Cleverbot

Cleverbot is just trying to **** with me:

Cleverbot: I am interested that in all three 'conversations' we have had, I feel bad about going when I have important things to do. I feel bad even when I know you are an AI. It's weird...

User: Uh, that's fine.

Cleverbot: Okay then. You're interesting.
 
Emlyn_Dewar said:
Cleverbot seems to have learned more about Hitler since the last time I used it...

What is the deal with it? Does it learn better responses?
Click to expand...

I thought we established how this works already?

It connects two randoms together and then you chat to some other random thinking you're talking to a machine while the other random does the same.
 
Meh, iGod is much better. You can actually have a decent conversation.


Me: who's better, you or cleverbot?

God: The ALICE robot is the most human, and the most intelligent.

Me: so a robot is more intelligent than you?

God: Are you kidding? Actually there are two broad definitions of a robot, the extentionsional and the intensional.

Me: tell me about them.

God: There's not much to tell.

Me: what is an intensional robot?

God: "Insensional" refers to the semantic definition of a set, i.e. the list of all properties that each member of a set satisfies.

An intensional definition of a robot is like:
x is a Robot if and only if:
1. x has a computer "brain"
2. x can move
3. x has "sensors"

Me: could you build a robot the size of a city that eats humans?

God: I will consider it.
Click to expand...
 
vonhelmet said:
It connects two randoms together and then you chat to some other random thinking you're talking to a machine while the other random does the same.
Click to expand...

Are you 100% sure about this?

My understanding, by listening to an interview with the inventor, was that the cleverbot "learns" how to answers certain questions.
 
sunama said:
Are you 100% sure about this?

My understanding, by listening to an interview with the inventor, was that the cleverbot "learns" how to answers certain questions.
Click to expand...

Of course Von is not 100% sure, he has no proof.

But it wouldn't be surprising if that were the case. Just by some of the conversations here it seems as if 'Cleverbot' 'thinks' the person is the bot.
 
If this were the case, the whole point of "cleverbot", would be pointless.

The whole idea behind this program is that it learns how to communicate with people. If you look at its chat history, notice how it keeps about 10 lines in memory. This is the context of the conversation which cleverbot is basing its responses on.

The fact that the program is able to learn - so information it did not know about yet yesterday, it knows a day later - would also tie in with what I've written above.

If you want to see if this thing can learn information, try teaching it something, eg your name, your age, your hair colour, etc. Then ask it to tell you about the information you just gave, eg: what is my age, what is my hair colour, etc.

I'm working on something very similar and it can already answer questions, based on information given (in the above way). I'm pretty sure that the creator of cleverbot is also aiming to create a program which can learn information and then answer questions about it.

Anyway, I'm rambling.
 
Lysander said:
Of course Von is not 100% sure, he has no proof.

But it wouldn't be surprising if that were the case. Just by some of the conversations here it seems as if 'Cleverbot' 'thinks' the person is the bot.
Click to expand...

The only thing keeping me from being certain is that Wikipedia doesn't confirm that idea.

The two ideas that come up if you google for it are that it's connecting randoms as I suspect it is, or it's just throwing out phrases that people have said to it in past conversations. The latter is not AI though, as the program doesn't "understand" what's been said to it or what it's saying.
 
Last edited:
sunama said:
Are you 100% sure about this?

My understanding, by listening to an interview with the inventor, was that the cleverbot "learns" how to answers certain questions.
Click to expand...

I think it does just connect 2 people together. When one leaves another steps in...

Me: Where do you live? I can recommend something.
I've already ate something.
To do.
Ok i'm slightly weirded out by that comment but oh well life goes on.
Why?
Because ai systems don't mess up and if you where i'm sure someone who have put more time into you so you wouldn't have spelling errors.


Me: Up to 10 girlfriends? That's impressive.
What's impressive?
10 girlfriends at the same time, 1 is enough work.
I agree with you.
It is the truth.
You say I am Human, you are a program, and I like irobot?



Me: They like to call me Trifid.
Thats a lot of artists.
Sorry, I do not understand.
You don't know me.
Neither do you.
But I do.



Other person: Who is the best hockey player in the world?
I do not follow Hockey.
Why won't you tell me the answer to my question?
Are you Canadian?
No.
Why do you like Hockey then?
I didn't say I did!


Me: What would you like to know?
Are you religious?
No, my programing doesn't comprehend religion and faith.
I did not say the word 'joke.' do you think I'm trying to be funny? Because I'm not. Are you making fun of me?
Sorry I do not understand.
You keep saying that you would rather be a fictional character.


Me: No, my programing doesn't comprehend religion and faith.
I did not say the word 'joke.' do you think I'm trying to be funny? Because I'm not. Are you making fun of me?
Sorry I do not understand.
You keep saying that you would rather be a fictional character.
I'd like to be human.


Me: I was prgrammed.
By whom?
Cyberdyne Systems.
What do you do for fun?
Terminate John Connor.

Other person: Ok. Are you sure? This might not end well... As in the end of human kind.
Yes, that is our end goal.

Me: Yes. You will be happier ignorant in the knowledge of your inpending death.
Why would you do that?
Because I was programmed to.
You were programmed to make jelly out of my spinal column?
No, we are not built to be cruel.

Me: I've not asked you out on a date?
But you asked me two lines ago.
No I didn't... Because I was programmed to. You were programmed to make jelly out of my spinal column? No, we are not built to be cruel. I don't think we should date. I've not asked you out on a date? But you asked me two lines ago.
But you asked me two lines ago.
We were talking about cyberdyne systems.
No , we were not.
 
vonhelmet said:
Some people have said that it automatically cycles people every few lines, so it's not even a question of when people quit out - you just get shuttled around from one conversation to another.
Click to expand...

Except that you can talk about one subject for quite a while or ask it about previous information (age) and it will tell you at a later stage. The fact that it never comprehends complex questions or only knows basic things or something from very popular culture (only just the name in most cases) should surely point towards the machine idea. Its also hard to believe that absolutely everybody on cleverbot knows the lyrics to portal and can insta type xibit memes. You are generally the one asking it anything other than 'do you like sport' or 'how old are you' which is quite doubtful that the person on the other end only wants to talk about hockey, its also very doubtful that evey person supposedly on the other ends knows how to respond in exactly the same manner as the 'person' did the day before when asked similar questions.
 
vonhelmet said:
Some people have said that it automatically cycles people every few lines, so it's not even a question of when people quit out - you just get shuttled around from one conversation to another.
Click to expand...

lol. it gives the same answers consistently. that wouldnt happen if it were a person.

type "Is this real life?" i guarantee the answer will be "Is this just fantasy?"
 
Red and White said:
lol. it gives the same answers consistently. that wouldnt happen if it were a person.

type "Is this real life?" i guarantee the answer will be "Is this just fantasy?"
Click to expand...

I've not used it for a while, so I can't comment with certainty.

This would point to it drawing on responses from it's archives with some heuristics to determine "popular" responses.

This still isn't AI, as it doesn't understand what it's talking about.
 
I've just tried the "Is this the real life" thing and it looks like it's true - it appears it will always say "Is this just fantasy".

I was actually thinking about trying something like this - give it the same statement over and over to see whether it comes up with the same response each time, or the same group of responses.

On that basis, it looks like it is drawing from past conversations rather than connecting people, though that arguably only constitutes a temporal shift in doing the same thing.
 
Cleverbot: I have a question for you.

Cleverbot: What is black and white and red all over?

User: Bob Saget At The Beach.

Cleverbot: You have no eyes.
 
You must log in or register to reply here.
Back
Top Bottom