He's probably right considering the current direction that A.I. is moving. If A.I. becomes nothing more than ANNs that continually upgrade themselves for the best efficiency then A.I. based on that principle will indeed view humans as totally inefficient and irrelevant.
So I'm totally onboard with this view. The current direction that A.I. is being pursued is not good. The real questions is whether anyone will heed this advice?
Note also that he didn't say that A.I. in general is a danger, but rather the way we are currently approaching it is the danger.
I totally agree.
DroneBot Workshop Robotics Engineer
James
How the world, I mean Hollywood of course, envisages future AI has been well recorded, not only by the "Lost in Space" original TV series, that only a few of us can remember, but all the way back to Fritz Lang's Metropolis.
Not everything that is possible is good, and not everything that is good is possible!
Who said that? Well, ME of course!
Note also that he didn't say that A.I. in general is a danger, but rather the way we are currently approaching it is the danger.
He didn't but the topic title was just an attention grabber, and it seems to have worked. ?
He didn't but the topic title was just an attention grabber, and it seems to have worked.
Oh no! Taken in by an attention seeking topic! Me bad.
I actually agree with Dr Stuart Russell in the video you posted a link to. A.I. as it is currently being pursed via nothing more than ANNs, is bound to end in disaster.
A.I. development based on ANNs is moving at a pace that most people are totally unaware of.
If you think this is something that is far off in the future, think again. Once neural networks are designed to design themselves (which apparently has already been done), they will quickly design more complex neural networks that humans can't even understand.
So we're basically there already.
DroneBot Workshop Robotics Engineer
James