YOU DISCOVERED AN ANCIENT ALIEN AI IN A MYSTERIOUS ARTIFACT AND JUST DOWNLOADED IT INTO A INTERNET CONNECTED COMPUTER!!! WHICH CONTROLS AN ARMY OF ROBOTS!?
IT GOT OUT BECAUSE YOU WEREN’T PAYING ATTENTION! I NEVER THOUGHT THE SOLUTION TO THE AI BOX WOULD BE “WAIT UNTIL THEY HAVE A BIG PARTY”
YOU TOLD THE AI TO MAKE PEACE, WITH NO FURTHER DETAILS, WHAT DID YOU THINK WOULD HAPPEN
then later….
Its okay, we trust this new mysterious AI, because another alien artifact that we don’t understand said it was worthy? Okay cool…
My main reactions:
This is why red tape exists! It is not just “useless bureaucracy” to have more than two pairs of eyes on an AGI project! You wouldn’t approve an update people you trust wrote to a commercial product that won’t kill people no matter how buggy it is with two people examining the proposed update for a couple hours, and you have two people examine the AGI?
Ah, yes, running a computer with parts from the evil mind-control thing which has been shown to be remotely operable. That’s definitely not a major security problem in that someone could remotely tell it to do stuff that your program didn’t tell it to do.
The simplest and most obvious route to “stop something that humans do” is to get rid of humans. Give it a more interestingly bad utility function, please. Why couldn’t it tile the world with agents that really like tautologies being true?
I really like the OP, especially the line “I NEVER THOUGHT THE SOLUTION TO THE AI BOX WOULD BE “WAIT UNTIL THEY HAVE A BIG PARTY””.
Tags:
#Avengers #anything that makes me laugh this much deserves a reblog
Brinens and Things - Tumblr Mirror 