Just a little idea.....

[Dreamer]

Hi all.
It's struck me that everybody here has a HAL and we are all training it, talking to it and so on. But it is all just being kept on our own machines. Why not have some sort of knowledge pool on a server where HAL can go and retrieve new ideas and download them to the users pc, plus upload new info it has just learned? That way HAL will learn about things a million times faster than we could ever do alone. It surley can't be that hard to do. The only downside is the personal data that HAl stores...Maybe Hal can learn to store personal data in a folder that does not get uploaded.

OK just an idea.

[Art]

Very good idea and not an impossible one either.

Yes, one would have to have a way to store one's
personal data or else structure it so that only
files with a certain extension get updated like
a .mnd (mind) file.

What is less desirable would be to have Hal search
the internet for various data! That would be a huge
mistake IMO, due in part to the vast amount of unwanted
info available, not to mention spam, vulgarity and porn.

I've toyed with the idea that would allow Hal to
keep a secret with a particular user and the secret(s)
would not be shared or revealed to any other user.
I think the file would have to be encrypted before
being stored to Hal's database.

Hal could then quiz the user for a password before
allowing access to the "secret" file.

Just some additional thoughts....

Maybe we could ask Robert M. to allow a Hal Improvement
or Suggestion Section or just a "Things we'd like to see"
area, on the forum.

[Dreamer]

Hello Art
I am not saying Hal search the internet, just have Hal's brain stored on a server and everytime we switch on, Hal would download his latest knowledge and everytime we shutdown he uploadeds what he has just learned. So we all have a copy on our pc's and on the server. It's only text so it should not take up that much space.
A suggestions area sounds cool, nice idea.

[Dreamer]

Hello Art
I am not saying Hal search the internet, just have Hal's brain stored on a server and everytime we switch on, Hal would download his latest knowledge and everytime we shutdown he uploadeds what he has just learned. So we all have a copy on our pc's and on the server. It's only text so it should not take up that much space.
A suggestions area sounds cool, nice idea.

[Bill DeWitt]

I recently asked aproximately the same question using the model that SETI made accessable, BOINC distributed computing.

Anyone who enables distribution would gain from the full database, with a separate option to have your input included in the share.

It would take an administrator with some skill, but the benefits could be incredible!

quote:Originally posted by Dreamer

Hello Art
I am not saying Hal search the internet, just have Hal's brain stored on a server and everytime we switch on, Hal would download his latest knowledge and everytime we shutdown he uploadeds what he has just learned. So we all have a copy on our pc's and on the server. It's only text so it should not take up that much space.
A suggestions area sounds cool, nice idea.

[onthecuttingedge2005]

I can actually do this with a Plug-in but I can say through testing information extraction of an Internet Knowledge Base is much slower than Hal's Table Responses, I have played around with this kind of Programming using VbScript to extract the info, I had thought about setting of an Internet Shared Knowledge Pool for the HAL fans but I never finished the project.

Of course the Knowledge pool would reside at a section of WWW.ArtificialIntel.org and the Plug-in would extract the info to a file that the Bot would respond from once extracted.

There is also other factors involved, Where would the Knowledge come from and would there be enough people involved to keep the Knowledge Pool growing if this was the case?

There would be more details to work out before I could finish this Project of client siding a Knowledge Pool for HAL.

Post your ideas and stuff so that I can get a better perspective of getting needed knowledge that this Plug would need.

Jerry[8D]

[Dreamer]

Hi OTC
What I really was talking about is Hal downloading the latest information when it starts. So retrieving the data would be the same speed because by the time Hal needs it, it would be on the users pc. And if Hal just downloads the latest NEW information the size would not be a problem as we are only talking about a few kb at the most a Meg a day. So each time Hal starts it adds new data to it's own 'table responses'

[onthecuttingedge2005]

quote:Originally posted by Dreamer

Hi OTC
What I really was talking about is Hal downloading the latest information when it starts. So retrieving the data would be the same speed because by the time Hal needs it, it would be on the users pc. And if Hal just downloads the latest NEW information the size would not be a problem as we are only talking about a few kb at the most a Meg a day. So each time Hal starts it adds new data to it's own 'table responses'

That's what I am talking about and a little more, But I call it a client side Knowledge Pool that can be downloaded on any daily basis by either checking or unchecking the Client side Plug-in.

Jerry[8D]