⏱ 4-min read
This issue: AI ran a real chemistry experiment and found something human scientists missed. Taiwan committed $1.5B to AI hospitals. And one prompt turned into a working legal app. Here's everything that moved.
AI Just Ran a Real Chemistry Experiment
For the first time, AI model GPT-5.4, Molecule.one’s Maria AI, and an automated lab system completed a research loop to improve a Chan–Lam coupling. The AI proposed hypotheses, designed reactions, analyzed results, and identified an additive that increased yields, later confirmed by chemists.
Image source: OpenAI
Highlights
An AI model helped design, run, and interpret real wet‑lab chemistry experiments
GPT‑5.4 worked with Molecule.one’s Maria AI and automated lab hardware
The system explored thousands of reaction conditions programmatically
It identified an additive that improved Chan–Lam coupling yields across substrates
Human chemists validated the result independently at bench scale
Why it matters: Scientific AI is shifting from “advisor” to “lab partner.” Instead of stopping at theoretical suggestions, models are beginning to participate in the experimental cycle itself — proposing, testing, refining, and handing humans results that actually work.
Design, schedule, and publish your marketing with one tool
PosterMyWall gives small business owners a full marketing suite: AI-powered design, email campaigns, social scheduling, event promotions, and marketing graphics all in one place.
Build branded content in minutes using 3 million+ customizable templates, then publish directly to your channels without switching apps.
Whether you’re launching a product, promoting an event, or growing an audience, it handles the creative and distribution side so you can stay focused on the business.
No designer, no agency, no problem.
The Supplement in Your Cabinet May Need a Second Look
A new UF Health study found that glucosamine use was linked to faster progression of Alzheimer’s disease. That matters because glucosamine is one of the most common over-the-counter supplements, often taken daily by older adults for joint health.
The latest AIHealthTech Insider also covers Yale moving 700,000 radiology exams into AI-assisted reporting, Philips finding AI saves clinicians 16 working days a year, and Taiwan’s $1.5 billion AI hospital push.
Want it weekly?
8 levels of context maturity in AI-native engineering
More MCPs give agents access to information, not understanding. You agents are still missing the context to generate output you can fully trust. This 8-level model explains where teams get stuck and what fixes it. Join this live webinar on June 24 (FREE) and grab the free assessment to see where your team stands.
🔥Big News in AI
Image source: Reve
🔹Reve Is Putting $100K Behind AI Images
Reve just launched a $100,000 image contest, asking creators to submit their strongest set of 10 images. The challenge is simple: show what you can make when AI image generation is pushed beyond one-off prompts and into a polished creative collection.
🔹ChatGPT Just Made Scheduled Tasks Easier
OpenAI is rolling out a new Scheduled page for ChatGPT tasks.Scheduled tasks are now faster, more reliable, and easier to manage across web and mobile for Go, Plus, Pro, Business, and Enterprise users. That means ChatGPT is moving from “answer me now” to “help me later, automatically.”
🔹Claude Code Shows Who Wins With AI Agent
Anthropic analyzed around 400,000 Claude Code sessions and found that coding skill may matter less than people think. In most sessions, humans made the big planning decisions, while Claude handled most of the execution — writing code, editing files, running commands, and figuring out how to get the work done.
Cap table tools founders and finance teams love
As you grow, you need a cap table management platform that you can trust to keep up.
From issuing grants to 409A valuations or ASC 718 reporting, Pulley gives you the tools to manage equity. Experience a platform built for you: transparent, reliable, and designed to put your company first.
Claude Design Is Becoming a Real AI Design Partner
Claude is rolling out new updates to Claude Design, and the big shift is that it can now stay closer to your actual brand system. You can import your design system from a repo, design files, or codebase, then Claude builds with your real components and checks its output before showing it to you.
Image source: Claude
The editor is also getting more practical for daily design work. You can drag, resize, and align elements directly on the canvas, sync projects with Claude Code, hand designs off to build, and export to PDF or PowerPoint.
One Prompt Can Now Build a Working AI App
Base44 takes one specific idea and turns it into a real web app — upload flow, UI, AI analysis, results screen, and deployment included. We used it to build ClauseCheck, a simple contract-review app that summarizes legal documents, flags risky clauses, and suggests questions to ask before signing.
The workflow covers how to define a narrow problem, write the full Base44 prompt, structure the app screens, and use built-in AI without setting up APIs, code, or hosting.
Anthropic Just Put $350M Behind the AI Job Shift
Anthropic is funding a $200M Economic Futures Research Fund and a $150M Claude Corps fellowship — a clear signal that AI job disruption is no longer theoretical. At the same time, Google’s free 5-day AI Agents Intensive is back, with 1.5M past learners and a new focus on vibe coding.
This Sunday Special shows where to move next — two open hackathons, free AI courses, a no-code legal app built in 30 minutes, and six AI career opportunities. The ladder is changing fast, but the new openings are still visible if you start building now.
Want it every Sunday?
New here? Get the next issue Thursday.
70,000+ AI builders read The AI Entrepreneurs every week.
