I've noticed that asking for a specific length sometimes makes the response much tighter. Without a limit, it tends to keep adding explanations and examples until the original point gets buried.
I'm curious whether others deliberately use word or sentence limits when working with AI.
I use mostly for writing fiction, and there are certain patterns that keep showing up.
Things like characters touching foreheads, everyone wearing leather jackets, dramatic coffee cups falling, and way too many characters saying “That’s dangerous.” What other AI writing clichés have you noticed?
Between SEO recommendations, engagement optimization, readability scores, and AI-generated suggestions, there are now endless ways to make a piece perform better according to some metric.
But a piece can score well on all of those things and still be boring.
I wonder whether we're gradually optimizing writing for systems that measure writing instead of for the people who actually read it.
A first draft is supposed to be rough. It's where you explore ideas, make mistakes, and discover what you're actually trying to say.
AI can immediately produce something polished-looking, which can make it tempting to accept the first version. But a clean paragraph isn't necessarily an interesting one.
Sometimes messy human writing has more potential than a perfectly structured answer.
AI can improve grammar, structure, clarity, and flow. But after enough rewrites, you can end up with something technically better that doesn't really sound like you anymore.
That's the strange tradeoff. You want help improving the writing without replacing the person behind it.
Do you ever reject an AI suggestion simply because it sounds too polished or “not like you”?
My approach changes quite a bit depending on what I am writing.
For something casual, I mostly use AI to clean up confusing parts. For something more structured, I might use it to challenge the argument or suggest missing points.
I used to have one general prompt that I used for almost everything, but that seems less useful now.
Does your AI workflow change depending on whether you are writing an article, post, story, email, or something else?
Having fewer restrictions is useful for certain kinds of creative work, but I would still choose a model with better memory, dialogue, and character consistency over one that is simply more unrestricted.
For people who use uncensored AI writing, what matters most to you once freedom is already covered?
I started doing this recently with pieces I considered finished
Sometimes the feedback is useless, but every now and then it catches a weak transition or an assumption I completely overlooked
The interesting part is that I do not necessarily accept the suggested rewrite
I just use the criticism to decide whether the original needs another look
Anyone else use AI more as a final reviewer than a writer?
A calculator doesn't automatically make someone bad at math, and a spellchecker doesn't mean someone can't spell. I see AI writing tools in a similar way: they can help with ideas, structure, editing, rewriting, and getting past writer's block. The real difference is whether you're using the tool to support your thinking or simply replacing it entirely.
Personally, I think AI becomes most useful when you still have your own ideas but need help expressing them better.
Sometimes the response is accurate, clear, and well structured, but it still does not sound like something I would actually say.
I have started caring more about whether the writing feels natural than whether every sentence is perfectly polished.
Do you usually keep a good AI response as it is, or rewrite it until it sounds like your own voice?
started adding this to some of my prompts because I noticed that AI often tries too hard to validate whatever idea I give it.
Getting actual criticism makes the conversation much more useful, especially when I am working on an argument.
Do you deliberately ask AI to challenge your ideas?
I can definitely produce a draft faster now, but I also spend more time comparing versions, checking details and deciding which suggestions are actually worth keeping
Sometimes I wonder if I am saving time or just moving the work to a different part of the process
How has AI changed the amount of time you spend on writing?
Sometimes a model seems to pick up on the way I phrase things after working together for a while. Other times I start a new chat and it feels like I have to explain everything again.
I am curious how much people actually rely on conversation history for maintaining a consistent writing style.
One thing I enjoy about AI is being able to test styles I would probably never attempt manually. Sometimes the result is terrible, but occasionally it makes me realize that I have been writing in the same way for years.
Has experimenting with AI ever changed the way you normally write?
I used to build huge prompts with instructions for tone, structure, vocabulary, examples, and things to avoid. Sometimes the output was worse than when I simply explained the idea and let the model figure out the rest.
Now I try to give it the goal first and only add instructions when something actually goes wrong.
Have you experienced the same thing, or do detailed prompts consistently work better for you?
As AI makes producing content easier, originality, taste, experience, and a unique point of view may become even more important.
What do you think humans can bring to writing that AI still cannot replace?
No brainstorming, no rewriting, no grammar suggestions, and no asking AI whether an idea is good.
Would your writing get worse, improve, or simply become slower? What do you think you would notice first?
I used to ask AI to generate things for me constantly. Now I often get more value from asking it what is wrong with something I already wrote.
Finding weak arguments, missing context, repetition, and unclear sections has become one of my favorite uses.
Has your AI workflow changed in a similar way?