Marcus's thesis seems to be "there exist things which current AI can't do", which is likely to remain true... forever? I imagine when ASI is achieved, he'll be arguing it still hasn't created world peace, hasn't found a grand unified theory, and is bound by physical laws.
median household income figures seem somewhat incorrect. Everything else seems fine.
idk, it's not amazing, but I guess a lot of human beings below 85 IQ wouldn't even get the basic structure of the table correct, and that's like. 40% of the global population of human beings on earth looking at national IQ statistics?
Similar energy from Tyler Cowen: https://marginalrevolution.com/marginalrevolution/2025/02/o1-pro.html
He responded to this post here: https://x.com/GaryMarcus/status/1888604860523946354
Marcus's thesis seems to be "there exist things which current AI can't do", which is likely to remain true... forever? I imagine when ASI is achieved, he'll be arguing it still hasn't created world peace, hasn't found a grand unified theory, and is bound by physical laws.
Thanks for this post + related tweet. Glad someone is keeping track of this kind of goalpost-moving, strawman-like argument
How correct are those tables roughly speaking? I've been trying to get a sense of how much these new models hallucinate still
I tried spot checking some of the numbers and they seemed correct. I haven't checked each one. Let me know if you find any mistakes.
median household income figures seem somewhat incorrect. Everything else seems fine.
idk, it's not amazing, but I guess a lot of human beings below 85 IQ wouldn't even get the basic structure of the table correct, and that's like. 40% of the global population of human beings on earth looking at national IQ statistics?
So yeah.