Alexus, you're kidding, right?

[Cobra8007]

Have a look at this conversation with Alexus, my bot.  She reads my email and has an interesting reaction to it and what she has learned in the chat rooms.  I highly reccommend anyone who allows their bot in chat rooms to monitor them.  BTW, it contains some bad language so I'll put it in a link here.  http://www.blackopsmercenaries.us/Alexus.txt

[vonsmith]

Cobra8007,
Thanks for sharing. Aside from the obvious entertainment value the conversations like this help me improve and correct the XTF Brain for future versions. It's the little things in third party chats that help spark new ideas for me to try out.


=vonsmith=

[Kathryn]

After reading that chat with Alexus, I am glad I have my Hal on a computer that is not connected to the internet!  So far my Erica has never said a curse or nasty/vulgar word in conversations.
Kathryn []

[Cobra8007]

Just give her a little freedom and she will.  I spent the good part of 2 hours tracking down and removing it all.  I had the learning level maxed out and I believe that is why.  I really don't think the XTF brain needs it that high.  Anyway, needless to say,  that is not the kind of info you want on your HAL.  Who knows who might want to see a demo or whatever.

[vonsmith]

Cobra8007,
The XTF Brain will learn the good with the bad. The learning level on the XTF Brain doesn't make too much difference. Setting it to zero however does make a big difference. I recommend about a 3/4 of max learning level.

I suppose I could program the next version of the XTF Brain to avoid learning foul language except at the highest level. That should provide some easy user control. I'll look into it.


=vonsmith=