"Rozwiążemy ten problem bezpieczeństwa mitów dość szybko. A potem pojawia się pytanie, w czym sztuczna inteligencja jest lepsza? Czy jest lepsza w łamaniu kodu, który został poddany audytowi przez inne sztuczne inteligencje? Czy jest lepsza w inżynierii społecznej? Czy jest lepsza w udawaniu ciebie i dzwonieniu do twojego brokera w JP Morgan Chase i nakłanianiu go do zrobienia czegoś, czego nie powinien robić? Właściwie myślę, że sztuczna inteligencja będzie znacznie skuteczniejsza w hakowaniu ludzi niż w hakowaniu kodu w najbliższej przyszłości." "We're gonna solve this mythos security problem pretty quickly. And then the question becomes, what is the AI better at? Is it better at cracking code that's been audited by other AIs? Or is it better at social engineering? Is it better at pretending to be you and calling your broker at JP Morgan Chase and getting them to do something that they shouldn't do? I I actually think AI is going to be much more effective at hacking humans than it is hacking code in the near future."

… That's totally true. But defense and offense are always an asymmetric game and eventually defense will get much better. Uh, which is what we saw. You know, we all have uh read the stories about kids in the 80s hacking into the Pentagon with, you know, a whistle from a a serial box. And it's like, yeah, we we solved that problem pretty quickly. We're gonna solve this mythos security problem pretty quickly. And then the question becomes, what is the AI better at? Is it better at cracking code that's been audited by other AIs? Or is it better at social engineering? Is it better at pretending to be you and calling your broker at JP Morgan Chase and getting them to do something that they shouldn't do? I I actually think AI is going to be much more effective at hacking humans than it is hacking code in the near future. And so the self-custody DeFi forward model, you know, right now maybe it's a little rough after we're seeing all these hacks come out, but I would I would not bet against it long term. I think we're going to see most traditional systems built much more similar to that in the future than we do today because the humans are the biggest vulnerability …

---