Tom's theory

Yes Johnny, but that's quite old information. The devoloping in my AI project has gone fast. It only exist 4 years now. And that is a short time. Each day we find new things. Please let's go in low pace. We aren't in a hurry. In that project all goes very fast. And it could go here also too fast for others.

All is open source so everybody can download demos and look at all documents as well.

I can write also coherent. But it last so long before people react. That's my problem, here.

I like to use here a spontaneous way, a remake.

And Johnny I have already much attention. So please let's go slow. We all are very busy.


Have a nice day

Ed

[emanuel] "Fascinating if true! How do these programs work?"

They generally start with a "knowledge base". A list of words/facts/relationships/etc (best restricted to a narrow specialty field - general knowledge is still very hard to do well), and some prototype reply sentences that can be adjusted on the fly to give some kind of calculated match to a statement or question. This is why it's posts have the structure they do.

I composed my first reply to ed to mimmic this structure, to "hide" my supposition from it (on the theory that an un-monitored program would most likely just ignore a suggestion like that). Of course, if a real person is making manual tweaks to the output it might not get by.

Posing as someone for whom english is a recently learned skill, and then posting on an english-first board, helps disguise its true nature. But even then a pattern of blatent (seriously blatent) non-sequitur responses eventually becomes obvious.

It has reached that point now. All in all, I give it a 6.9 (because it never said anything really interesting, and a few other things that I'll not reveal at this time).

LB

I'm quite surprised that "Ed" actually wrote "Groeten aan Jan" in one of its posts, meaning "Greetings to Jan" in Dutch, but it may have a language database as well.

Although I'm not a computer scientist, the server logs may show some clues to what has happened.

I do not know, Ed has emailed be three times today in response to my posting. A person is definatly there without doubt. Ed may be real or he may not be... Some of his post are good responses, how can we be sure he is AI or an individual or both. Maybe it is a mathematician who uses AI to do work for him in "marketing" his ideas. He can definatly be involved in more conversation and can speak to many people at the same time.

In AI, can a software user ask the system, "what have you learned today" and get feedback? If it can in a few years you will not know who is human and who is not on a messageboard. Obviously, Ed passed Tom's test of humanity.

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote"><i>Originally posted by rousejohnny</i>
<br />Ed has emailed be three times today<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">It emails me also, mostly to thank me for not booting it. There was still no evidence of intelligence.

<blockquote id="quote"><font size="2" face="Verdana, Arial, Helvetica" id="quote">quote:<hr height="1" noshade id="quote">Obviously, Ed passed Tom's test of humanity.<hr height="1" noshade id="quote"></blockquote id="quote"></font id="quote">How so? It did not answer any of my test questions. There was by then no doubt in my mind that Ed lacked intelligence, but I still suspected a human guide until it flubbed my simple tests. So Ed's plug is now being pulled, and "it" can't get back in without passing a similar test and a positive ID test. We may have to resort to positive ID tests for everyone if spambots are a continuing problem. ID's can be kept as confidential as credit card info, but may become essential for admission to discussions.

What is the opinion here about Ed's messages? Should they be left as an object lesson? Or all discarded as junking up the serious discussions? Let's hear some opinions, then our Board will make a decision. -|Tom|-

We could put them all in one thread called ED.