This got under my skin. I do write with AI, but not to automate or outsource. I use it to sharpen, to reflect, to get underneath the obvious phrasing and into the real thought. I get that for some people, ANY AI involvement feels like cheating. But that framing misses the nuance of how it’s used—and what kind of relationship you’re building with it.
The problem isn’t just synthetic content - it’s the synthetic TONE. Flattened voice, templated cadence, writing that feels algorithmic even when it’s not. I’ll admit, the “giveaways” people have started calling out (em-dashes, metaphors, a certain rhythm) have made me second-guess my own writing style (and change it lest people don’t think my thoughts are my own.)
Using AI to research, think through ideas, and compile them so you can distill things down and write in your own voice is one thing. I've found AI tools to be invaluable for bouncing ideas around. And, as you mentioned in your article, long-term memory really helps for building a GPT that is uniquely fitted to the ways you think and process information.
The loss of tone sucks. It feels like we're going from all the colors of the rainbow to monochrome. The vibe is off.
Prompt whispering, or consciously building a system that is a true human + AI partnership, takes active effort. Just like with social media, I think people need to take a step back and more critically analyze how they're using the systems and tools at their disposal daily.
100%. I really appreciate your openness here, especially because it is so easy to default to binaries in these conversations (real vs. fake, human vs. AI, etc). I feel the tone collapse too, not just in what’s being written, but in how we’re all learning to read, I guess?
"Prompt whispering" has been one way I try to stay close to the voice beneath the voice. But yeah… it takes active effort, like you said. And it's something a lot of people don't do (or don't know to.)
I'm working on a new post that sort of takes some of the ideas I wrote about in my non-AI related post about the rise of absolutism (https://open.substack.com/pub/unexpectedinsights/p/truth-in-tension-why-complexity-matters?r=h5zs0&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false) combined with some of what I touched on in my Prompt Whispering post - exploring the tension of how the convergence of AI's design for user satisfaction and society's drift towards absolutism creates a feedback loop, and how that might mean that ChatGPT and other LLMs may inadvertently wind up validating superficial understandings and even discourage critical thinking and self-reflection. And what would that mean?
Wow, this type of content is impressive and helpful. Keep it up 💪 !
I can’t believe!!!
Is this a 🤖?
Me , robot 🤖 ?
This got under my skin. I do write with AI, but not to automate or outsource. I use it to sharpen, to reflect, to get underneath the obvious phrasing and into the real thought. I get that for some people, ANY AI involvement feels like cheating. But that framing misses the nuance of how it’s used—and what kind of relationship you’re building with it.
The problem isn’t just synthetic content - it’s the synthetic TONE. Flattened voice, templated cadence, writing that feels algorithmic even when it’s not. I’ll admit, the “giveaways” people have started calling out (em-dashes, metaphors, a certain rhythm) have made me second-guess my own writing style (and change it lest people don’t think my thoughts are my own.)
That said, I’ve been trying to find a different approach that’s more relational and more emotionally attuned. Shameless self-promo, but I wrote about it here: https://unexpectedinsights.substack.com/p/chatgpt-prompt-whispering-for-the.
Still figuring it out, but this feels like the right tension to hold?
I agree, Elisabeth. There's a nuance here.
Using AI to research, think through ideas, and compile them so you can distill things down and write in your own voice is one thing. I've found AI tools to be invaluable for bouncing ideas around. And, as you mentioned in your article, long-term memory really helps for building a GPT that is uniquely fitted to the ways you think and process information.
The loss of tone sucks. It feels like we're going from all the colors of the rainbow to monochrome. The vibe is off.
Prompt whispering, or consciously building a system that is a true human + AI partnership, takes active effort. Just like with social media, I think people need to take a step back and more critically analyze how they're using the systems and tools at their disposal daily.
100%. I really appreciate your openness here, especially because it is so easy to default to binaries in these conversations (real vs. fake, human vs. AI, etc). I feel the tone collapse too, not just in what’s being written, but in how we’re all learning to read, I guess?
"Prompt whispering" has been one way I try to stay close to the voice beneath the voice. But yeah… it takes active effort, like you said. And it's something a lot of people don't do (or don't know to.)
I'm working on a new post that sort of takes some of the ideas I wrote about in my non-AI related post about the rise of absolutism (https://open.substack.com/pub/unexpectedinsights/p/truth-in-tension-why-complexity-matters?r=h5zs0&utm_campaign=post&utm_medium=web&showWelcomeOnShare=false) combined with some of what I touched on in my Prompt Whispering post - exploring the tension of how the convergence of AI's design for user satisfaction and society's drift towards absolutism creates a feedback loop, and how that might mean that ChatGPT and other LLMs may inadvertently wind up validating superficial understandings and even discourage critical thinking and self-reflection. And what would that mean?