[DemonRaven]

But Facebook’s chatbots might get very smart very soon. On Wednesday, Facebook unveiled DeepText, a natural language-learning network engine to help its chatbots and other software better understand and mimic human language by scanning “several thousand posts per second” by Facebook’s global users.

http://www.newsweek.com/finally-there-facebook-chatbot-actually-works-466387                                                                          What the story is describing is a virtual assistant not a true chatbot. Chatbots by their nature and the nature of written language are hard to program. Many well known experts have been trying various types of natural language engines to make them more intelligent in their conversation. Language evolves and written language is especially problematic as many use abbreviation and misspell words. So all I have to say to facebook is good luck. I am going to be on the side lines laughing when it falls on its face just like microsoft's chatbot did.

[Freddy]

I can see the privacy outrages already.

[keghn]

   Build a Chatbot - ML for Hackers #6:
https://www.youtube.com/watch?v=5_SAroSvC0E&feature=youtu.be

[Art]

Aside from first appearing like a goof, he turned out to deliver quite a good video. Good find!

[8pla.net]

   Build a Chatbot - ML for Hackers #6:
https://www.youtube.com/watch?v=5_SAroSvC0E&feature=youtu.be

While that made me laugh , nevertheless it was informative.

[8pla.net]

So all I have to say to facebook is good luck. I am going to be on the side lines laughing when it falls on its face just like microsoft's chatbot did.

It seems, we all share the same A.I. Dreams.

[DemonRaven]

If they do more then just design a mimic program and go in and fix the errors it will be okay. But just making a mimic program has been tried in the past with no success. There is no quick method of creating them. You can not just let it learn on its own.   Just like with a  neural network it has to be trained by humans.At the very least they have to go in and correct mistakes. If they don't then they get some really raunchy racist and sexist chatbots . Had microsoft bothered to ask us "amateurs" and hobbyists we could  have warned them.

[8pla.net]

When it comes to A.I.  we're all subject to the same set of limitations.

[ivan.moony]

If you put a human baby in the wrong environment, bad things could happen when he grows up.

[Art]

Much like stories of Feral children, raised by wolves in the woods then later being found. Adjustment to normal behavior for them was always elusive.

As the twig is bent, so shall tree grow.