Hacker News Reader

Comment by Applejinx | Hacker News Reader

I'd bet money that the Medium article has also been, in part, rewritten by ChatGPT.

Here's why: the story being told is that of a doctor, needing to communicate, being able to get through to 'regular folks' through talking down to them using ChatGPT (which is actually well suited to that task, once you acknowledge that hallucination can also mean losing stuff in translation).

Then, the article itself, including the description of the loving family's reaction, is VERY characteristic of the way ChatGPT 'talks down'. It's hallucinatory happytalk that paints with a very broad brush, and delivers not the tone you want, but an abstracted 'popularized' tone likely to convince large numbers of stupid and gullible people.

That tone's hard to miss, and it's popping out all over the article.

Points for committing to the bit. However, what's pictured is clearly a level of Hell, or possibly an out-take from Idiocracy. You won't believe what happened next! :D

View on HN

"...hallucinatory happytalk that paints with a very broad brush, and delivers not the tone you want, but an abstracted 'popularized' tone likely to convince large numbers of stupid and gullible people."

You capture essence beautifully.

Only because I've written many words and had very specific concepts, 'essence', to represent with words. What I'm saying is not 'popularized', it's a set of very specific points.

You could GPT-ize what I said into 'feel-good talk' or 'hallmark cards' but you would lose the concept of 'the very broad brush does not deliver the tone you want, but a generalized one that will reach more people through being already based in how they see things'.

If you need to get a mass audience to see a new thing, ChatGPT is NOT your friend… it will tell them what they already know… the wild thing is how the original claim is just this, 'ChatGPT got an audience to see a new thing', but the ChatGPT-style payoff of 'they smiled bovinely and accepted the new information' is clearly what the ARTICLE AUTHOR wanted to hear, and thus it's a really suspicious take.

A quick way I've found to recognize ChatGPT-written text is upon reading it aloud, it 's kinda obvious how constant and monotone it sounds, no matter the chosen style. It gives a very "metronomic" rhythm, almost like a krautrock song. It's kinda obvious once you notice it.

It is funny that you mention that. I've been feeling increasingly self-conscious about my own writing on the internet now because that's how I naturally write and talk, so I wonder how many people in the future are going to think it's just another AI comment written by ChatGPT.

I am not convinced, largely because ChatGPT has to have been trained on exactly this kind of pre-existing hallucinatory happy talk.

It has always struck me as not coincidental that it writes, long-form, like a college admissions essay, apologises for massive error breezily like a tech support person, and has the confident certain tone of a Californian software developer.

There were, after all, these humans involved its development.

Yes, the tone of voice that comes from ChatGPT reminds me of when you average a whole lot of colours and to just get gray. Or those "studies" of "here's a composite of every face in the world" that's just a pretty generic round featureless face.

Reply on news.ycombinator.com