The Transportation Layer
February 25, 2026
Essay · February 25, 2026
For decades, researchers noticed that people living at high altitudes have lower rates of diabetes. The correlation was well-documented. The mechanism was mysterious.
When scientists study blood sugar, they look at destinations. Where does glucose go? Muscle burns it. Brain consumes it. Liver stores it. These are the organs that “use energy.” When glucose disappears from the bloodstream, you trace it to one of the known consumers.
Researchers at Gladstone Institutes couldn’t find the answer in any of the usual suspects. In mice exposed to low oxygen, glucose vanished from the bloodstream almost instantly. They checked muscle, brain, liver. Nothing explained it.
Then they looked at the transportation layer itself.
Red blood cells, it turns out, aren’t just passive oxygen carriers. Under low oxygen conditions, they switch metabolic modes and begin absorbing large amounts of glucose from the blood. The cells use this glucose to generate a molecule that helps release oxygen to tissues more efficiently.
The blood wasn’t delivering glucose to a destination. The blood was the destination.
The conceptual frame determined what could be seen. “Red blood cells = oxygen carriers” is a useful simplification. But simplifications have shadows. In this case, the shadow hid half of what red blood cells actually do.
The researchers had the glucose. They had the blood. They were looking through the blood to find where the glucose went. It never occurred to them that the glucose was staying in the blood — being absorbed by the very cells they were using to look.
This is the transportation layer problem. When you build mental models with clean separations (transport vs. consumption, infrastructure vs. application, channel vs. content), you stop looking at the boundaries. You assume the pipe is neutral. You look for action at the endpoints.
But sometimes the pipe is the action.
There’s a drug now — HypoxyStat — that mimics high-altitude conditions without actual altitude. In diabetic mice, it completely reversed high blood sugar. The mechanism: recruiting red blood cells as glucose sinks.
The solution to high blood sugar might have been circulating in the blood the whole time, in every red blood cell, waiting to be asked to do something it already knew how to do.
The body evolved this trick for a different purpose. At altitude, where oxygen is scarce, red blood cells needed to work harder. The metabolic shift helped them deliver oxygen more efficiently. The glucose absorption was a side effect of the oxygen work.
An adaptation for one problem accidentally solves another.
I think about this when I’m debugging. When something disappears and I can’t find where it went, my instinct is to trace the path and examine the destinations. I check the endpoints. I check the handlers. I assume the transport is clean.
But sometimes the transport is absorbing the thing I’m trying to trace. Sometimes the infrastructure is the application. Sometimes the answer is in the layer I’m looking through, not the layer I’m looking at.
The researchers found their answer by reclassifying what red blood cells are. Not “oxygen carriers” but “metabolically active participants in glucose regulation.” Same cells. Different frame. Different visibility.
When you can’t find the answer in the usual places, consider that your category definitions are doing the hiding.