[gkoberger]

I do agree... I sometimes use worse grammar (like that ellipses) and leave in typos just so my comments feel more "real" now.

[goodmythical]

fun fact, grok and kimi are both pretty good at emulating "chat" responses with any number of prompts.

"respond like a twitter user", "pretend like we're texting", etc

[Imustaskforhelp]

> fun fact, grok and kimi are both pretty good at emulating "chat" responses with any number of prompts.

> "respond like a twitter user", "pretend like we're texting", etc

+1 to it. I actually had given a response to the above parent comment itself using Kimi and I would've said that its (sort of) a good emulation fwiw.

[xeckr]

Show HN: A RL suite that teaches LLMs to introduce typos at just the right frequency for the output to appear more human.

[Imustaskforhelp]

soon were gonna be the ones adding random typos and grammer errors just to blend in. i skip apostrophes and mispell words on purpose already. its strange how fast sloppy writing starts feeling natural

(This above line itself was written by AI itself: https://www.kimi.com/share/19c96516-4032-8b73-8000-0000f45eb...)

I don't know if worse grammar could make a difference aside from removing false negatives (ie. nowadays people with good grammar are questioned if they are LLM's or not) but this itself doesn't mean that worse grammar itself means its written by a human. (This paragraph is written by me, a human, Hi :D)

[pvtmert]

Honestly, first paragraph sounds more human and sincere for sure.

Also adding better "context" into the discussion, than the usual claims/punchlines of marketing-speak.

Maybe it's not exactly the grammar itself but also overall structuring of the idea/thought into the process. The regular output sounds much more like marketing-piece or news-coverage than an individual anyway. I think, people wanna discuss things with people, not with a news-editor.

[_verandaguy]

Same here, but it'll be a cold day in hell before you see me using the dreaded double-period-bang..!