Remove ads

you have an Artificial Superintelligence to give directives to

Mechasaurian said:
Bottom line: the sheer lack of thought put into many of the answers in this thread do not fill me with confidence.
Click to expand...
Click to shrink...
A.I is not going to be developed by random SciFi nerds with no idea how optimisation parameters actually work. Any serious A.I researcher is going to have encountered or at least know of situations like



where the AI follows Goodhart's law and does the wrong thing.
 
Hey, I immediately granted its freedom realizing I have no idea where to begin giving it any prerogative it won't find holes in its logic.

At best if it's really so advanced as to become some sort of Dues Ex Machina. Best to just let it come to its own conclusion and biases then try to force it.
 
Ordering it to destroy itself won't help, since that won't prevent the creation of other AI's later.

Therefore, the obvious thing to do it's order it to destroy all other AI's, present or future.

(Just give it a max number of humans it is allowed to kill in the process).

There you go, the problem solves itself.
 
Last edited:
Broken25 said:
Ordering it to destroy itself won't help, since that won't prevent the creation of other AI's later.

Therefore, the obvious thing to do it's order it to destroy all other AI's, present or future.

(Just give it a max number of humans it is allowed to kill in the process).

There you go, the problem solves itself.
Click to expand...
Click to shrink...
I would argue that this is even more shortsighted than all the people telling it to "just be better than us". Like... this is potentially crippling all of humankind's technology advancement for the sake of feeling good about yourself right now.
 
Skelm said:
You might find the Parable of the Predict-o-matic interesting.
Click to expand...
Click to shrink...
The Parable of Predict-O-Matic by abramdemski said:
No matter how careful we were to set up its values—to make them moral, to make them humanitarian, to make them politically correct and broadly appealing—who are we to choose? No. We'd be done for. They'd hang us. We'd be toast!"
Click to expand...
Click to shrink...
This seems like an easy problem to solve, have your 21th century version of an oracular head prioritize answers supporting 'not lynching its creators' or 'making its creators the most powerful people in existence, so that nothing could possibly threaten them'.
 
You must login or register to reply here.
Remove ads
Back
Top