[justinl33]

The comparison to weather satellites is misleading and oversells this technology. Weather radar works by detecting water droplets at known atmospheric heights - it's a fundamentally different problem than trying to measure glucose concentrations in blood through layers of tissue. The real breakthrough here isn't the radar tech (which has existed for years), it's the machine learning pipeline that can extract meaningful glucose data from extremely noisy radar returns.

[DoctorOetker]

I don't believe the claim is radio echolocation, but radio spectroscopy.

I didn't read the paper yet, but I predict from the comments in this HN thread, that the proposed system is essentially a dielectric spectroscopy setup optimized for glucose detection (or any number of proxy byproducts/complexes/etc..)

Check this wikipedia page for example: https://en.wikipedia.org/wiki/Dielectric_spectroscopy

Look at the picture on the right, from typically lower to higher frequencies there is the motion response of ions, the reorientation response of molecules with a dipole moment, the excitation of vibrational modes in a molecule and the electronic excitation of electrons switching orbitals...

EDIT: since I have not read the article, I do not vouch for its authenticity (above my paygrade)