this post was submitted on 19 Apr 2025
11 points (86.7% liked)

Typography

287 readers
2 users here now

Type design, setting, fonts, etc.

Civility is aggressively enforced: do not be rude, troll-y, flame-bait-y, etc. If someone else is, please report, block as necessary, and move on.

Comments that cannot reasonably lead to discussion or which have no analysis ("first!", "lol", [string of emojis]) will be deleted. Please do not post them.

founded 5 years ago
MODERATORS
all 14 comments
sorted by: hot top controversial new old
[–] Diplomjodler3@lemmy.world 7 points 3 months ago

I use em dashes all the time. How dare you call me a robot, filthy meatsack?

[–] davel@lemmy.ml 7 points 3 months ago (1 children)
[–] otacon239@lemmy.world 4 points 3 months ago* (last edited 3 months ago) (1 children)

I was going to say, Mac and iOS make it pretty easy to get to and it looks better than a regular dash in context. Then again, I’m one of the few people that I know that only writes in full punctuation at all times.

[–] AnUnusualRelic@lemmy.world 4 points 3 months ago

They're also easy to enter with a compose key which should be part of any normal Unix-ish desktop setup.

[–] alt_xa_23@lemmy.world 4 points 3 months ago

With default settings, libreoffice will autocorrect hyphens to em dashes in some cases

[–] j4k3@lemmy.world 4 points 3 months ago

This is so stupid. I have my AI server up right now in another tab and setup for creative writing. I currently have these tokens banned: 3221, 265, 28845, 382, 650, 28716, 16913, 1101

Those tokens correspond to specific word fragments and punctuation. Most are due to some sentence starting patterns I don't like. I can ban any tokens and the model will not use them. All models have this functionality in the API for any user that actually knows what they are doing at a basic elementary level. Like the tokens for - – — are 28773, 764, and 1040, for my mixtral model.

[–] MadMadBunny@lemmy.ca 1 points 3 weeks ago

I use them all the time.

[–] Opinionhaver@feddit.uk -1 points 3 months ago

It's not proof but it's a red flag. However, I assume that everyone who uses AI to assist with their writing, such as myself, probably edits them out for this very reason.