Kierkegaard found an application for the three-em dash here: “I have just now come from a party where I was its life and soul; witticisms streamed from my lips, everyone laughed and admired me, but I went away ⸻ yes, the dash should be as long as the radius of the earth's orbit ⸻ and wanted to shoot myself.”

I think this is cool and am happy to see the post to learn more about punctuation. LLMs have really brought the en/em (and beyond because there are so many) dash into the spotlight in a negative way. At a previous dev job I handled copy being sent for translation and got feedback from writers about inputting strings with the incorrect dash.

This is industry-standard punctuation with real use cases, obviously there's a saturation point but that is more LLM induced than anything else.

From a coding standpoint I'm surprised devs are not more interested in punctuation like this because there are so many different operators and syntax across programming languages.

I'm going to make a super-slop model ⸻ train it on text that gpt-4o-mini improved five times ⸻ and this is going to be the dash that it uses.

Wait till you see Asterism and “Cyrillic Letter Multiocular O”

Sloppy: "–"

Sloppier "—"

Sloppiest ⸻