Shockingly bad journalism here. The 'hoax' that Pullum wrote about was the specific claim that Inuktitut has a lot of words for snow. This is false, and the referenced paper agrees (pointing out that "discussion of lexical elaboration has been tarnished by unscholarly claims about Inuit words for snow"). The detailed results of the paper are perhaps of some interest, but at base it is making the very unremarkable claim that languages differ in how many vocabulary items they map to different areas of conceptual space. It could hardly be otherwise.

English has plenty of words and phrases for snow, depending on if you're describing falling snow or snow on the ground.

Just talk to anyone who skis, snowboards, or lives someplace with a lot of snow.

Explains why multilingual llms are so much more capable. If each culture chooses to capture different conponents of the world at a higher resolution in their language based model of the world, then it makes sense to me that multilingual llms should understand the world on a deeper level than monolingual ones, as they need to approximate less.

> He likened it to the xenomorph from Alien, a creature that “seemed to spring up everywhere once it got loose on the spaceship, and was very difficult to kill.”

If only there was a better analogy to describe the viral nature of something.