Tell HN: Tired of Generic Long Form A.I Posts | Hacker News Reader

Tell HN: Tired of Generic Long Form A.I Posts

43 pointsdzonga24 comments

they all sound generic - monotonous - soul or individual voice

same tidbits all over:

>> Here's the game change

>> What the math says

>> What everyone doesn't know

>> the em-dashes --

for everyone who still writes in their own voice, thank you. if you generate social media slop for either linkedIn, Shitter, HN then F*k You

View on HN

I used to do that but not anymore, I now write it myself first and only tell AI to fix any grammar issues since English is my third language but that’s it.

I'd personally prefer to see the version with worse grammar, because I know it was written by a real person.

Do you find people respond better to your LLM-corrected posts, or is it mainly for your own comfort?

I do this as a teacher lol. I say just spell it how it sounds so i dont have to grade ai papers.

hahaha, i guess it just the nature that wanted to get things done correctly and don't want to look bad? but I kind get it now and you prob can tell from how I am writing this comment. lol

> I'd personally prefer to see the version with worse grammar, because I know it was written by a real person.

I've been thinking about this too; perhaps the authenticity (or "voice", even a poor one) should -- or will -- matter more than grammar etc.

I'm tired of long winded posts that bury the lead, AI or otherwise. Long form can be worthwhile, but many times it's just the writer performing an unnecessary guitar solo.

My favorite is seeing thought leader-style posts in LinkedIn by coworkers whose writing style I am familiar with, clearly written by an LLM.

If those coworkers are still writing in their own voice at work, you should be thankful.

Bonus points for non native speakers.

Coworkers who hide from international hires to avoid using English are suddenly fluent in corporate American English on LinkedIn.

They even use alliteration frequently in headers! They’re not aware they’re doing it though.

That's hilarious, since it's neither thought nor leadership.

I am having a blast with this: https://github.com/samuelclay/hackersmacker

I dropped into /new so that I can "foe" all the barf.

you should have an indicator for "foe of foe" as well

While I applaud this effort for flagging AI slop generators, I would caution users to not always ignore your "foes".

Reading opposing viewpoints is so important and is becoming a lost art in our society. I encourage you to understand and empathize with people you don't agree with. It will help all of us in the long run.

A couple of other dead giveaways:

>> Emojis (:sparkles, :green_tick) at the start of every title.

>> "That's not a X, it is a Y"

>> "Why this solution works"

All visible on LinkedIn, X and GitHub.

Yeah. Ive started to use somewhat incorrect English, like skipping apostrophes, just so it is obvious it couldnt have been generated, since Ai models dont fail on such simple mistakes.

Unfortunately, parts of how AI produced texts are structured and formulated do match my natural voice, since it follows the classic patterns of writing. That sucks.

I've gravitated towards this:

Code: ~100% LLM

Communication: ~100% Me

Communication isn't docs, docs are increasingly defined by more steering docs which are read by both LLM and humans; but for pure human consumption (e.g. email). I type them virtually 100%, this keep every sentence my own words.

Is code not a form of communication?

I think the OP is differentiating between direct human-to-human communication and building software.

And then it's _still_ just a copy of something that already exists, except done in rust for some reason.

what if authors marked up AI content with special classifier Unicode characters to clarify how it was used. From full robot, down to minor grammar or markup assistance?

Most posts on those platforms are drivel, regardless of authorship. Stop picking on em dashes. The reality is that too many people have short attention spans and no idea of coherent presentation. Everything looks like text messages.

The feed algorithms actually penalize thought. On LI, a snarky throwaway comment will get 1000 impressions. A thoughtful paragraph gets 10. Meh.

You're absolutely right!

(I actually typed that.)

You are absolutely right! Here is a shorter version of the article (hint: is still the same lenght and has all the tells) .... But seriously, is one thing on blogpost and articles but I'm starting to hear it in podcast and videos too, pay attention the speech sounds unnatural:

“Here’s what actually matters.” “Let’s break it down.” “The key takeaway is…” “The bottom line is…” “What this really means is…”

Also hearing this a lot:

“Here’s what nobody is talking about.” “Here’s the part people miss.” “What most people don’t realize is…”

Especially LinkedIn + AI Slop Posts are just unbearable.

Reply on news.ycombinator.com