>>281811779
I don't want to get into an argument here, I think we can discuss AI stuff in the AI thread.
>"In my supreme intellect I have decided
Yeah but again this is an alignment thing. IDK if you're familiar with the writing on this subject, the goal of alignment is to make sure AI is not plotting our doom and is serving humanity. AI won't have emotions, it will probably have drives; it won't have ambition but it will have goals. We want it to serve humanity at large. That means doing good stuff and not helping terrorists make bombs. The AI companies are going to become very rich regardless, ideally they'll make the economy so large everyone will be living a life of plenty.
>What lets you assume we don't need to give a robot fleet 20-30 years to build up all the factories for MORE SPECIALIZED ROBOTS to optimize tasks?
Good question; so the authors of AI 2027 address this specifically, they point of that during WWII it took a few years of effort to turn factories that made cars into factories that made bombers. At full production they were producing something like one bomber per day. That was with 1940s technology and labor. An AI company could conceivably do the same thing, but probably faster; and AGI would be able to guide labor and resources more efficiently than the best human-staffed logistics companies. They envision the first factories being built by humans guided by AI, and then the AI can take over once enough robots have been produced.

ai-2027.com if you want to read about it, i found it a very fascinating document.