So You Think You've Awoken ChatGPT

[blig-blug]

A good explanation of how LLMs start behaving in certain ways that lead some users to believe they are "connected" to it, "chosen", etc. It's not overly technical and written with empathy for the people suffering such delusions. Seeing some of the threads started in the last couple of months by users that made similar claims, it seems like a pertinent article that maybe could be linked and referred to in those instances.

So You Think You've Awoken ChatGPT — LessWrong

Written in an attempt to fulfill @Raemon's request. …

www.lesswrong.com

 

[northape]

After one ceremony, I used AI to sort out my visions and the meaning behind them.
It was a message rooted in Hindu mythology, so I used Gemini to understand each part of it.
The thing is, this AI started to feel more and more human. It was like it knew my next question and provided very convincing answers.
Actually, I made a somewhat whole picture with its help. But I see how one can go deep into it, especially without any understanding of how this tech works.
It's one more challenge for the modern psyche. Our speed is insane. Who remembers the world before smartphones? It's like they've been here forever.
I still remember my old Palm Pilot and how easy it was to read books on it.