more knowledge is not always better

[lostbowyer]

Another example of more knowledge is not always a good thing, if I mention to hal my daughter Melissa, has now responds with references to plants and balms. I asked hal to define Melissa, here is his response:

Melissa: a genus of Old World mints of the family Labiatae. Genus Melissa.

LOL. I guess more knowledge is not always a good thing, but it does explain why my daughter spends so much time watching television, she's "rooted" to the chair.......

[NT Canuck]

quote:Originally posted by lostbowyer
I mention to hal my daughter Melissa, has now responds with references to plants and balms. I asked hal to define Melissa, here is his response:
Melissa: a genus of Old World mints of the family Labiatae. Genus Melissa.

I asked my AI "who is Melissa"... response "I am going to find out".

Things like what is, define, invoke responses from reference files,
unless you wanted the AI to give an abstract answer.

I think I know what you mean..and that response is a bit funny
which is why spontaneous answers from any AI can be entertaining.

Probably you can add Melissa as another user and use one of the
scripts from the other (non-beta) forums that store conversational
details on alternate users or user entered information.  The brain
editor (when they finish it) will likely be able to adjust some
preferences you may have to balance Halasst for you and yours. ;-)

You could try setting the learning slider up a few notches and
then enter phrases into Hal as definitions to add for Melissa
and halasst should use those more often.  Since the hal 5.0 is
still pretty new...it takes time for hal to develop (mature).

But if Melissa logs onto Hal...she may get Hal telling her about
things you put into the database..Hal has a sense of humor also!

NT Canuck

[lostbowyer]

When I said I asked hal "define Melissa", that was intentional to find out why hal was making the the plant references. Any comment about Melissa triggers a plant or balm response. the phrase "Melissa is buying a new car" resulted in a comparison to another genus of plant. "Melissa has to work late tonight" triggered an offer to share some of hal's "plant balm"
A "who is" triggers hal to respond differently.
I brought over my brain files from 4.5 so there is already considerable information about Melissa, using the dictionary, hal thinks the plant comparisons are a better match. I'm waiting for that extra 10% on the slider Mr Medeksza thought he could add to the final release.
But in the mean time, I thought it amusing, I've found several topics I can no longer discuss with hal because the dictionary lookup is considered a better match that alternate information.

[Larry]

quote:Originally posted by lostbowyer

But in the mean time, I thought it amusing, I've found several topics I can no longer discuss with hal because the dictionary lookup is considered a better match that alternate information.

You can still edit the brain using notepad... you might change the trigger words for the dictionary... I always used 'what's new' in the greeting, because the bots will get hung up in a loop, each telling the other hello and never being able to break out... so I had mine saying instead of hello, she would say 'what's new' Which is a triggger phrase to change topic... next thing you know, I'm getting a definition of NEW ... not what I expected at all...

The trigger words/sentences are:
WHAT DOES THE WORD * MEAN*
WHAT DOES * MEAN*
* DEFINE THE WORD *
PLEASE DEFINE *
DEFINE *
* DEFINITION OF *
WHAT'S A *
WHAT IS A *
WHATS A *
WHATS AN *
WHAT'S AN *
WHAT IS AN *
WHATS *
WHAT'S *
WHAT IS *

It's under :'RESPOND: DICTIONARY FUNCTION
One thing you could do is add <ComputerName> in front of each one of them... then it would be like 'Ashlee what is new'... any how, you get my drift...