The magic of intelligence is that such a machine is in fact possible. Humans are proof of it -- evolution didn't program us to play chess or search for exoplanets or shitpost on Twitter.
Show this thread
The central question in AI is -- can we make a machine that's capable of doing things we didn't program it to do? It's reminiscent of alchemical transmutation -- something out of nothing. A system that's fundamentally more than what went into it.
-
-
-
Quick note for those who bring up cellular automata or other systems with unpredictable behavior: CA don't "do things" (implies a task). "Being more" than the parts implies being more useful and valuable (subjectively). Intelligence is emergent but emergence is not intelligence.
Show this thread -
Unpredictability / complexity alone (e.g. cellular automata) isn't useful or valuable. Intelligent systems do things (tasks) they weren't programmed to do. They produce solutions that weren't contained in their programming instructions.
Show this thread
End of conversation
New conversation -
-
-
Seems obvious to me that we can't. If we program them to learn, and they become better than us in some behaviour which is based on their learning, they are still doing what they were programmed to do.
-
Progress will continue in chaos; phenomenon from action before understanding or even intention. A feature, not a bug, may well allow for run away phenomenon that feels, to us, more like true creativity than anything we encountered. When it’s beyond us to understand, who knows?
End of conversation
New conversation -
-
-
Wasn't transmutation something into something?
-
Something into something qualitatively different (we thought at the time)
End of conversation
New conversation -
-
-
AI purpose driven, rather than instructions driven
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Is the Age of Ultron still a possibility then ?
pic.twitter.com/xYeER6nZzUThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Yes we can. It just needs access to a world simulation rich enough to find effective sub-objectives in service of the larger objectives programmed into it.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.