The “AI Threat” CEO warnings are sure getting frothy (https://www.theregister.com/2023/05/24/openai_superintelligence_global_agency/)! I really can't tell if they're so high on their own EA fumes that this is for real, or if it's just a brilliant (?) ploy to avoid near term product liability by muddying the dialogue.
Google still can’t tell black people from gorillas, but the real threat is AI superintelligence? 1/
I think the advances of the past few years are stunning, but we're entering Penrose-like microtubule nonsense here. Weird==weird, advanced==advanced, logical reasoning not required. (See also today: https://qwantz.com/index.php?comic=4051) 2/
@jered You're giving them too much credit. Everyone and his dog is now writing an LLM, but if there is tight regulation then the barriers to entry will strangle most of them at birth.
A company that, say, got in first, had Microsoft's resources behind it and was in a position to be taken seriously by the people making the regulations, would be likely to be able to cope with new regulations just fine. The competition, not so much.
