The new chemical treatment essentially removes the lignin from wood. The precise nature of the process isn't mentioned in the paper, which suggested it might be nightmarishly complex or involve extremely toxic chemicals.

I'll go on a limb and assume that treatment of wood is some derivative of Kraft process (one for making paper pulp), likely using sodium hydroxide and sodium sulfite to chop the lignin to make it easier to remove. Boiling in hydrogen peroxide (with or without metal catalyst) further degenerates pre-shredded lignin via an analogue of Dakin reaction, which is also well understood – and is especially efficient against free-phenolic bits of lignin obtained in most variants of Kraft process.

What's left is likely susceptible to moisture, as the original publication makes it a point worth mentioning numerous times. You let that cellulose soak up too much water, and there goes your structural strength. They even had to treat it in fluorosilane to make it superhydrophobic when testing outdoors. In Arizona.

It's damn interesting, but I'm sceptical about it being viable outside of extremely arid places.

EDIT: Here's a great video explaining similar chemistry behind making transparent wood, which IMHO is a fascinating material:

So, I finally had the time to go through the article's supplementary material, and I have to say that my previous point about susceptibility to moisture might not be as valid as I thought. While I'm not yet 100% certain about the specific, fluoro-silane bonds to hydrophilic OH groups throughout the volume of the sample instead of being a simple coating (which I wrongfully assumed), and they have treated the delignified wood sample in it before the pressing took place. They also conducted emissivity tests on both hydrophobic and untreated samples to see if it made any difference, which showed negligibly changed spectra. Anyway, while doing my due diligence I also found this fun paper in references, and will probably write something more once I'm done with those.

My bad, Hubski. I shouldn't have written my previous comment hastily, even if the Ars Technica author clearly didn't go through the trouble himself either.

Still think that the material is pretty damn neat, though.