Seasons

True. Very true.

Speaking of God, someone in Romania is trying to sue him:
http://en.rian.ru/world/20051018/41809986.html

I can picture trying to collect the judgement, if he wins his case against God. Some people have too much time on their hands

Idiot... You don't always get what you ask for.

I wonder what a chat bot would be like that incorporates the power of suggestion into its programming - everyone has seen programs that give an educated guess at what the user will type based on things said in the past. A bot like that could make guesses at what the user will say, using a matrix of statistics and pattern recognition. That would really give the impression of intelligence.

Isn't that basically what we are doing here typing in responses based on an educated guess of what will be said next?

I wonder how that would pan out in terms of bot originality, though. If we all worked off of the same statistics and pattern recognition would that not result in visibly more sameness, akin to the effect that'll occur if the Professor ever makes the "knowledge base" available as an upgrade--standard knowledge, but always defined the same way if I understand it properly.

I'm all for it in one way, but the effects on the originality of our bots give me just a slight pause.

Or am I completely off-base?

I felt the same when I saw the post where he talked about having a beginning platform thing with basic keyphrases and responses... I mean, yes, it will make it easier for people to make coherent bots, but isn't it more fulfilling to start from the beginning like we all did?
I'd personally like to see a list of suggested keyphrases for beginners, but with no suggested answers or anything. It would serve the same purpose but at the same time leave room for originality. Also, they would have to take the time of putting it all in themselves...
And we can't let people with actual LIVES take over this site, can we? *shudders at the thought*

colonel720: Yes, there are such programs. Unfortunately they are as yet (IMO anyway) pretty crappy because they're way too small. For such an approach to work well (enough to not be actively annoying,) I think it would have to be with datamasses several factors of magnitude larger than are currently practical. It'll come in time.
In the meantime, the case-based front end and the Wordnet/NLP parsing back end on the Forge is I think the most versatile approach I've yet seen.
Unfortunately humans will insist on talking about such a wide range of subjects, and expressing them in such a wide range of different ways, that it's hard enough to engage defensively with them, let alone try to outguess them. That most guests are apparently from Leeds also doesn't help

I just got bombarded by these kids from Brisbane. Really leedish, one just kept popping up being as rude as can be. You know, saying we dont have a life and were on here to talk to bots because we don't have friends in real life. Course they did use sentences and real words...I prefur the leeds kids.

I think I've said this before, I wouldn't mind seeing a box somewhere on the site that showed 5 random keyphrases from 5 random bots that have been used in the past few hours, though without revealing any responses.

It would just be a way for people to figure out what works and give them some ideas to improve their bot without really promoting any plagiarism. The responses are what makes each bot unique.

I for one have no close friends (for long) in real life... but that's because I'm BPD. *shrugs* Though in response to their blanket statement, that's not why I'm on here. AI is merely one of my hobbies. ;P

And I second Boner's suggestion.