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'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.

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

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

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.

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…”

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.

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

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.)

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.

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?

Especially LinkedIn + AI Slop Posts are just unbearable.