Comment by HarHarVeryFunny
8 hours ago
> minting zero days, running cyber attacks and defenses at machine speeds, discovering new science, advancing the frontier of mathematics
Maybe more people than Karpathy realize see exactly the above, which is what he and the privileged 5000 are seeing, except unfortunately the "discovering new science" part of it seems to be wishful thinking.
So, Anthropic and OpenaAI have built a "DeepBlue for math" that is also super-human at hacking, as well as being useful for coding.
I'm not sure what Karpathy's point is? Are we meant to be more impressed? More worried? More eager to buy it?
He's telling that some guys are about to create a theory of everything and open Pandora's box, while some of us are still grunt workers delivering 2ndary value to advance society
It seems like Karpathy thinks people are misunderstanding the capacities of LLMs when actually different people are interacting with LLMs in different contexts and so seeing different capacities.
I mean, most of the people in world wouldn't be impressed by chatbot solving their math problems since calculators can solve the majority of the average person's math problems already (and spreadsheet would solve the rest).