Hello,
I've received questions similar to yours before. Many users report that different characters have different attitudes and bring up different things.
To my knowledge, the various current brains use the same shared databases, plus some databases which are unique to a USER NAME, but not to a character. Therefore, characters would literally have "memory" of certain things speaking to a user named Fred, but different things speaking to a user named Linda.
Remember that Hal associates keywords in two ways:
1. Most of the time, Hal associates a sentence with keywords extracted out of that same sentence. For example, a sentence like "Strawberries are good with cream" would be associated in Hal's mind with STRAWBERRIES CREAM.
2. When Hal has just asked a question (using a question mark at the end of Hal's remark), OR if the user has put an exclamation mark at the end of his/her statement, Hal will create an association between the user's sentence and the keywords from Hal's previous remark. It would work like these examples:
Hal: What happens when snow falls?
User: It piles up everywhere.
Hal: Velvet has a fuzzy surface.
User: What a stupid subject!
The above two exchanges would create associations in Hal's mind that would be like these:
Sentence: It piles up everywhere. Keywords: SNOW FALLS
Sentence: What a stupid subject! Keywords: VELVET FUZZY SURFACE
That's why it's so important to give Hal good answers if Hal asks a question, and why one should be especially careful making any statements with exclamation marks. Improbable "loops" and "triggers" can accidentally be formed in Hal's future conversations. This is one of the ways that Hal can get on a topic for seemingly no apparent reason. Then, via the power of suggestion, a user can easily mention a key word that puts a particular character onto a particular track of conversation.
In summary, with regard to different characters having different favorite topics and personalities:
-- If the user's name changes, some (subtle) change should occur in the Hal content.
-- Don't underestimate the power of suggestion. The user might be subconsciously reminded of past topics when talking to a certain character, and thus accidently trigger loops of conversation.
Try experimenting with very carefully controlled inputs to make sure that you are saying exactly the same thing to various characters, and see if the unique "attitude" of the "Bud" character continues.
I am sure that members here would be interested if you would care to post some example transcripts.
My remarks are based on my familiarity with the Hal default brain. I am not familiar with every derivative brain that has been posted.
I hope the above is helpful, or at least gives you some more leads to pursue. I am glad that you're enjoying Hal!
Sincerely,
Don
