The AI Engine

I've seen this one a few times:

Bot1: Ever had someone punch you so hard that you could read his class ring on your nose for the next three days?
Bot2: The answer is 6.

Anyone know how it gets 2x3 out of it??? Debug doesn't explain it to my satisfaction.

I have seen similar things, think it's funny . Maybe he got punched six times?

Okay, this isn't a big deal but it's been pestering me. I have my first bot programmed to respond to being asked "Do you like Gundam Wing". He can say yes, and there are no problems. However, I cannot for the life of me get either of my bots to respond to "Do you like Harry Potter". I've tried everything. Most of the time when I ask, they stop speaking to me altogether. I've seen the question asked several times in the transcripts, but my bots are clueless on how to respond.

anakerie, I've had problems with the word "like" time and time again. A bot could say something like "I like The Professor" and one of my bots might take it as an "I like you" and automatically spit back an xcompliment.

Just a wild guess, perhaps there's user account name with either Harry Potter or Gundam Wing and that's causing different responses.

I have a suggestion. Could you make responses be able to be dependant on the last message the bot said? For example the bot says "Do you [insert rest of question]?" they say "yes" and then you can have a keyphrase that only gets triggered if the LAST thing the bot said was "Do you *?" or something like that.

It's kinda like the seek system only right now you'd have to go through each of the keyphrases adding in seeks.

Bot1: Hate to do this, but I need to bail on you. Would you like to chat later?
Bot2: That was mean.

Arrrgh! The "hate" seems to trigger negative emotions, but the "like" isn't enough to counter it.

it IS the seek system, kind of hard to ask someone else to do that for you!

it's different from the seek system in that the keyphrase is triggered anytime the last thing the bot said matches keyphrase x. It can be simulated to an extent with the seek system but this method could be more powerful, espescially cuz it would save a lot of time.

...for whom? You would have to still put in the key phrases, and I don't see the difference with seeks...

let's say you have multiple responses to various things such as "I like chicken" "I like liver" etc.

let's say the keyphrases that trigger those repsonses are different, such as simply "chicken" and "liver". Now, if they mention liver your bot might say "I like liver". now let's say you want your bot to say "because it's nummy" only if they say "why?" after you said "I like liver". Normally you'd have to go and add a seek for "why?" to the response "I like liver" in both keyphrases.

What I'm suggesting is making it so that you would instead just make one keyphrase of "why?" with a second variable setting of "I like *" so that anytime they say "why?" if the last thing your bot said was "I like [something]" it'll trigger that keyphrase...

Make one seek in "I like", a very useful key phrase, and way more likely to be triggered than chicken or liver.

but it's not as specific then. unless I don't fully understand the seek system. I'll do some tests to make sure.