AI makes even well-adjusted adults weird.
For instance, I use a coding model running locally on a DGX Spark machine. Like all LLMs, it can’t stop talking, and its verbosity drives me up the wall.
So I have this habit of always adding “Be terse” at the end of my queries, or if I know it’s going to apologize and butter me up when I catch it being wrong and I ask it to correct its previous answer, “Quit whining”.
One day one of my most AI-happy colleagues at work walked into my office and read the query I was typing to the model, and he said with a genuinely sad face: “Aaw… You’re treating your LLM like shit. Poor thing.” I looked at him funny for 10 seconds, and I realized he wasn’t even joking.
Turned out, he also thanks his models and praises them when they come up with the correct answer. Like… for real. The mind boggles…
he also thanks his models
I have seen people saying Hello and Goodbye to LLMs. It’s weird but also shows that they don’t know they are using a computer application. No need to say hello because the program is there, and as for goodbye, it’s useless too, just close the damn window.
Yeah but the weird thing is, this guy is a computer engineer. He’s firmly convinced that being nice to a LLM makes it perform better. Which goes to show that even people with a solid technical grounding aren’t immune to anthropomorphization a magical thinking.
Oh no, Scotty’s one of them!
We’re not gonna make it, are we? People, I mean.
