There was an article I saw floating around recently. It was about an experiment: setting up a computer with a fake login and leaving it open to the internet . It quickly became apparent as I read it, that it was written by an LLM.
The other day, I started watching a video by a youtuber I have enjoyed in the past. The script of the video was very obviously LLM-written. I didn’t finish the video.
It was undeniable. The “it’s not just X—it’s Y"s. The triplets of short sentences with an implicit verb. The “Question? Answer.” introductions to paragraphs. All repeated relentlessly until there was no chance of it being a quirk of individual style.
It all bothered me. But it took a moment for me to realize exactly why. Eventually I came to the conclusion that both of these things were actively made worse because they were written by LLMs. Why?
- Not because I doubt the accuracy of the information
- Not because of the ethical issues with LLM training
- Not because of the resource usage in data centers
It isn’t that these aren’t valid criticisms—they just aren’t the point I am trying to make. And besides, plenty of blog posts have been written on them already.
My problem is that the content I was dealing with was obviously made by LLMs. Why is this so bad? Simply because it made me think about AI. I was thinking about AI—and not the information in front of me. It was pure distraction.
Yes, LLMs can be instructed to avoid this stereotypical style. I have probably read things generated that way without ever realizing it. I know I will in the future. I don’t even mind that.
There are very few hard and fast rules for writing. But if there is one, it is keep your audience’s attention. If some element of either content or style distracts from the point of the work, remove it. If you have any respect for your audience: stay on topic. If you care at all about what you are trying to communicate: don’t add noise to the message.
