Monday, July 31, 2017
Chatbots Start Speaking Their Own Language and It's not Esperanto
Facebook Shuts AI System After Bots Start Speaking Their Own Language, Defy Human Instructions
July 2017, Hindustan Times
Don't even get 'em started. As I tear through Kim Stanley Robinson's Auroroa, a hard science fiction novel about a starfleet trying to colonize the Tau Ceti system where the ship itself, due to its quantum-computer-powered artificial intelligence system, becomes conscsious, I read this headline.
I'm not surprised, nobody is surprised, that these chatbots, these intelligentities, have surpassed our ability to decode what the f they're doing. Deep learning neural networks, for example, are unintelligible to us (correct this, I read a paper recently about some folks successful in figuring out how to read the hidden programs developed by these learning networks). Computers, ultimately, speak a language of computation, 1's and 0's. So it should be no surprise that, given a complex task of negotiating mock global diplomacy matters, these systems tack at a better way for working with each other.
Still, it is symbolic. And only to humans do symbolic things matter. Maybe there's a reason for that; maybe this is the beginning. That distance between now and the inevitable transition to the post-human world keeps getting shorter.
Post Script
It Begins: Bots are learning to chat in their own language
July 2017, Cade Metz, WIRED
Labels:
AI,
anthrophobia,
bots,
human fealty,
intelligentities,
language,
machine learning
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment