It really doesn't sound like ChatGPT’s default voice, though it is pretty good at taking on different voices so in a sense you could say that about almost anything. It does use em-dashes, which people have recently started way over-indexing on as a ChatGPT tell, but lots of posters on HN have been using em-dashes for longer than ChatGPT has existed.
It reads like pretty much every piece of tech marketing/evangelism in the last several decades. Which, sure, ChatGPT nails pretty well if you tell it to do that, but... I don't think that has high specificity as a ChatGPT tell.
And yet, over the last year or two, people using ChatGPT to write their comments stand out like a sore thumb. The overall structure, the specific style of using em-dashes, semicolons, and colons...it's blindingly obvious.
Yep, I read through some of their comments -- it is strange. I would certainly like to see people improve their grammar, punctuation, and general consistency; but, let's face it, people rarely care to.
Call me paranoid (because, let's admit it, I am) but... after all, it's the Internet, and it's 2025! There's been enough controversy about the political power of speech over the past decade alone, that I can see people running their stuff through ChatGPT just to stay on the safe side and make things sound blandly "professional": just so they can avoid being taken the wrong way by a random reader who happens to strongly object to some particular aspect of their communication style.
(Goodness knows I've found myself on either side of all that at different times -- personally, I find it highly inauthentic to make noncommittal "positive" statements in lieu of plain observations. It's absolutely grating; while some other people seem to require it, and can be indeed quite self-contradictorily harsh about it.)
I can definitely see a major use case for LLMs there -- though I do find the implications quite terrifying. Call it political correctness, call it jamming stylometry, call it a day. Either way there's definitely some sort of power differential here that needs to be examined and I think the world is less prepared than ever to confront whatever its meaning turns out to be.
Which brings me to my other point:
>To be clear, I don't really care, I use ChatGPT all day every day, but just letting OP know it's often pretty obvious when you have it write for you.
Now this I don't quite understand. Pointing something out ("letting someone know") generally implies you want someone else to care about that something, even if you honestly don't. So, since you don't care -- why is it that you want others to? Honest question.
It really doesn't sound like ChatGPT’s default voice, though it is pretty good at taking on different voices so in a sense you could say that about almost anything. It does use em-dashes, which people have recently started way over-indexing on as a ChatGPT tell, but lots of posters on HN have been using em-dashes for longer than ChatGPT has existed.
It does read like marketing material, though.
It's not just the em-dashes for me. It's actually more these parts:
But here’s the point:
That’s the win:
Those sound exactly like ChatGPT when I tell it to write in a more direct, opinionated style.
It reads like pretty much every piece of tech marketing/evangelism in the last several decades. Which, sure, ChatGPT nails pretty well if you tell it to do that, but... I don't think that has high specificity as a ChatGPT tell.
Generic marketing speak is generic.
2 replies →
It does, and it muddies the waters a lot. Why does it read like a sales pitch?
ChatGPT learned that voice from actual people, you know.
And yet, over the last year or two, people using ChatGPT to write their comments stand out like a sore thumb. The overall structure, the specific style of using em-dashes, semicolons, and colons...it's blindingly obvious.
If you just go back a couple months and read OP's comments, they sound very different from everything they've posted today: https://news.ycombinator.com/item?id=42734300
To be clear, I don't really care, I use ChatGPT all day every day, but just letting OP know it's often pretty obvious when you have it write for you.
Yep, I read through some of their comments -- it is strange. I would certainly like to see people improve their grammar, punctuation, and general consistency; but, let's face it, people rarely care to.
Call me paranoid (because, let's admit it, I am) but... after all, it's the Internet, and it's 2025! There's been enough controversy about the political power of speech over the past decade alone, that I can see people running their stuff through ChatGPT just to stay on the safe side and make things sound blandly "professional": just so they can avoid being taken the wrong way by a random reader who happens to strongly object to some particular aspect of their communication style.
(Goodness knows I've found myself on either side of all that at different times -- personally, I find it highly inauthentic to make noncommittal "positive" statements in lieu of plain observations. It's absolutely grating; while some other people seem to require it, and can be indeed quite self-contradictorily harsh about it.)
I can definitely see a major use case for LLMs there -- though I do find the implications quite terrifying. Call it political correctness, call it jamming stylometry, call it a day. Either way there's definitely some sort of power differential here that needs to be examined and I think the world is less prepared than ever to confront whatever its meaning turns out to be.
Which brings me to my other point:
>To be clear, I don't really care, I use ChatGPT all day every day, but just letting OP know it's often pretty obvious when you have it write for you.
Now this I don't quite understand. Pointing something out ("letting someone know") generally implies you want someone else to care about that something, even if you honestly don't. So, since you don't care -- why is it that you want others to? Honest question.
2 replies →