A person leans on the titanic intellect of a trillion dollar company's most fearsome LLM, only to be corrected by a random commenter with a link to Wikipedia.

We live in interesting times.

Well Gemini and Wikipedia produced the same conclusion.

Linguistic questions were one of the first knowledge categories I trusted LLMs to be able to answer well - quite literally being models of language. It would be pretty shocking for a ~frontier model to get something like that wrong in the last like 3 years at least.

I was referring to the LLM giving the wrong definition of grain (i.e. not an individual grain), which while not as popular as it once was, is still widely used for ammunition.

I don't know how Gemini is wrong so much more often than anything else. Its actually quite impressive.

And the LLM likely referenced Wikipedia

And yet still misunderstood how much a grain is.

It's not an old English measure, because Latin has "granum salis" (a grain of salt) in medical authors and Pliny. There's no indication that it was a measure; it was a cube or crystal of salt.

Interesting, although could those not be different things? The linked page - https://en.wikipedia.org/wiki/Grain_(unit) - shows that it absolutely was a foundational measure in England. I guess it's probably down to the etymology, which would tell you when and where the phrase came from, then you could tell whether it from the English or ancient meaning, although it's not especially relevant to the meaning. The linked article does call it an "English idiom", so you'd need to show a similar ancient idiom for that to be the source.