For any task in ARC, it's possible for a human to write a reasonably short computer program that can handle new pairs. Everything is straightforwardly computable -- unlike, say, classifying MNIST digits. There's no human advantage in any single task.https://github.com/fchollet/ARC
-
-
The single biggest misconception in AI is the confusion b/t unintelligent output artifacts (skill programs) & the process that generates them. "Artificial intelligence" has been focusing purely on task-specific automation, entirely disregarding intelligence. Time to change that.
Show this threadThanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Doesn't this argument apply to the 5-year-old also? It has no intelligence that wasn't placed there by evolution? Let's rather say that dealing with novelty is a central challenge for AI that is beyond current capabilities
-
Yes... And then to define 'novelty'... :) To be honest, all this seems just words and claims around fuzzy intuitions about what 'intelligence' is supposed to be. Sure let's investigate programs that can solve various tasks, but that hardly seems like a huge insight...?
- Show replies
New conversation -
-
-
Can very young children (probably younger than 5) solve ARC tasks? If not, does that imply the need for additional knowledge priors? (which in humans might be learned or develop later innately)
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.