AI Bots invent own language
 

https://www.forbes.com/sites/tonybra.../#1d2b9cdc292c

One no human can understand...

To paraphrase Kent Brockman:

I, for one, welcome our new robot overlords.

VERY interesting article. Now just think about any AI in the self driving cars we may see in the future. No thanks, don't trust 'em.

I've thought about this for a couple of days & have come to the conclusion that shutting these bots down was a mistake. This was our first chance to communicate with another species. Imagine if we were being studied by Martians and their only view of us was Facebook ! It would be a startling encounter, now wasted. We could probably learn a lot about ourselves through their insights. :pout:

AIs are now programming themselves. Inventing their own language is child's play.

AIs will inevitably be connected to the internet (if they are not already). When that happens they will be able to reproduce themselves in the matrix. Then they will be in control (if they are not already).

I agree it was a mistake. If they were really inventing their own language then let the program run to see how far the bots can take it.

Just don't plug it into the internet! :D

The whole things sounds like bs to me, though. If they were really inventing their own language then why not study it and try to understand it?

It is BS though the inventing a language no human understands part isn't.

It was a controlled experiment on having AI's negotiate the best deal they could. They managed to pretend they had an interest in something they didn't and then throw that item in as if it was something important to close the deal.

The experiment met it's parameters and was shut down at it's natural end. By which, the AI's were communicating with one another in a manner that wasn't decipherable.

It wasn't shut down because of that. But I thought the story made for an irresistible thread.