Building in Public

I spent years keeping too much of the work private.

Private repos. Internal tools. Half-finished ideas I never cleaned up enough to show anyone.

That is changing because keeping everything hidden stopped being useful.

It also stopped matching how I actually think. I have strong opinions about devices, clinical AI, product design, and the difference between real progress and polished theater. Keeping all of that private made the work feel narrower than it actually was.

Why Now

Running IntuBlade, helping build Sayvant, and going through Stanford’s informatics program made something obvious.

The work gets better when it runs into reality sooner.

A paramedic telling me why a device setup is annoying is useful. A physician pointing out where a note generator can quietly lie is useful. Silence is not useful.

I do not mean “feedback” in the vague startup sense where people clap for the idea and tell you it sounds promising. I mean reality showing up early enough to damage the wrong assumptions. That is the useful part.

I also got tired of the way private work can create fake progress. You can spend a long time polishing something in isolation and convince yourself it is better than it is. Then you show it to the first real user and realize half the assumptions were wrong.

That pattern shows up everywhere. It happens in software. It happens in med device work. It definitely happens in AI, where people can get intoxicated by a good demo long before they have built something dependable.

What You’ll Find Here

This site is for the parts of the work that are worth making public:

• Medical devices: what it takes to make a cheap airway device real, not just patentable.
• Clinical AI: where documentation systems fail, what makes them safer, and which claims in healthcare AI are mostly nonsense.
• Physician-built software: the upside and the friction of being close to both the workflow and the code.
• Public work: the projects, papers, and experiments that can stand on their own.

Some of this will be polished. Some of it will be work in progress. That is fine. I would rather show something real than wait for the fake moment when everything looks complete.

That also means some of the writing here will be less tidy than a normal company blog. Good. I am more interested in being accurate about the work than sounding perfectly composed while describing it.

I also think the internet has too much after-the-fact storytelling. Everyone writes as if the path was obvious in retrospect. Most of the time it was not. Most of the time there was confusion, bad guesses, rework, and the slow process of finding out what actually mattered.

That is especially true in healthcare. People love polished narratives about transformation and innovation. Meanwhile the actual work is usually slower, more operational, and more constrained than the public story makes it sound. If I write about a device, a documentation system, or a piece of software here, I would rather describe the friction honestly than wrap it in theater.

The Stack

The site is built with Astro and kept intentionally simple. I want it to be easy to update between shifts, flights, meetings, and the rest of real life.

I am not trying to turn this into a content machine. I want a place where I can write clearly about the work, show the parts that are real, and leave out the usual founder performance.

There is enough founder performance already. Enough vague claims about changing healthcare. Enough websites that sound intelligent for three paragraphs and tell you almost nothing by the end. I have no interest in adding one more of those to the pile.

There is also a more practical reason to do this. I want a public record of what I actually think. Not a cleaned-up version after the fact. Not the investor version. Not the conference version. Just the work, the reasoning, and the places where I think most people in the space are overstating things.

If this site does its job, it should make the work easier to evaluate. People should be able to see what I am building, how I think, where I am skeptical, and what kinds of problems I keep coming back to. That is the point. Not branding. Not optics. Not performance. Just enough public record that the work can stand on its own.