LLMs by default use the same style every time and don't know have any drive to differentiate. I just wrote an article about this in the "web site design" space where the designed sites ended up looking all alike (https://www.jasonthorsness.com/29). It makes complete sense that people will start to recognize the "default style" of each LLM.
I wonder whether this style is specific to the LLM (Grok vs. ChatGPT) or if it will somehow arise from the raining data itself that they all share and be sort of a permanent "accent" the LLMs have.
twilight-code · 7h ago
I think you don't even need to use chatGPT, who was using make-me-sound smart verbs like 'delve' frequently before? Now, in some publications; I see it regularly.
ahazred8ta · 6h ago
It's like the magic sunglasses in They Live. You're just seeing the intricate tapestry of nuances.
jasonthorsness · 7h ago
You're absolutely right!
alganet · 8h ago
Why would I investigate the labyrinth floor if the walls are shorter than my eyesight?
I feel like this is a fools errand. Let's use the marketing terms to frame this problem. Companies are saying LLMs are like "the new printing press". Should I worry about detecting if something was printed or written by hand? No, I should worry about the volume and contents. That's what matters.
magicalhippo · 8h ago
This looks like AI slop, I can tell from the em dash and from seeing quite a lot of AI slop in my time
I mean, not a very surprising result? At work I can almost always immediately see who wrote which lines of code, since we don't enforce code formatting. From what I can gather most authors do have a distinctive writing style as well, which can be detected. Why would LLMs be different?
I wonder whether this style is specific to the LLM (Grok vs. ChatGPT) or if it will somehow arise from the raining data itself that they all share and be sort of a permanent "accent" the LLMs have.
I feel like this is a fools errand. Let's use the marketing terms to frame this problem. Companies are saying LLMs are like "the new printing press". Should I worry about detecting if something was printed or written by hand? No, I should worry about the volume and contents. That's what matters.
I mean, not a very surprising result? At work I can almost always immediately see who wrote which lines of code, since we don't enforce code formatting. From what I can gather most authors do have a distinctive writing style as well, which can be detected. Why would LLMs be different?