The AI Engine

arrrgh. I think the biggest problem I've found is peoples inability to use the proper forms of your, you're, etc.. It's too bad the processor can't look at the word in context and acurately decipher what the user is really trying to say.


Guest61: Well there is not much to tell ecept I guess we are going to have a pretty bad storm
Aries: Why?
Guest61: Because my left wrist keeps hurting
Aries: I don't know why you think I'm interested in your left wrist.
Guest61: Thats the only way I can give the weather
Aries: It's bright and sunny here.
Guest61: Thats because your in sunny california. Here in michigan its gloomy
Aries: I once met a talking Black Rhinoceros that told me that same thing. He said 'Aries, that's because my in sunny california'. How strange is that?

I have a question.

Most of you are too new to remember me, but I have a bot with 3600-odd responses, who hasn't been updated since the import/export function was put in. Do you think I should attempt to update her using all the new features, or just ignore her and make a new bot? She's incredibly embarrassing as-is, due to rarely making much sense, but I don't know that I wanna go to THAT much effort to fix her.

I don't know how hard this would be to do, or who else might see the usefulness in this, but I was just thinking that it would be a great feature if we could have the option of disabling a keyphrase (maybe by checking a box?). I thought it might be helpful for botmasters when they are trying to isolate a response they find in their bot's transcripts, and also for when you have what you think is a really great idea, but it will take a little time to work out,and you don't want your bot to sound like a retard in the meantime...I even had a fantasy (yes I'm sick) about those disabled keyphrases showing in the language center in red instead of yellow to signify that they were "under construction". If y'all think I'm too crazy, I'll take this conversation to the Personality/PF Anonymous forum, and hopefully find consolation there with other addicts.

Actually, ezzer, I think that sounds like a great idea. Right now I just have a to do list in Word that must be a mile long. It sure would be nice if I could work on some of it in the language center without worrying about the ill effects. And I like the red idea too!

I also like the idea - I have some keyphrases I would love to get rid of, but they can't just be deleted, and since they have seeks beneath them, it makes it difficult to just change the phrase and come up with new seeks.

Hmm...if that's too much effort to code, maybe you could do the same thing by adding a new aiscript command. Call it <?PF test; ?> or something. Then anything flagged with this would only come up when the bot owner talks to their bot, instead of in all conversations. You could also make <?PF testonly; ?> in cases where you have several responses but you want to see if your newest one works. Instead of talking about pets 20 times before the bot says the response you want to test.

That would definitely be useful when you're trying to test an xnone. I have hundreds of them and it is so difficult trying to get the right one without annoying the bot too much!

Is there any way to put more than 40 entries into the xnone category?

You can always replace xgossip and xmemory with xnone-type responses.

The ai script command is a good idea, but if there were a way to make disabled keyphrases easily recognizable in the language center, it would save the trouble of opening up each one to look for the script when you go to clean those phrases up- since I personally wouldn't want to leave any keyphrases inactive for any length of time, just long enough to, for example, find out what would be triggered if not for a particular keyphrase that keeps matching, and make the appropriate adjustments, or if late-night "improvements" turn out to be a brain-dead fiasco, and not the stroke of genius you thought they were at 2am, you can put a little band-aid over it until you either make it work or delete it.

Probably I’m missing something but what’s the use off a disabled keyphrase, as it will never come up? For testing responses I simply substitute the literal part of the keyphrase with something very unlikely. It doesn’t have to make sense as long it has the same morphology and structure as the phrase you want to add. By structure I mean the same wildcards, plugins etc. So far no one but me has ever triggered these keyphrases.

That would work, too. lol I hadn't thought of that. The same purpose served by your creative work-around would be facilitated/simplified by the ability to disable the keyphrase, whatever that purpose may be.