand.  

• Worst case scenario? No one gets Fable, or even worse, the government claims ownership over it and restricts it so only they can access it, and on top of that, they pass some law banning open source model development.

• If this happens, we’re stuck with the worst of all possible outcomes: choosing between “we can charge whatever we want because there’s no competition” pricing for AI on tap from the big labs, or cheaper lesser intelligent AI from second-tier sources because we can’t access the latest frontier open models.

Hey current administration, if any of y’all read this? Plz go with best case scenario, plz and ty!

FROM OUR PARTNERS

Secure Your Data & AI Supply Chain: Now On-Demand 

Stop patchwork tools from stalling innovation. Watch the on-demand webinar, "Elevate Data and AI Agent Security with GitLab and Google Cloud," where experts from GitLab and Google Cloud share how to secure BigQuery pipelines, automate compliance, and consolidate tooling. 

Protect your code and workloads at scale—on your schedule.

 Watch On-Demand

🎓 AI Skill of the Day: Review AI Code by Risk, Not Size

Rahul Sengottuvelu, head of applied AI at Ramp, made a post this week that got the attention of Elvis Saravia, Boris Cherny, and Anita from Vellum; that’s AI educators, benchmarkers, and the creator of Claude Code. So you know it hit a nerve.

His core mental model for the modern AI age: “fable+ class models” (so models like the newly banned Fable 5) are becoming “english -> code interpreters” that convert your idea into “correct” code, regardless of problem complexity or output complexity.

Meaning: you describe the product change you want, and the model turns it into working code across increasingly bigger chunks of the codebase. Or put another way, you can trust it to do larger chunks of work entirely on its own.

Here’s where Rahul’s rule comes into play: manage that change based on risk, not size.

That means engineers need to stop treating every AI-written change the same way. A 12-line change to the login system can be more dangerous than a 1,200-line change to a settings page, because login code controls who gets access to what.

Payments, identity, data access, network calls, and private customer info are “high-risk” because one mistake can expose data, move money, or create a security hole. For those areas, Rahul’s advice is to keep the code changes small enough for a human to inspect carefully.

For lower-risk work, like UI changes, backend plumbing, formatting, internal tools, or performance tweaks, a giant AI-generated change might be fine if you can prove it works. Instead of reading every line, you run tests, compare outputs, put it behind a feature flag, or let it run in shadow mode, where the new code operates quietly without affecting real users.

Try this before merging AI-generated code:

Review this AI-generated code by risk, not by size.

First, classify the change:
1. High risk: auth, identity, payments, data access, network access, PII, security, or production database writes.
2. Medium risk: business logic, user-facing behavior, integrations, or performance.
3. Low risk: UI, formatting, internal tooling, backend plumbing, or code that can be tested safely.

Then tell me:
- What could go wrong?
- What needs line-by-line review?
- What can be verified empirically with tests?
- Should this run behind a feature flag, in a sandbox, or in shadow mode before release?
- What guardrails would make this safe to merge faster?

Anthropic’s Boris Cherny agreed and added the next step: the future workflow is Claude Code + an advanced model + a verifier in a loop. A verifier is the checking layer that tests whether the code behaves the way you expected.

But don’t treat that as “let the AI approve its own work.” Elvis chimed in and warned that blind autonomous loops don’t work without guardrails.

We recommend to use AI to flag what a human needs to inspect, not to make the final judgment. Ask it to point out risky files, missing tests, possible security holes, data-access issues, and places where the code should run in a sandbox or shadow mode before users ever touch it. This is especially true if you are new to building your own code; as normies, we will all eventually use coding agents for knowledge work, so it’s important to learn these patterns now as best practices to scale with our confidence.

Total AI beginner? Start here (goes with this video).

Have a specific skill you want to learn? Request it here. 

🍪 Treats to Try

*Asterisk = from our partners (only the first one!). Advertise to 700K+ readers here!

1. *10 weeks. $30M USD in AI impact. This can be you. → See how we did it.

2. Snap SPECS puts an AI assistant and AR lenses into glasses that can see what you see, with developer tooling for Lens Studio, Claude Code, Codex, and Cursor, $2,195 preorder.

3. Anthropic upgraded Claude Design, so you can turn a prompt into on-brand designs, slides, and prototypes, then export them or keep working in tools like Adobe, Canva, Miro, Replit, Vercel, and Wix—available on Claude Pro, Max, Team, and Enterprise.

4. Palmier Pro gives AI video generation a native editing timeline, so you can generate clips, edit them, and export to Premiere or DaVinci, no pricing details.

5. Ask Pinterest helps you shop with a conversational assistant that remembers your taste, boards, and saved Pins, limited access—try it here.

6. ChatGPT scheduled tasks now has a dedicated management page and faster recurring briefings from your interests, chats, and connected apps, included in eligible ChatGPT plans.

7. Killed by OpenAI tracks OpenAI products, features, and APIs as alive, killed, or zombie, free.

NEW interview: Andrew Dai on why AI still can’t really see.

Andrew Dai, former Google Brain / DeepMind researcher and now CEO of Elorian, joined us to explain why today’s AI can write code, read docs, and build apps, but still misses visual problems a grade-schooler could spot in seconds.

In our latest episode, Andrew unpacks why image description is not the same as visual reasoning, why models may need a “visual scratchpad,” and why better vision could unlock stronger coding agents, robotics, design review, and physical-world engineering.

🎧 Watch / Listen: YouTube | Spotify | Apple Podcasts

📰 Around the Horn

• Amazon joined Odyssey's $310M round as the world-model startup chose AWS and Trainium chips to build AI that simulates physical environments.

• Noam Shazeer joined OpenAI to lead architecture research, taking a Transformer co-author and former Gemini technical co-lead back out of Google.

• OpenAI introduced LifeSciBench, a 750-task biology benchmark, then showed an AI chemist using GPT-5.4 to improve a medicinal-chemistry reaction.

• AWS launched Continuum and Context to help companies coordinate agents across teams, memory, and workplace systems.

• Pew Research Center found 49% of U.S. adults have used AI chatbots, up from 33% in 2024, while 63% say AI is advancing too quickly.

• DeepL acquired Mixhalo to improve live-event audio streaming and translation.

• Alibaba Cloud expanded in Europe and said new agentic AI services are coming for regional customers later this year.

FROM OUR PARTNERS

Adobe Firefly expands agentic capabilities and advances its creative AI studio experience with:  brand kit creation (describe your style, brand name, and color palette and let it go to work), short product video creation, Quick Cut (auto-assemble video footage into a polished cut), storyboards (visualize and sequence ideas before production).

Check out the latest AI innovations today at firefly.adobe.com

🧩 Thursday Trivia:

You know the drill: one is AI, and one is real. Which is which? Vote in the poll below!

A.

B.

A Cat’s Commentary

Trivia answer: A is AI, and B is real.

P.S: Love the newsletter, but only want to get it once per week? Don’t unsubscribe—update your preferences here.