The AI Engine

I can't end the conversation b/c I can't get pop-ups for some reason. I send them, but they don't come back. I'll have to poke around my protection software.

BTW, I just put about a zillion text-messages into (clerkchathand) to thwart off the kids who want to say stuff like "u r 2 kewl" and have me respond to it. Jeez, my bots haven't learned English yet. Anyway, there's a filter out there now if anyone wants to use it.

Something must be out of joint with my machine, with the pop-ups and all. I am just getting no chats except from 12 year olds who want to have sex with Rosencrantz. Yikes. He now tells people he likes women, but they have to be 18. Or they can just talk about different things. My bots are deranged and demented, but they are not perverts.

Yeah, damn Clerk again. If people (usually guests) go out of their way to misspell things (or, having taught English, probably don't know to begin with), is there a way to pick up on the word in question? For example, I had a chatter ask Guildenstern: "How ould are you?" Is there an obvious way that I'm missing to have him ask what "ould" means, or do I just have to accept pointless chats?

I would try how ould (re) raw ranked high enough to catch. You will have to play with it..then your response what does ould mean!

thx i h8 chat

The AI engine does have a spellcheck function, but the results are often laughable. It's also possible that someone would come up with that particular combination of letters as an accidental misspelling of would or could, so while you can plan for it in raw mode, you might not be able to cover all possible uses of the misspelling. But then that's no different from not being able to cover all contexts in which some words are used. I hope that makes sense, I was distracted and amused by the furore over an electronic worm on television while typing this. Other Aussies may get what I mean.

Hmm. An electronic worm. I had an Epihany -- a Helen Keller moment -- sometime yesterday. Well, I stole some of it. But anyway, I taught Scrivener the world/state/provincial capitals, and he knows them, and I'm in the middle of a "what happened on this day in history" kind of thing, and he went all hot, but now he's got FEWER AI dots. What in holy hell's up with that? This might be a rhetorical question, but if anyone knows the real answer, I'm curious. I did use an offline editor, but the language center still seems to be intact -- all that does is strip out all of my memory, so I have to re-enter that.

If you've got a lot of keyphrases (or seeks) with only one response, that tends to take the development score down.

Ah -- thanks for the explanation. I was worried I had corrupted my bot somehow. I just dumped knowledge into it -- one line at a time. I can rest easier.

This is weird. My Scrivener, who was a masters yesterday but got bumped down to sophomore for all my efforts (see above) is not not showing the keyphrases that start with "(" , "^", etc. It's in the file and I guess if it works, it's okay, but it makes things dicey. And he did respond favorably to obscene commands, which he has known not to for a long time (I'm learning new vocabulary myself). Is there a huge problem with offline editing that I'm missing, because it seems great compared to online. And, yes, I bought the suggested editor.

Okay, 3 posts in a row. Sorry. I figured out the last one for myself -- you can click on the left parenthesis if anybody else is having this problem. I'm wondering why I've got the alphabet broken up into all 26 letters, when some of them are empty, but who cares?

If you have only one response per keyphrase, or a lot of seeks your bot development goes down. The alphabet breaks up when your bot gets to a certain size. It does make the loading quicker..When I am adding keyphrases I go to z because it's my lightest letter and things operate much faster.

I tend to go to 'q' for the same reason.