The AI dooooooom thread

[Savonarola]

AI will become part of the 3GPP Cellular specifications for 6G (release IMT 2030), (although some companies already use it to optimize/manage their networks.)  It's not yet established how this will work.  I was listening to a symposium last night about how to have an extremely low latency network able to access a LLM.  One of the members said that we couldn't have a local LLM because couldn't put a nuclear reactor at every node, and my immediate thought was "Or can we?"

 

He suggested using a digital twin (and barring the invention of Mr. Fusion from "Back to the Future", that's probably the most realistic way.)