Did it occur to you that if superintelligent AI turns out not to do what humans want that it is more likely correct than us?
-
-
Replying to @Plinz
If we trust that its underlying value function is sound, sure. But many value functions are possible, and we can't rely on our ability to judge a value function a-priori with respect to the behavior it gives rise to in the agent. We may be forced to infer soundness from behavior.
1 reply 0 retweets 0 likes -
Replying to @vlthr
My point is that we cannot claim to be able to judge value functions better than superior minds.
2 replies 0 retweets 6 likes -
This Tweet is unavailable.
-
Replying to @jingle__belle @vlthr
I think that means that you have not yet understood your own values?
1 reply 0 retweets 1 like -
This Tweet is unavailable.
Replying to @jingle__belle @vlthr
In a way. Let's deconstruct the human part
9:54 AM - 7 Nov 2019
0 replies
0 retweets
1 like
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.