Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing

> The bug gets fixed. Your mental model doesn’t move.

> The symptom vanishes. You ship.

> The tool didn’t determine the outcome. The posture did.

Here's a free prompt if you can't come up with one that avoids this awfulness yourself: https://news.ycombinator.com/item?id=48100213



Here's a Claude code skill for that https://github.com/blader/humanizer


Or use software that only lets the AI critique/give you feedback, but not write for you.


> Even as someone who uses a lot of AI, if you can't be bothered to at least give it a prompt like "Go through the documentation and comments in detail and remove any obvious AI shibboleths like emdashes, it's not x it's y, rule-of-three, 'delve', excessive grandiosity and flourishes, boldness, bullet points, etc", you should receive a brisk kick in the rear.

...is it that easy? I mean, wouldn't the AI providers just include this in their system prompts and/or harness?


It's better that they keep those verbal tics to make low effort AI generated output identifiable. Those that care will put in the effort to clean it up for their purposes.


From a societal perspective, I agree (and admittedly, maybe I shouldn’t be giving them ideas). But from a product perspective, this doesn’t seem optimal.


I get good results doing that, but it's all about being the human who has the taste and making a tasteful result. change the prompt, try again, change the temperature, whatever it takes.


Are those sentences untrue?


No, but the construction is painful and repetitive. My issue is the stodgy, repetitive, flavorless prose, not the content per se. The actual content could probably be compressed down into one or two paragraphs - "give me the prompt". Hemingway, not advertorial.




Consider applying for YC's Fall 2026 batch! Applications are open till July 27.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: