The most important shift right now is not a single breakthrough model or product, it is who is actually building. Clinicians are beginning to move from end users to creators, using AI to turn everyday frustrations into working tools. For the first time, the distance between “this should exist” and “I built this” is short enough to cross without a technical background.

Here’s what stood out:

Plus: Tool of the week

Let’s dive in.

🧑‍⚕️ ChatGPT for Clinicians: OpenAI launched ChatGPT for Clinicians for verified U.S. clinicians, with support for documentation, medical research, and cited medical answers. The bigger story is access, because individual clinicians can now experiment without waiting for a hospital-wide rollout.
Why this Matters: This is exactly how we start closing the gap between clinical insight and technical execution.

🛠️ Doctors Coding Clinical Tools: Healthcare IT News highlighted a shift I think every clinician should pay attention to, AI is becoming simple enough for physicians to build custom workflow tools themselves. That does not mean everyone needs to become a software engineer. It means our clinical judgment can finally shape prototypes directly.
Why this Matters: The person who understands the broken workflow can now start building the fix.

🧠 Aidoc Raises $150 Million: Aidoc raised $150 million to expand its clinical AI work in imaging, where its tools help analyze CT scans and X-rays for emergency triage and other findings. The company has 31 FDA clearances, which shows how far radiology AI has moved from novelty to infrastructure.
Why this Matters: Clinical AI is becoming a core layer of hospital operations, not a side experiment.

🩺 LLMs Still Struggle With Clinical Reasoning: A JAMA Network Open study tested 21 off-the-shelf LLMs across clinical workflow tasks and found they still have important limitations, especially in early diagnostic reasoning. The message is not “AI is useless,” it is “do not hand it the wheel.”
Key Takeaway: Clinicians need to guide, test, and constrain these tools before trusting them in care.

📚 Show Us the Evidence: Nature Medicine argued that medical AI needs stronger evidence of clinical value, not just impressive demos or retrospective performance. That is especially relevant for clinician-builders, because small tools still need thoughtful validation.
Key Takeaway: Build fast, but evaluate like patient safety depends on it, because it does.

⚖️ AI Licensure Framework: A new JAMA Perspective argues that autonomous clinical AI may need its own licensure framework as these systems begin overlapping with physician cognitive tasks. This is where the conversation gets serious, because autonomy changes accountability.
Why this Matters: Clinician-builders need to know when a tool is support, and when it starts acting like a clinician.

🔬 FDA Focuses on AI in Drug Development: The FDA reiterated that AI and machine learning are increasingly used across drug development and that risk-based regulatory frameworks are needed to support innovation while protecting patients. This is the quieter but important backend of healthcare AI.
Why this Matters: AI governance is becoming part of the full medical product lifecycle.

Zapier connects tools together so a form submission, email, spreadsheet row, or database update can trigger an AI-assisted workflow. I use it to think through the boring glue work that makes a clinical idea actually usable. Prompt to try: “Design a Zapier automation that takes a completed intake form, summarizes the key clinical points, flags missing information, and sends it for clinician review.”

Agentic AI is getting easier and easier to implement. Automated workflows make our jobs easier. Try it out for yourself!

For years, physicians were told to stay in our lane while other people built the businesses and tools around us. AI gives us a chance to stop waiting for permission. Start small, build safely, test honestly, and make the thing you always wished existed for your patients.

What are you building? I have some half-baked ideas. Let me know yours and lets collaborate!