Need help related to AI article prompting.

GhostSEOO

GhostSEOO

Newbie
Joined
Sep 9, 2025
Messages
30
Reaction score
13
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
 
Advertise on BHW
GhostSEOO said:
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
Click to expand...
I think the issue is less the ai but more that it wants to sound “good” as opposed to clear. In editing AI content, I cut anything that does not add meaning. If a sentence does not make the idea easier to understand it goes. I have also found that if u give the AI a specific audience and ask it to write as if explaining the topic to one real person u get much more natural results than asking for high quality writting.
 
voxeltoon said:
I think the issue is less the ai but more that it wants to sound “good” as opposed to clear. In editing AI content, I cut anything that does not add meaning. If a sentence does not make the idea easier to understand it goes. I have also found that if u give the AI a specific audience and ask it to write as if explaining the topic to one real person u get much more natural results than asking for high quality writting.
Click to expand...
Thanks for taking out some of your time.
I so the aame just rewrite the whole sentence as they are not perfectly explained. I gave him the audiences, tone and specific target topic with outline still the results was same maybe i need to recheck my prompt and see if i can find something that is not helping me and sabotaging me.
 
Habib-SEO said:
Free AI models love filler. Tell it: "Plain English. No metaphors. No fluff. Every sentence must add value."
Click to expand...
Did that most of the time, result were same. I guess i need to take some time to learn how to give ai good prompt. Thanks btw
 
Try different ai model.
Ai models are like everything is fixed inside ai model. if the ai model is weak and outdated it may not give good results
 
The issue usually isn't the model, its that the prompt is asking it to "write well" instead of giving it clear constraints.

A few things that helped me, tell it to use plain, concrete language and avoid poetic or metaphorical phrases, explicitly ban filler like "imagine", "picture this", "heavy eyed", "journey", "unlock" etc. Make it explain concepts directly instead of creating atmosphere and have it write a first draft than run a second prompt that critiques and rewrites anything that sounds like AI or vague marketing copy. If you have examples of your writing use several samples instead of just one so it learns the overall pattern better.

For your example I would simply write, "This is the transition between wakefulness and sleep, known as the hypnagogic state, during this period your thoughts become less analytical and more associative making it easier to relax and process ideas, theta brain waves naturally occur during this stage as well as during deep meditation." Its much clearer and gets the point across without the unnecessary fluff.

Hopefully this helps :D
 
GhostSEOO said:
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
Click to expand...
AI models needs a hell of a lot of data. Try feeding it more input. And also run through two models one to generate and one to cut whatever you dont want in it. So that way you automate what your doing manually.
 
GhostSEOO said:
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
Click to expand...
Totaly agree, AI tends to drift into "poetic" filler when it doesnt have a clear voice to follow. What helped me was giving it 2-3 paragraphs of my own past writing as reference insted of just describing the style in words. Cuts down the generic fluff alot.
 
One thing that actually cut AI fluff for me was adding a negative example directly in the prompt. Paste an example of bad output and write "never write like this" right next to it, AI respond much better to concrete examples than abstract instructions like "be clear." also try ending prompt with "if you catch yourself writing a metaphor, delete it and restate as plain fact instead." That one instruction alone reduce the poetic filler significantly.
 
i think the biggest mistake is expecting AI to get the tone right in one go. I have had better results treating the first draft as raw material. If a sentence sounds like something I would never say in a real conversation, I either rewrite it or tell the AI exactly why it feels off.After a few iterations, the output ususally gets much closer to what I want.
 
I start by generating a large list of titles first. Then I remove the ones that are too similar, would produce nearly identical content, or are too generic.

After that, I ask the AI to assign each title a core topic and a writing tone. Once that's done, I give it competitor URLs and ask it to generate the article structure based on those. Finally, I generate each article separately instead of creating everything in one go.

I never generate full article in one go. I generate each part of article separately. Just so that i can check and reevaluate it myself.
 
Do not request AI to write well. Ask it to gather facts and write based on those facts. In addition, make it stop using unnecessary fillers, figures of speech, and vague descriptions of emotions without any justification
 
GhostSEOO said:
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
Click to expand...
totally agree i faced the same issue but what helped me was to give it a clear self made structure and 2-3 lines for introduction and first h2 heading
 
I ran into the same problem. AI loves adding dramatic or poetic phrases that sound nice but don't actually tell the reader anything. What helped me was telling it to use simple, direct language and explain things the way you'd explain them to a friend. I also ask it to rewrite any sentence that sounds vague or doesn't add real value, which cuts down the fluff a lot.
 
John Chin said:
i think the biggest mistake is expecting AI to get the tone right in one go. I have had better results treating the first draft as raw material. If a sentence sounds like something I would never say in a real conversation, I either rewrite it or tell the AI exactly why it feels off.After a few iterations, the output ususally gets much closer to what I want.
Click to expand...
How many iterations though?
 
I have found that the model usually reflects the examples you feed it. If the prompt is too broad then it tends to fall back on generic or overly descriptive phrases like this.
What helped me was giving it 2-3 real writing samples, telling it exactly what to avoid and asking it to use simple and direct language. It is also easier to edit a rough draft than expect perfect output in one go.
 
It really feels like you are battling against the natural poetry writing tendency of the model. The model will continue to go towards poetic fillers if not restrained properly because it statistically occurs more during training.
The approach here doeis not involve giving tone based instructions bt to ensure that the model adheres to the structure without any poetic elements like metaphorical laenguage or description of emotions, among others. Moreover, it needs to be done step by step, from research to outline and finally sections
 
Our flow looks like this. Initial generation with ChatGPT, -> Claude as an editor -> Manual check.
 
GhostSEOO said:
Hey, BHW members
I need your help with ai article writing. I let my AI do the research and writing, and I gave him a detailed prompt on what to follow and not. But in the end, it still writes generic shit that has no meaning to it. I usually rewrite myself, but sometimes it takes a lot of time to rewrite in my style. I even gave him my style to replicate, but he still wrote something else. ( I use claude and chat gpt Free*)
Look at it...
(Here's what it feels like: heavy-eyed and inward, like the moment right before a dream starts, the fog right after you wake up before you've opened your eyes. Sleep scientists call this the hypnagogic state. Thoughts stop feeling sharp, and time gets a little loose...
This is the state from last section, the one where the critical faculty softens and stops arguing with everything you say. Theta shows up on its own during deep meditation, in the minutes right before sleep, and sometimes right after you wake up, before the day fully starts.)

"What's the point of heavy-eyed and inward"
How do you guys create output which avoid this type of writing?
Click to expand...
This type of output is usually generated because the model is attempting to sound “smart” instead of being useful. You need to force it to employ concrete, not stylistic constraints.
Some suggestions for what could work well: “Do not use any metaphors or flowery language, write in a way that a tenth grader can understand, one concept per sentence.”
 
Advertise on BHW
You must log in or register to reply here.
Sign up now!

Main Menu

Home Main Forum List Black Hat SEO White Hat SEO BHW Newbie Guide Blogging Black Hat Tools Social Media Downloads

Marketplace

Account Selling Content / Copywriting EBooks / Methods Hosting Misc Proxies For Sale SEO - Link building SEO - Packages Social Media Social Media - Panels Web Design More

Making Money

Affiliate Programs Hire a Freelancer IM Journeys Making Money Pay Per Click Site Flipping Want to Buy

BlackHat World

Dispute Resolution Forum Suggestions & Feedback Introductions Lounge
Back
Top