Why I Started Writing This Down

Private work can lie to you. It lets you confuse motion with progress: private repo, internal note, half-finished thing that would be great if you just got one more clean pass at it. I have a lot of those. Some are useful. Most are hiding places.

I started this site because the work got too spread out to keep pretending it was one clean lane. Emergency medicine. IntuBlade. Sayvant. Stanford. Clinical AI. Airway devices. Benchmarks. Small tools built between shifts because the thing I needed did not exist yet. That is not a tidy brand. Good.

The Work Gets Better When Reality Hits It

The best feedback I get is usually not polite. A medic says the setup is annoying. A physician finds a note that sounds right and is wrong. A hospital leader likes the clinical idea and then tells you the budget process will murder it before anyone gets trained. That is useful.

The fake version of feedback is people saying the idea sounds interesting. The useful version is someone damaging the bad assumption early enough that you can still fix it. That has been true across almost everything I have built.

IntuBlade got better when EMS people told me what would actually survive a truck, a station, a bag, a shift change, and a procurement process. Sayvant gets better when physicians show exactly where the chart is wrong, where the model overreached, and where the product created cleanup labor instead of removing it.

Clinical AI gets better when the demo stops being the goal.

What I Am Putting Here

This site is for the work that can stand in public without a pitch deck holding it up: medical devices, clinical AI, physician-built software, research, benchmarks, and the public artifacts that make claims easier to inspect.

Some of it will be polished. Some of it will be messy. That is fine. I would rather show a real edge than wait for the fake moment when everything looks complete.

What I Am Not Doing

I am not trying to turn this into founder theater. There is enough of that already. Enough vague language about changing healthcare. Enough websites that sound intelligent for three paragraphs and leave you with nothing you can actually evaluate.

Healthcare is especially bad at this. People love the clean narrative: unmet need, elegant solution, transformation, scale. The actual work is slower and more operational. It is cost, training, trust, shelf space, chart review, billing behavior, procurement, clinician fatigue, and the gap between what a demo can show and what a department can tolerate.

That gap is where I spend most of my time, so the writing here should feel closer to the work. Less inspirational. More specific.

If I write about an airway device, I want to talk about why a charger disappears, why cleaning workflows kill adoption, and why a cheap tool sometimes matters more than a better-looking one.

If I write about clinical AI, I want to talk about what happens when the model invents one physical exam finding and the physician stops trusting the whole note.

If I write about software, I want to talk about the part where the code meets the person who is tired, behind, and still responsible for the outcome.

The Stack Is Intentionally Boring

The site is Astro. Static pages. Markdown. Minimal moving parts. I wanted something I can update between shifts, meetings, flights, and the ordinary mess of running companies while still practicing medicine.

No content machine. No brand funnel. Just a public record of what I am building, what I think is overhyped, and where the work is actually hard.

If this site does its job, a reader should be able to answer three questions fast:

What has he built? How does he think? Where does he call bullshit?

That is enough.