The $600 Billion Problem Hiding In Plain Sight (And Why I'm Building RareGap)

# Weeknotes: The $600 Billion Problem Hiding In Plain Sight (And Why I'm Building RareGap)

If you've spent any time in biotech, you've probably felt it - that mix of awe and frustration that defines rare disease research. On one hand, it's heroic science. On the other, it's chaos wrapped in bureaucracy.

Rare diseases are individually *rare*, but collectively, they form an invisible superpower.

**300 to 446 million people** - 1 in 17 humans - are living with something you've likely never heard of.

And of the 7,000-10,000 known rare diseases, **95% have no FDA-approved treatment**.

That's not a rounding error.

That's a systemic failure - scientific, economic, and moral.

And that's the problem I'm tackling with **RareGap**.

---

## The 7-Year Ghost Hunt (a.k.a. The Diagnostic Odyssey)

Imagine this: you're sick, but no one knows what you have. Not your doctor, not the specialists, not the tests. On average, it takes **4 to 7 years** for rare disease patients to get a definitive diagnosis.

That's seven years of false starts, dead ends, and "we're not sure yet."

The clue - the one genetic marker, lab anomaly, or family history pattern that holds the answer - exists somewhere. But it's **scattered** across PDFs, case reports, and decades-old registries.

It's a data problem, not a science problem.

And that's where things start to unravel.

---

## The Scientific Haystack Is Filthy and Fragmented

Here's what most people don't see: rare disease research isn't just rare - it's *disorganized*.

Each insight sits on a different island:

* A 2013 French study that never made it to PubMed.

* A registry managed by one hospital in Finland.

* A trial dataset sitting in a forgotten Dropbox.

If you're a researcher, you spend **150+ hours** manually connecting dots - combing through scattered evidence, deduplicating papers, cross-referencing outcomes - just to produce a single usable insight.

At $50/hour, that's **$7,500 of human effort** before you even form a hypothesis.

Multiply that across 7,000 diseases, and you start to see the absurdity.

AI is supposed to fix this. But generic LLMs don't know which parts of a 12-page paper matter. They hallucinate when precision matters most. So the "AI revolution" hasn't reached this frontier - yet.

---

## The Repurposing Gap (Billions Sitting on Shelves)

Here's the wild part:

I've discovered thousands of **approved drugs** already on the market.

Statistically, some *should* work for rare diseases - but no one knows which ones.

Why? Because the tools to match drugs to new indications are brittle, fragmented, and outdated. The mapping between molecular targets, clinical phenotypes, and gene pathways is buried across incompatible databases and paywalled journals.

So drug repurposing - the fastest path to saving lives - becomes guesswork.

That's the gap I'm attacking first.

---

## The $600B Market No One's Building the Right Tools For

Despite all this friction, the rare disease market is booming:

* **$216B in 2025**, projected to exceed **$600B by 2034**.

* **20% of all global prescription sales** will come from orphan drugs by 2030.

* Governments practically *beg* companies to solve this: tax credits, fee waivers, and up to **10 years of market exclusivity**.

So why isn't there more innovation?

Because data fragmentation kills velocity.

Researchers know where to look - they just can't see fast enough.

The bottleneck isn't intelligence; it's time.

---

## What RareGap Is Building

RareGap is an **AI system that maps the unseen research gaps** in rare disease science - the ones hiding between PubMed abstracts, clinical trials, and drug repurposing databases.

Think of it as a radar for discovery:

* It ingests fragmented scientific evidence across languages, registries, and literature.

* It surfaces actionable "research gaps" - areas where biomarkers, outcome measures, or repurposable assets align.

* It turns weeks of detective work into **minutes of clarity**.

A human researcher spends 88 hours building a literature map.

RareGap does it in four minutes - with citations, metadata, and connection reasoning.

That's not automation. That's acceleration.

---

## Why I'm Building This

I didn't build RareGap because it sounded noble.

I built it because it's *inevitable*.

If AI is going to touch every field, then **life sciences is the one where it literally saves lives**.

And applying serious product thinking to this frontier seem inevitable.

Everything looks like a research paper. Nothing feels like a system built to move.

I'm building that system.

I'm not aiming to cure every disease tomorrow. I'm aiming to collapse the timeline between *possibility* and *progress* from years to hours.

Changing the world can start with one person doing something small but meaningful.

---

**Status:** Building. Early winter in 2025.

**Mission:** To help researchers close the rare disease gap - one discovery at a time.

Because if I can surface what's been hidden all along,

I might just save a few million lives in the process.image.png