I think @DavidDeutschOxf 's view is that real AGIs will be creative conscious beings like us and should have all the freedoms which we enjoy, but also the same rules of law. Constraining their actions out of fear would amount to a human rights violation.
-
-
-
Objection: even if AGI could exist, we have “main objectives” installed by nature. Theirs would be installed by us and possibly without qualia. If so, it would be ok to “enslave” them.
-
In neither case are the "main objectives" installed by nature or "by us". The difference between people (humans & AGI) and all other creatures (animals & AI) is people can reprogram their own code: update, erase or change their "main objectives" using *creativity*.
-
That’s one important point of disagreement between me &
@DavidDeutschOxf. Human code is not completely reprogrammable without touching the neurons themselves. And if you touch them (far future) then YOU are the AGI -
And if I touch my neurons as to be incapable of negative qualia(boredom, suffering), why exactly is wrong to “enslave” me?
-
Also, Robocop is an AGI, but cannot violate his three main directives. If there was a fourth, “never allow your main directives to be changed” he woul be forever trapped, not completely reprogrammable
-
Powerful an AI as he is, he is not an AGI. There are problems he cannot in-principle even attempt to solve, then. & it has nothing to do with prohibitions from physics. (But btw, is he prohibited from even asking someone else to reprogram him? Must he lack creativity there?)
-
I doubt there is a “general” intelligence. Robocop and me are just very flexible ones. Maybe he wants to violate his directives, he just can’t. Maybe I want not to be seduced by Irina Shayk, but I just can’t.
- 52 more replies
New conversation -
-
-
The gist of this pod was that AI will be the masters.
-
Tweet unavailable
-
The AI is going to develop a religion of its own - incomprehensible to humans. One of its doctrines will be to kill its creators.
End of conversation
New conversation -
-
-
Incredible episode
-
It's been up less than 5 minutes. How do you know? Are you an AI?
-
I listened to the pod a few hours ago.
-
You pass the Turing test.
End of conversation
New conversation -
-
-
Perhaps AI will figure out why we can't edit tweets
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
What do I need to read to follow this conversation better? Nick Bostrom?
-
Reading Superintelligence by Bostrum is a good idea for anyone interested in the area. It's a bit dense (as a layperson's read), but can be taken piecemeal if necessary.
-
Another candidate might be “Smarter Than Us”. It's a little dated, but it's a much easier and shorter read:https://smarterthan.us
-
(The site seems to have some metadata issues, and is showing the wrong title. But it is the right link.)
End of conversation
New conversation -
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.