Hacker News Reader

It doesn't "lie" (there's no intent to deceive). It makes "mistakes" [1]. It's a tool that is increasingly a part of the modern world whether you like it or not. Students who learn how to use the tool, which includes learning its limitations, will benefit overall.

[1] technically it doesn't even make "mistakes", it's just an LLM doing its thing, but we humans like to anthropomorphise

View on HN

when a child makes up things that it knows not to be real there is no intent to deceive either. the child lacks the capacity to understand the difference. yet we will tell the child that it is lying.

just like a child, chatgpt can not tell the difference between reality and fiction.

chatgpt lies because it presents fiction as fact. there is no intent to deceive necessary.

We do like our anthropomorphisation, and I prefer to say that ChatGPT and its pal likes to "bullshit". They're neither "trying" to tell the truth nor trying to lie - they trying to sound plausible, with a cheerful disregard of the value of truth.

I'm done anthopomorphising now.

Reply on news.ycombinator.com