LLM generated posts lack clarity and originality

This title was summarized by AI from the post below.

Full transparency...I use LLMs to help me write posts online. But i'm seeing a trend and maybe its been there for a while but I'm just barely noticing it. Everyone is writing these short sentences or breaking up large paragraphs with single line sentences. Of course there is nothing wrong with that but all the post look the same now...trend of uniformity. I asked LLM to draft 30 days of posts on cyber, systems theory, possibility science and antifragility. On the surface, it looked really great but then once i started posting them, I noticed the same short sentences and the clarity wasn't there. I had to rewrite 90% of what the LLM originally produced. I'm curious, are people just copying and pasting what their LLM of choice produces? 🤔 #LLM #systemtheory #antifragility

I write one line sentences on social media for conciseness and clarity. Attention span on social media is not that long. I typically skip long written post unless its from someone I personally know or not a topic of interest.

Keep it human ✊🏼 I’m already seeing enough slop in other places… 😉 Otherwise, the AI fabric project by Daniel Miessler is pretty interesting to customize your automation with unique prompting via markdown adjustments. https://github.com/danielmiessler/fabric

AI to keep me from rambling, then I edit to make sure the point I intended is the one actually coming across

No, but I see the same trends as well. I guess it would be better if people used it as an idea generator rather than a post machine. Use it to expand on your ideas, not to completely rely on it to think for you.

Yea, I write short one liners myself (no LLM) but honestly the LLM is doing its job if its feeding on how people are posting and adjusting itself to write in that style.

I use AI to generate ideas but I almost always rewrite parts of it/restructure it. If I can spot posts that are entirely AI written, I’m she others can too.

See more comments

To view or add a comment, sign in

Explore content categories