The AI created by big tech companies today is a fake AI that cannot even use the word “artificial intelligence.” LLMs can be an excellent way to mimic human language, but that is not intelligence.
If we define intelligence as that of an LLM, then the behavior of infants and young children is instinct itself, animals without sound language are not intelligent, and you end up making all actions based on statistical judgment as like LLM.
The technology currently called AI is merely a computer capable of ultra-fast parallel computation through extensive computer hardware.
Big tech wants to hide this fact and argues that regulations should be imposed to prevent it.
A real threat occurs when genuine artificial intelligence is developed. AI that generates current images and videos does so simply by copying and pasting existing information into the required elements.
True intelligence will create based on creativity, and like Einstein and Hawking, it will advance humanity’s productivity and creativity. But if such [intelligence] is used by just anyone without any restrictions, it is like anyone having a nuclear weapons design blueprint. This is why regulation is necessary. The goal should be a social consensus on the distinction and operation of the “usable people” and the “nonusable people” in [Real Intelligence].
Occam's Razor says anything LLM-related that talks about AGI is 99% hype, but I think it's plenty easy to create threats -- even existential ones -- without anything approaching AGI. It's already fucking up production databases and Iranian civilians due to "misalignment" with the intended goal of the user (well, the databases anyway), I see no reason why that kind of unchecked automation on a larger scale couldn't pose a serious risk to humanity without any actual "intelligence" involved.
That's a nice remark. You were referring to how the existing threat is the distorted propaganda of misguided big tech, isn’t it?