I got an email recently. HHS was congratulating me for my work on UDS+—a project I hadn’t touched in over a year.

They were “proud” of my contribution.

Except I never finished it. I gave up.

And I’m not sad about that. I’m relieved.

The Illusion of Progress

UDS+ was supposed to be a step forward. It was meant to improve how community health centers report their outcomes. A noble cause. But the process was bureaucratic insanity wrapped in compliance theater. Every feature had legal implications. Every spec revision introduced new obligations that had nothing to do with patient care.

And just when I thought I was nearing the finish line, I attended a federal health conference in St. Louis, Missouri. There, Micky Tripathi—National Coordinator for Health IT—proudly announced the finalization of a 1,000+ page spec of new features and obligations to be “forced” onto EHR vendors.

That’s when I realized the truth:

There is no finish line.

Seven Kinds of Homelessness, Zero Kinds of Outcomes

At that same conference, I asked a question during the Q&A:

“Why do we care about classifying seven kinds of homelessness? Do these categories lead to different health outcomes?”

No answer. Silence.

According to the 2024 UDS Manual (Table 4: Selected Patient Characteristics), health centers are required to classify patients as:

Seven kinds of homelessness. But no correlation to health outcomes. No purpose beyond paperwork. These are distinctions without a difference. Just data points filed away to satisfy auditors and report writers.

That’s when I knew: this wasn’t science. This was ritual. Compliance theater, nothing more.

Medicine Is Not Science—and That’s the Problem

Later that day, I asked a second question—this one to the AI panel:

“Why doesn’t the National Science Foundation recognize medicine as a branch of science?”

They didn’t answer that one either.

But the answer is public: it doesn’t. The NSF has seven research directorates—biological sciences, engineering, computer science, geoscience, math, social sciences, and education. Not one for medicine. Because medicine, as practiced in America, resists accountability, predictability, and reproducibility—the very essence of science.

Instead, medicine charges for time, not outcomes. It hides behind authority instead of evidence. It aggregates metrics that are orthogonal to outcomes. Metrics like “homelessness not living in a car” versus “homelessness living in a car”—as if that will yield predictive power in clinical care.

You can’t build real AI on top of non-scientific foundations. If medicine won’t ground itself in reproducibility and measurable outcomes, no amount of machine learning will make it better. It will just make it faster at doing the wrong thing.

Why I Walked Away

I didn’t quit UDS+ because it was hard.

I quit because it was hopeless.

The road was never going to end. Every year, a new spec. Every spec, a new demand. Every demand, a step further from real outcomes.

There’s no innovating inside a system that’s designed to survive its own failure. There’s no saving a machine that feeds on compliance, not care.

The Path Forward

I’m not building for that system anymore.

I’m building for places where innovation is still allowed. Where founders can move fast, ship product, and actually be accountable to users—not auditors. I’m focused on outcomes now. On science. On systems that can prove they work—or be dismantled if they don’t.

If you’re stuck inside a broken system like I was, there’s another path. It starts with choosing to leave.

Want to build something that works?

Schedule a call. Let’s talk.