Good morning. It’s Friday, June 19th.

You read. We listen. Let us know what you think by replying to this email.

Six people doing the work. Your headcount is one.

Your finance close runs in #finance. Stripe and QuickBooks reconciled, runway updated, posted Sunday night without you asking.

Engineering review lands in #eng. Viktor pulled the open PRs, left comments on auth-refactor, flagged a dependency blocking api-pagination.

Campaign brief lands in #growth: Meta CPA up 18%, recommendation to pause broad match, a draft landing page already deployed for the variant test.

You hired him on day zero. He lives in Slack and Microsoft Teams alongside your contractors and investors, connects to 3,000+ tools, pushes back when you ship something dumb.

"Viktor is now an integral team member, and after weeks of use we still feel we haven't uncovered the full potential." Patrick, Director, Yarra Web.

Get Started for Free

Midjourney is building a full body ultrasound scanner and a luxury spa to put it in

Midjourney is jumping from generative AI into healthcare with a wild new project: a full-body ultrasound scanner. The Midjourney Medical device trades radiation-heavy CT scans for sound waves. In a technical dive shared on 𝕏, the team revealed that the system uses 40 Ultrasound-on-Chip modules to pack 358,000 ultrasonic elements into a water-filled ring.

Stepping into the device triggers a massive data dump, pulling 17 gigabytes per second. To reconstruct these high-resolution bodily slices, Midjourney sends the raw data to 21 cloud servers packing two petaflops of compute. Right now, bandwidth bottlenecks mean scans take 20 minutes, but the engineering goal is a 60-second full-body map. Interestingly, the prototype uses raw physics simulations rather than AI to build the images.

To deploy them, the self-funded lab is building a massive 25,000-square-foot luxury spa in San Francisco by late 2027. They are pitching it for basic body composition tracking first while working through the long, complex FDA approval process. Read more.

Anthropic brings real-time enterprise dashboards to Claude Code

Claude Code now supports Artifacts. This means your terminal sessions can turn into live, auto-updating webpages for things like PR walkthroughs or interactive dashboards. It uses your full session context from code, logs, and chat history, and updates in real time at a single URL with full version history.

To make life easier for IT admins, Claude Enterprise added centralized authorization for Model Context Protocol (MCP) connectors. Instead of making every employee manually authorize tools like Figma, Asana, or Supabase, companies can now manage access centrally using identity providers like Okta.

Watch: Artifacts in Claude Code: share your work as it happens

Claude Design also got a major tune-up to keep layouts on-brand. You can now import design systems directly from GitHub or local files so the AI doesn't drift from your style guide. A new /design-sync command links your design canvas directly to Claude Code, and the update cuts down on token usage while adding export support for Adobe and Canva.

Meanwhile, Anthropic's internal "Project Fetch" experiments show just how fast its software agents are evolving. In tests using a robotic quadruped, Claude Opus 4.7 completed hardware programming and sensor integration tasks under 10 minutes. That is nearly 19 times faster than a Claude-assisted human team and 37 times faster than humans working alone. The AI wrote less code but found more efficient solutions, though it still stumbled on complex physical tasks like adaptive, continuous ball retrieval.

Anthropic expects to restore access to Claude Mythos and Fable 5 within days as executives work with regulators to quickly resume services.

The company is also addressing the heavy environmental footprint of its systems. Acknowledging the massive power demands required to run these frontier models, Anthropic became the first pure AI startup to join the Frontier coalition, helping boost total pledges to $1.8 billion for long-term carbon removal.

Frontier models and product moves

Agents and the agentic stack

Business, labor, and institutions

Hardware and infrastructure

Research and benchmarks

Robotics

Media and creative AI

Watch

goldfish delivers instant, context-aware message replies on your Mac in your own personal tone with one click.

Swytchcode is a secure execution layer for AI agents to reliably connect, manage, and run 2000+ APIs.

Rendervi provides architectural AI rendering to generate consistent, photorealistic visual designs from your models in minutes.

Qursor delivers precise UI context and code selectors so your AI agent can fix interfaces fast.

Framer 3.0 introduces canvas-native design agents and teamwork branching to help creators build websites faster.

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email or DM us on X!

Thinking of starting your own newsletter? AI Breakfast readers who sign up with Beehiiv receive a 14-day free trial and 20% off for 3 months.