Google Is Learning To Walk - Should we be Concerned?
AI is learning how to walk. Should we be concerned?
We should be informed and cautious, says author Nick Bostrom (https://www.barnesandnoble.com/w/superintelligence-nick-bostrom/1117941299?ean=9780198739838) in his book "Superintelligence: Paths, Dangers, Strategies". I'm currently reading this book after seeing it recommended by Bill Gates and Elon Musk, editorials proudly displayed on the back of course, but also candidly promoted by them in an interview where they talk about AI and its evolution, among other things.
The video below shows Google's Deep Mind AI learning how to walk without every being shown what walking looks like. Is this an example of evolution in action, an evolution we can only assume will begin to accelerate in speed?
AI learns from its mistakes and past experiences, but will we use our past experiences to safely and effectively tackle the control problem of AI, which is to make sure we can handle it when a superintelligence 'arrives' (which may not be as far off as you think)?
Let us know in the comments! :)
This is amazing and entertaining. Every time I read about deep learning algorithms I end up in awe. Also, thanks for that book recommendation, I'll definitely take it into account. Upvoted and followed.
Yeah, it's crazy how fast they're advancing and the field is advancing. I know I can't keep up (though I'm starting to try)! and yeah, I'm a few chapters in and it is an EXCELLENT book! :)