The third footnote in that article, in its entirety:

> Claude, by the way, estimates that 30-40% of all mashed potatoes eaten in the US are the instant kind. ChatGPT says 25-35%.

Is this what passes for a reference now?

The text that the footnote is attached to is:

"Large Language Models can gall on an aesthetic level because they are IMPish slurries of thought itself, every word ever written dried into weights and vectors and lubricated with the margarine of RLHF." I infer 'IMPish' as meaning 'like Instant Mashed Potato'.

I read that footnote as a somewhat oblique criticism of two LLMs, rather than on the statistic itself - which may indeed have just been fabricated by the LLM as opposed to an actual statistic somehow dredged from its training data, or pulled from a web search.

Really depressing to think that people trust statistics from these models and soon the models will be ingesting statistics they themselves made up as training data.