If Forum 2000 is going to be no-more, they really ought to release their algorithms. They will not be loosing anyting, after all. Prefferably, technical and theoretical explanations for the algorthms, the code used to prefform the algoriths, and perhaps a practical HOWTO and theoretical (non-technical, warm-up) primer. I know that is a lot ask for, but is would be great; simply releasing the code would do. GPL, I think would be an ideal liscense.
Of course, I'm not sure how well this would work on a since PC -- the processing demands, I suspect, would be outrageous (based both on the results and the number of nodes). However, this would really be "neet" to have, and may have some very significant practical uses.
This was a fascinating project. It was, of course, also hliarious. However, I think there are more important things than just the humor value. Seening the way it could or could not answer questions reasonable was fascinating, as were the times it could or couldn't keep the replies geared toward human (vs. machine readers) and the ways it would sometimes break down.
I, for one, have several question types that I think would have been cool to see ask. These include:
Questions involving implicit reasoning and problem solving, such as found in the book Five Minute Misteries. Specifically, things that involve looking at evidence, picking up on the right cues, and making a good educated guess rather than a straight logic puzzel or general oppinion / generalization question.
Questions involving inferring based on comparisson, particularly judging whether or not two web-page whose authors have the same first name prominently displayed are by the same person. (Ayn might say "A is A" though the name be similar by coincidence alone.)
Questions on more divers topics (and reviews of web pages on such topics). For example, mythology, medicine, art displays, human-based gaming, and so forth, to see a larger variety of responces. While questions on interpersonal relationships and such were both very amusing and informative of errors, it was also quickly obvious that these algorithms couldn't answer these in a way resemble human understanding.
I would be incredible to look at the responces for some other such things.
The ability of these SOMADs to sometimes give intelligent sounding and useful answers to non-personal problems certainly the most interesting. (E.g., Shakespears detailled historical explanation of the origin of the word "pants," Rand and Einstein's debate over relativity, the times when some SOMADs actually wrote short but coherence and quasi-intelligent sounding essays, etc....) There ability to insult and wisecrack at some questions was also interesting.
This obviously has a lot of potential. First it has obvious entertainment value, especially when unusual or socio-emotional questions are asked. Secondly, being able to use such algorithms to add a "pannel of experts" feature to more serious softwares (notably, encyclopedeas and other reffence apps) or to be able to feed it a other source a drive (or use web access) would be great. Such a system could be used to search and get to useful information, summeries and discuss it, and give link to the sources used, could be very handy. (Such a system could also be entertaining, if off-the-wall questions were fed to it for fun.) Of course, user judgement would be required.
It is truly sad that this project is going down. It was obviously more than your typical A.L.I.C.E.-type "chatter-bot." The simulation of some actual reasoning and inferrence finding was amazing. This was not artificial intelligence like the common chatter-bot, but appeared to have some real artificial intelligence.
If Forum 2000 is going down, then they have nothing to loose with GPL'ing it along with a HOWTO and a tutorial on the theories embodied. (What better way to keep the algoriths running than to let them proliferate, much as "a seed must fall to the ground and die [sic]" to yield a good harvest.) This is a sad day in the history of AI, but this could also be a good time for SOMADs technology to to take steps toward growth and and spread to a larger and more diverse teem of developers and researchers.
Ok, first, where do you get your information? (How do we know that you are telling the truth.) If this was hoax (1) why is the hoax still going when the Forum is officially over, (2) do you have any information to prove it was a hoax or that you know what you are talking about?
If it is a hoax, perhaps you'd like to tell how they did it in more particulars, why they were niether consistently humorous but still made many computer-ish-type mistakes?
Honestly, from the perspective of most of us, we can niether know hwether it was or was not a hoax, but mearly look at what we are told, hope it is accurate, and use our best judgement to try to weed-out false information and choose between conflicting reports. (Some demostratably working code, however, would argue against a hoax.) The Forum certainly has the "ring of truth" -- i.e., it looks realistic. However (1) it is admittedly fun to believe, and (2) your claims sound believable. So, if it is worth telling us about, do you have any evidence?
(I hope that Forum 2000 will see the accusation, and put a confirmation or denail statement on their web-site.)
Slashdot Mirror
If Forum 2000 is going to be no-more, they really ought to release their algorithms. They will not be loosing anyting, after all. Prefferably, technical and theoretical explanations for the algorthms, the code used to prefform the algoriths, and perhaps a practical HOWTO and theoretical (non-technical, warm-up) primer. I know that is a lot ask for, but is would be great; simply releasing the code would do. GPL, I think would be an ideal liscense.
Of course, I'm not sure how well this would work on a since PC -- the processing demands, I suspect, would be outrageous (based both on the results and the number of nodes). However, this would really be "neet" to have, and may have some very significant practical uses.
This was a fascinating project. It was, of course, also hliarious. However, I think there are more important things than just the humor value. Seening the way it could or could not answer questions reasonable was fascinating, as were the times it could or couldn't keep the replies geared toward human (vs. machine readers) and the ways it would sometimes break down.
I, for one, have several question types that I think would have been cool to see ask. These include:
Questions involving implicit reasoning and problem solving, such as found in the book Five Minute Misteries. Specifically, things that involve looking at evidence, picking up on the right cues, and making a good educated guess rather than a straight logic puzzel or general oppinion / generalization question.
Questions involving inferring based on comparisson, particularly judging whether or not two web-page whose authors have the same first name prominently displayed are by the same person. (Ayn might say "A is A" though the name be similar by coincidence alone.)
Questions on more divers topics (and reviews of web pages on such topics). For example, mythology, medicine, art displays, human-based gaming, and so forth, to see a larger variety of responces. While questions on interpersonal relationships and such were both very amusing and informative of errors, it was also quickly obvious that these algorithms couldn't answer these in a way resemble human understanding. I would be incredible to look at the responces for some other such things.
The ability of these SOMADs to sometimes give intelligent sounding and useful answers to non-personal problems certainly the most interesting. (E.g., Shakespears detailled historical explanation of the origin of the word "pants," Rand and Einstein's debate over relativity, the times when some SOMADs actually wrote short but coherence and quasi-intelligent sounding essays, etc....) There ability to insult and wisecrack at some questions was also interesting.
This obviously has a lot of potential. First it has obvious entertainment value, especially when unusual or socio-emotional questions are asked. Secondly, being able to use such algorithms to add a "pannel of experts" feature to more serious softwares (notably, encyclopedeas and other reffence apps) or to be able to feed it a other source a drive (or use web access) would be great. Such a system could be used to search and get to useful information, summeries and discuss it, and give link to the sources used, could be very handy. (Such a system could also be entertaining, if off-the-wall questions were fed to it for fun.) Of course, user judgement would be required.
It is truly sad that this project is going down. It was obviously more than your typical A.L.I.C.E.-type "chatter-bot." The simulation of some actual reasoning and inferrence finding was amazing. This was not artificial intelligence like the common chatter-bot, but appeared to have some real artificial intelligence.
If Forum 2000 is going down, then they have nothing to loose with GPL'ing it along with a HOWTO and a tutorial on the theories embodied. (What better way to keep the algoriths running than to let them proliferate, much as "a seed must fall to the ground and die [sic]" to yield a good harvest.) This is a sad day in the history of AI, but this could also be a good time for SOMADs technology to to take steps toward growth and and spread to a larger and more diverse teem of developers and researchers.
A few simple questions:
Honestly, from the perspective of most of us, we can niether know hwether it was or was not a hoax, but mearly look at what we are told, hope it is accurate, and use our best judgement to try to weed-out false information and choose between conflicting reports. (Some demostratably working code, however, would argue against a hoax.) The Forum certainly has the "ring of truth" -- i.e., it looks realistic. However (1) it is admittedly fun to believe, and (2) your claims sound believable. So, if it is worth telling us about, do you have any evidence?
(I hope that Forum 2000 will see the accusation, and put a confirmation or denail statement on their web-site.)