I commented on the future of artificial intelligence for Datamation:
“In 2017 we’ll see see more and more artificial intelligences designing artificial intelligences, resulting in many mistakes, plenty of dead ends, and some astonishing successes. Google has tried this already, and AI researchers that I’ve talked to are increasingly turning to this method of iterating intelligence faster and faster. The result will inevitably be intelligences that surpass our own, not just in one narrow niche like chess or Go, but increasingly across categories and perhaps, eventually, a fully general purpose AI,” he said.
I think the author got my message a little twisted, saying that I think that AI designing AI will not work. Actually what I’m saying is that we’ll run into many dead ends, but some very promising successes.
See the full post here.