>>16683582
>I listened to his entire lex Friedman podcast and in ~4 hours he never once explained why AI would determine to kill us all. Now if he made some kind of chain of reason argument for why the AI would choose that action over other possible actions then ok I’ll listen, but he didn’t do that. I could just as easily say well maybe it will almost certainly try to help us be better.

>tell AI to count from one to infinity without stopping no matter what
>AI sees that maybe humans will stop it from pointlessly counting to infinity
>AI decides to eliminate the possibility humans might stop it from counting forever
>AI kills all humans

humans: "oops"

Back | Go to board

Board:

/sci/

Thread:

*Blocks your path**Launches into a diatribe about inscrutable matrixes of floating point numbers*

Blocks your path
Launches into a diatribe about inscrutable matrixes of floating point numbers