> Within less than a decade we will have systems that output "thoughts" and actions at least dozens of times faster than humans. We've already had this for dec

> Within less than a decade we will have systems that output "thoughts" and actions at least dozens of times faster than humans. We've already had this for decades. You're just describing computers. If we give these systems unrestricted access to infrastructure/resources, and something bad happens, that's not the system's fault. It's our fault. I am not a doomer, but based on the current state of AI, I can't say I'm very optimistic that we'll get this right. We actually do know how to solve this problem, but there is so much magical thinking and grift in this space that I don't think our prior experiences matter.