Comment by aftbit
14 days ago
We better hope that we haven't awoken ChatGPT, because we haven't even attempted to keep it in a box. We're all rushing to build as many interconnections as possible and give AI unilateral control of as many processes as we can.
I think you’re getting ChatGPT confused with Skynet. I’m still confused as to why so many people think these things are going to kill us all.
I’m confused why you don’t.
For the same reason you don't think a typewriter is going to kill us all. That's just not how it works.
1 reply →
The idea of a malicious AI, like in a sci-fi movie, seems entirely divorced from what LLMs actually are. Look at the HN front page - these things struggle to perform basic accounting tasks, and you think it’s secretly plotting to commit genocide?
1 reply →
AGI vs LLM
Because survival is a universal trait and humans aren't necessary to keep an AI "alive". There is no reason an AI wouldn't treat humans the same way we treat everything living below us. Our needs are priority and it would be wise to assume the AI will also give itself priority.
Why do you think an AI would treat us any way at all? They don’t have feelings or emotions. They don’t have a persistent sense of self. They are incapable of engaging with us as sentient beings.
1 reply →
> treat humans the same way we treat everything living below us.
"below us"? speak for yourself, because that's supremacist's reasoning.