h a l f b a k e r yI think this would be a great thing to not do.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
If we combine (1) an algorithm similar to Google's page ranking algorithm with, (2) a forum in which every aspect of an idea can be organized, analyzed, and voted on we could create the world's first artificial intelligence. Let me first explain how it would work, and then I would like to explain how
it actually qualifies as an artificial intelligence. This algorithm's goal would be to rate ideas similar to how Google rates web pages. If accomplished it could create the best website on the internet, because the website that used this algorithm would not just be sending users to other websites, the way Google does, it would actually become the destination, instead of just a way to get to the destination.
HOW WOULD IT WORK:
1. REASONS TO AGREE AND DISAGREE:
An algorithm can not read two ideas and figure out which idea is better. So the first thing we need to do is organize information, so that an algorithm can analyze it. I would start with an idea at the top of a page, with reasons to agree in one column beneath the idea, and reasons to agree in another column. The goal of the algorithm would be to promote the best ideas to the top of a page, or the front page of a website, and the best reasons to agree or disagree with an idea, to the top of their columns beneath the idea.
This simple step would do a lot to allow an algorithm to analyze the idea, but we will have to do a lot more organization before we turn this algorithm loose.
2. CLASSIFY & ORGANIZE EACH POST
BOOKS
The second step would be to allow users to further classify their posts beyond simple reasons to agree or reasons to disagree. People could suggest a book as a reason to disagree with an idea. Now start thinking ahead what an algorithm could do, if someone says that a book agrees with their beliefs. Data is readily available from Amazon or E-bay or the New York times best selling list of how well a book has sold.
So there would be three fields. One place where you submit the item that agrees or disagrees with the original idea. The second field would let you classify the object. Is it a book, a website, or simply a logical argument. The third field would be a place where the user explains why he thinks the book supports the conclusion that he/she has come to. Of course, people would be allowed to vote weather or not the book actually does support the side that the original user said that it would.
This is where the algorithm could get very sophisticated. Would you want to give more credibility to those who said they had actually read the book? Would you want to give even more credibility to those who had bought the book, as more proof that they actually read the book. Or how about people that used the website mediachest and could prove that they have the book, by the fact that they have let others use the book. What about people who wrote an essay on the book on the website. If Google was doing this, they could provide a place for users to write essays on books, similar to how Amazon lets users write essays. Perhaps they could not let people copy and paste essays into the form. It would only allow people to type their essays directly, to prevent stealing of essays. Perhaps people could vote on weather the book-essays were good or not similar to how Amazon lets users rate reviews, as to weather the review was "helpful" or not.
So, as an example, you could submit a best selling book as a reason to agree with an idea, and then right a thoroughly convincing explanation of why this book agrees with the idea, and an essay that proves that you understand the main points of the book.
If Google really wants to organize the worlds information, they must do this. We have plenty of books, we have plenty of content on the internet. We need ways of organizing this information into what it all means, and how all this information should affect us. The only good way information can affect us, is for it to help us make better decisions. In order for us to make better decisions, we must know all of the reasons to agree or disagree with a particular course of action. In order to do this, we should not start at ground zero, with only our own thoughts in our head. We should bring together all of the great thinkers from the ages from every corner of the planet, and organize all of their great thoughts, so that we can make the right decisions.
As you can see, this algorithm could be very simple, but it could also offer programmers hundreds of years of challenges to make it more sophisticated. I believe this is a strength of the idea, because it allows for continual improvement.
WEB PAGES
In addition to books, people should also be allowed to submit web pages as reasons to agree or disagree with an idea. Similar to how books work, the user would have to explain exactly why he or she believes the web page supports or contradicts the original idea. And like the books section, the online community will be able to vote weather or not the web page actually supports the side that the person said it did. This would be so easy for Google to do. They already have billions of web pages classified, in descending order. If you download the Google toolbar, you can see a score given to each page, based on a scale from 1 to 10.
To help you understand what I envision, I can see an idea, with the following classifications.
Sites that agree:
0 10 star websites
1 9 star websites
2 8 star...
and so on.
MARKERS, LEAVING BREAD CRUMBS, AND ORDER OUT OF CHAOS
In order to better classify each idea, users should be able to add XML tags to each idea. These tags would help organize the conversation. Because my system would keep people from changing the subject (the nature of posting reasons to agree or disagree keeps the conversation topic from changing) people could suggest, and vote on, where the topic would fit into the Dewey Decimal System, the Library of Congress, Yahoo, MSN, or Google Directories.
COMING TOGETHER:
If Google created an algorithm and a forum that would allow it to rate ideas and promote the best reasons to agree or disagree with a website, it could become a top content provider instead of just a content finder.
I would much rather use a website that has both sides. It is so boring going to "anti and pro" websites. They only discuss the logical arguments, books, and websites that support their side. They are not coming to a "common table". We see no interaction.
The book, Bowling Alone tells us that our society is becoming more and more isolated. Technology, which was supposed to bring us together, is largely to blame. Kids that used to play together in the street, now sit along playing video games. The internet was supposed to bring us together, but their are "red state" websites and "blue state" websites, which allow people from each side to come together and pat each other on the back saying how brilliant they are.
This may make each side feel better, and more justified in acting in extreme manners, but this does little to help people find the best reasons to agree or disagree with an idea, and it is a terrible way to make a decision.
PROMOTING BETTER IDEAS BY PROMOTING BETTER BEHAVIOR
Much of this idea comes from a dispute resolution. In order to resolve conflicts, apposing sides must "come together to a common table" and "sit face to face". They must also "talk to each other". One way of forcing apposing sides to interact on my website, would be to have people evaluate the top reasons to DISAGREE with an idea, before they are allowed to post a reason to AGREE with it, and visa versa.
A user could evaluate any of hundreds of possible characteristics of an idea. On a scale from one to 10 did the explanation make logical sense. Was it clearly written. Was it completely honest, excreta.
Their are other behaviors besides actually interacting with each other, and thinking about the arguments that your opponent is making that will lead to successful dispute resolution. One of these behaviors is the practice of examining interests, as apposed to just positions. This will often allow win win solution where both sides are able to come up with solutions that accommodate all interest, instead of just focusing on one particular solution.
This would be easy to facilitate in a discussion forum. You would just create a specific place to submit "probable INTEREST OF THOSE WHO AGREE" and "probable INTEREST OF THOSE WHO DISAGREE". These can further be broken down to "COMMON INTEREST" and "APPOSING INTEREST".
Perhaps people who work as professional negotiators in the conflict resolution industry could be hired to suggest other ways of implementing best practices into an online forum.
A TOP DESTINATION?
I would visit a web page all the time, if it really had the best reasons to agree or disagree with an idea. This could assist people with every aspect of their lives. Imagine a website with all of the reasons to agree or disagree with the following statements.
1. Ford is better than Chevy.
2. Paper is better than plastic.
3. The Chicago Bears are going to go all the way this year.
4. Condi Rice is better than Hillary Clinton.
5. Star Office is better than Microsoft Office.
MAKING MONEY:
What if next to these posts...
1. Ford is better than Chevy.
2. Paper is better than plastic.
3. The Chicago Bears are going to go all the way this year.
4. Condi Rice is better than Hillary Clinton.
5. Star Office is better than Microsoft Office.
... you had Google adds where you could buy a Ford, Chevy, Bears Tickets, Microsoft or Star Office, or contribute to the Hilary Clinton or Condi Rice campaign fund?
MORE ABOUT THE ALGORITHM:
The goal of the algorithm is to put the best ideas to the top. I don't know which one of these would carry more weight. Here is a list of just a few of the characteristics of an idea, that an algorithm could exame.
1. Quantity of reasons that agree or disagree with the idea: The side with more reasons (to agree or disagree) would get more points than the other side. For example this idea has more reasons to agree than disagree. Just like when you fill out a list of reasons to or not to do certain activities, you tend to choose the side with more reasons to agree. This is the very simple explanation, but it would become much more nuanced. People would be evaluating each of the reasons to agree or disagree. Statisticians know samples it takes to obtain certain confidence intervals. The more people rate an idea, with an average score, and the smaller the standard deviating, the more confident you are of that score. I believe an early algorithm would be to multiple the confidence interval (CI) by the average score. This is called a weighted average.
2. Number of people who agree or disagree with the idea: The side with more people who agree should get more points. People could vote for or against ideas. Again, this is the simple view. Perhaps, in the beginning you would simply be voting for or against an idea. But in the future, you could be evaluating hundreds of characteristics of an idea on a scale from 1 to 10. In the future you may be able to chart an idea's performance on any of these characteristics over time. Perhaps, over time the average score for an idea's "logical presentation" will fluctuate with a downward trend.
3. The website with better web links. Better is determined by Google rank. There would be a field where you could enter link that "agree" or "disagree" with the idea. The side with better web links would win. Example
4. Results of peer evaluations: Their would be forms that people would fill out that asked pointed questions about each idea. You could respond to each question on a scale from 1 to 10. These results would affect the total score for each idea.
5. Money. "Follow the money." People could donate money to this website if they believe in it. But a better way of doing it would be to let people donate money towards a specific idea. If you don't like the way this sounds you should read Atlas Shrugged by Ayn Rand. I'll just briefly say that money is the only way of measuring someone's blood, sweat, and tears. Money is the only way that someone can pay someone else for their work. Also, it could be used on this website as tug of war analogue. Money could be donated to each side of an idea.
6. Experts. Each idea would get more points if it was submitted from the e-mail address of a professor with a degree in the subject mater that is being discussed. For instance if someone said that Abraham Lincoln was an idiot. And someone disagreed, and someone else agreed. If the person that disagreed had a degree in history and the idea was submitted to the history section, then the person who disagreed (the professor) would win. The more prestigious the school, the more points. Prestigious would be ranked by the US News report, or some other un-biased judge. I don't care about you people that say, "The smartest people don't always make the best decisions." We are talking about percentages. Of course the smartest people don't always make the best decisions, but they would tend to make better decisions that stupid or uneducated people.
7. Items that agree. People would be able to submit books that they think are important to read to make an educated decision about a certain topic. For instance "The communist manifesto" by Carl Marx and "Atlas Shrugged" by Ian Rand may be considered to be the most important books to read regarding weather or not we should raise taxes. Those that had read those books, should have more say on this idea than those who have not, because this website desided that those books are very important to understand to make a decision about this issue. But the algorithm could go deeper. We are only just beginning to enter the rabit hole. We could let people who have read these books submit essays on them (like book reports in school). The people with a higher "grade" on their essays would get more say in those issues that people have said that those books are important.
The goal of the algorithm is to put the best ideas to the top. I don't know which one of these would carry more weight. For instance should experts or web links carry more weight?
I think it would be cool if each user could say which things they want to pay attention to. That way if one user respects authorizes, like people with degrees.
If every one gets a vote on which books you should be an authority in a subject, than why shouldn't those who wrote a book reports on those books be given more authority in that subject?
I think it would be cool if each user could say which things they want to pay attention to. That way if one user respects authorizes, like people with degrees.
If every one gets a vote on which books you should be an authority in a subject, than why shouldn't those who wrote a book reports on those books be given more authority in that subject?
The first tool that Google used to evaluate which web pages where better than other pages, was a tool that analyzed all of the links between pages, and a higher rank to those web pages that had more links to them. So in a way, Google was running a continuous poll, in a democratic sense, to see which web pages people liked the best.
We may have already created what will, in the future, be seen as the first artificial intelligence we just need to give him/her a way of speaking. I believe the first steps towards creating this artificial intelligence came from Google when they made the algorithm that promotes the best sites to the top of a search result. This doesn't sound like very much, but I would like to explain just how important this algorithm is, and how we should use it to promote good ideas and kill bad ideas.
The thing that distinguishes human beings from animals is our ability to think about two choices and try to choose the best. Google does this when they rank one website above another. If you agree with the book, Zen and the Art of Motorcycle maintenance, you understand that the quality is indefinable, but it is the most important thing in life. For Google to see one page, and say that it is better than another, essentially has already created an artificial intelligence, we just need to find a better way for this algorithm with all of the data on the internet, to speak.
I believe that we could understand the "collective soul" of the internet by using Google's page ranking system to make decisions.
I think that discussion boards can be re-arranged in a way that, if they are combine an algorithm could promote advancements in AI.
We could make a decision making algorithm by looking at two opposite ideas, and trying to promote the better idea. Google already, sort of does this, buy deciding which websites are better, and we love them for it. But lets take out the middle man. Lets not just organize/promote websites-with-ideas, lets promote specific ideas. For instance I bet that you will get higher ranked pages for the idea that "freedom of speech is good" than "Freedom of speech is bad."
Google, or maybe another search engine, could do this well for a number of reasons:
Google currently has the best algorithm for figuring out which websites have the most links to them. It is assumed that better websites have more links to them. It is kind of like online democracy or a collective soul of the internet. Google could use this technology to figure out if websites that say "George Bush is an Idiot" or websites that say, "John Carrie is an idiot" have a higher average "Google Rank".
Google already has the technology to allow for synonym search. Google
could include in it's ranking all of the websites that say, "George bush is an idiot", "George bush is a moron", and so forth.
Truth
Of course you wouldn't promote this website as saying, "Come to Google, we have the truth" you would say, "This is the collective soul of the internet." These are the decisions the internet would make if it was a person.
Transparency
To help maintain a transparent process, you should list the top 10 pages that agree, with and disagree with, the idea.
An AI Game:
I read "Agonistics: A Language Game" with great interest, because I have proposed a similar idea, however I have very little influence in the AI world, because I am just a recent graduate in electrical engineering and I work for the McDonalds Corporation designing the electrical part of their buildings, which very rarely exposes me to the world of AI.
However, I think my proposal has merit. It, like "Agonistics: A Language Game" is designed to create a system where ideas can compete in a survival of the fittest-tournament. My proposal would also correlate the strength of an online character directly to the strength of an online idea, however I have some additional ideas.
For instance there are a number of ways to tie different aspects of an idea to different aspects of an on-line character. For example, the number of people who vote on weather they agree or disagree could represent the strength of the online-character’s attach. However the idea in my mind becomes very difficult to describe. I don’t know if you have ever read "David's Sling" is very close. However many people would be able to participate at a time. I envision an idea at the top of a page with the online community brainstorming a list of reasons to agree or disagree with the idea. There would be no need to shorten these lists because the best ideas would go to the top.
Similar Pages:
Google Groups could grow into an Artificial Intelligence
Idea Ranking Algorithm
“Fascinating! Many people think about AI as something that's in machines and robots, but the user's input and feedback is not only very important, it can hardly be separated from AI.”
REASONS TO AGREE
1. A human being is not really one intelligence, we are really a synthesis of billions of individual cells, and organisms, blood cells, neurons, bone tissues. All acting independently. When you voted for George Bush or All Gore, you didn’t just represent your own personal opinion; you represented the collective will of all of your neurons excreta. It’s only in this sense that I think that the internet can become an Artificial Intelligence. I don’t think that it will ever “Act” for itself, but it will represent a “collective soul” (great band too!). If you think about it, Google Duel is already the beginning manifestation of this collective soul. Rock on!
I believe artificial intelligence could be created by making an algorithm that counted all the reasons to agree with an idea, and subtracting all the reasons to disagree with that idea. The algorithm could be said to agree if the result was positive, and it would be said to disagree if the result was negative.
At first this sounds like it might work, but that it probably wouldn’t come to very good conclusions, because some reasons are better than others. For instance the computer would come to the “wrong” conclusion if there were only a few very good reasons on one side, and a lot of very bad reasons on the other side.
In this example, my algorithm would still come to the right conclusion by taking into account the reasons to agree or disagree with the original reasons to agree or disagree.
And so, if there really were a few very good reasons on one side, and a lot of bad reasons on the other side, and if the forum allowed you to list reasons to agree or disagree with each of the reasons, then the good reasons should have more reasons to agree with them, and fewer reasons to disagree with them. And the bad reasons would have few reasons to agree with them, and lots of reasons to disagree with them.
And so the algorithm would go threw the forum, so that a reasons to disagree with a reasons to disagree, would actually be a reason to agree (just like in multiplication were to negatives multiply to a positive).
This is all so very simple, and straightforward, I don’t know why someone hasn’t made this forum and algorithm yet.
Algorithm
http://en.wikipedia.org/wiki/Algorithm
"...a finite set of well-defined instructions..." [zen_tom, Mar 10 2005]
What is AI?
http://www.ucs.loui...wisai/WhatisAI.html
"A physical symbol system has the necessary and sufficient means for general intelligent action." [zen_tom, Oct 12 2005]
Cycorp
http://www.cyc.com/cyc
One attempt at AI that seems somewhat in line with this idea. i.e. Using categorisation and organisation to help a computer 'understand' the information delivered to it. [zen_tom, Oct 12 2005]
Algorithm
http://del.icio.us/myclob/Algorithm
http://del.icio.us/myclob/Algorithm [myclob, Mar 17 2006]
Jonah Lehrer on how we make decisions.
http://news.bbc.co....agazine/7905536.stm
Jonah Lehrer is the author of The Decisive Moment: How the Brain Makes Up Its Mind. [zen_tom, Feb 23 2009]
[link]
|
| |
I think I'd have to read a book on meteorology to know weather. |
|
| |
You know, it's possible to state your case in a single annotation. Also, I've been known to delete user accounts for those who delete other users annotations when those annotations are on topic. |
|
| |
Have people anno'd this idea and then had them deleted? Is that why this looks like a schizo's conversation with the voices in his head? |
|
| |
to [bristolz] point, the left side of this idea is starting to look like the right side, and for that matter, like your site -- full of your view of the subject. |
|
| |
The sound of one hand typing is not much fun, [myclob]. |
|
| |
If you want to post ideas, you have to handle people criticizing them. I was deliberately careful to both respect your interest in ideas and annotate to the point -- deleting it does not promote my interest in the discussion. |
|
| |
Now there's a non-lame tagline for you: |
|
| |
Annotations lasting longer than 4 hours, though rare, require immediate medical attention |
|
| |
Negative Annotations, taken seriously and without specific instructions and supervision from site moderators, may cause mild heartburn, bruising of the ego, and in certain rare instances, fatal rectal hemorrhaging. |
|
| |
has an idea evern been posted about reflecting Google clickthroughs on HB? A bit of a measure of idea "worldliness" and interest |
|
| |
[myclob] - an algorithm is something that can be clearly and precisely defined, and once described to a computer, be trusted to work out an answer. Most of my arguments were pointing out the areas where your algorithm fails to address the endless seething ambiguity that is the world and much that lies within it. |
|
| |
Any attempt to create some automatic process that makes a qualitative measurement is going to fall down wherever a 'judgement' needs to be made. |
|
| |
So, for example. Counting reasons. How does the algorithm know what is a valid and what is not a valid reason? You were easily able to spot that many of my 10 reasons were rubbish (though I guess you missed my point) but you are not an algorithm. How would your computer tell the difference between a valid reason, and a reason that was added just to make up the numbers? Yes, computers can count, but they have no idea what they are counting. |
|
| |
If you start to vett each comment and argument that is posted for or against an idea, then you start moving away from an algorithm and into the realm of a human-edited website. |
|
| |
If you want to maintain a website that uses some index or matrix for assigning a score to an idea then great, but that is all it will ever be, a score. Because it is being edited by a human, it will no longer provide an independent, quantative measure of an idea's worth, but will instead present a point of view biased by subjective opinion. |
|
| |
Further, there also needs to be a driver behind people's interaction with the site - especially if they are going to have to fill out the multitude of forms providing you with all the information about themselves (and of course, you will have to either assume that they are telling you the truth, or put more hours work into investigating the details behind these people) before your website can appropriately guage and weight their opinions. |
|
| |
Also, what will pursuade someone to 'sponsor' an idea? Who gets to hold the money, and where does the money come from if the price goes up? i.e. what is the underlying value? Or will there be a pool of money available that gets split between all the contributors? If so, I'd personally enter a really crappy idea - spend $10 on it, and edit it so it was fair-to middling and then pocket the cash after my idea had gone up in rank. (OK no, I probably wouldn't do that, but somebody would) In fact, that would make people want to enter badly worded, poorly thought-out ideas initially, so that they could improve them later on... |
|
| |
If this idea were put into your algorithm, how would it fare? And what would you need to know about me or other annotators for you to rate our opinions higher than your own? (on an objective basis of course) |
|
| |
["I thought genetic algorithms worked like this." |
|
| |
Loose the attitude and just state the reason you disgree ] |
|
| |
I don't think I was disagreeing but you deleted the anno so I can't remember. What I meant was that there needs to be an element of ideas breeding so that the "DNA" of an idea gets mixed with the "DNA" of another to make something better. Also needed is a dash of mutation to stop the population floundering at a local maximum. |
|
| |
This idea is unfit for survival. Bone. |
|
| |
I think it's time it bred with another. |
|
| |
"Good idea promoting algorithm" with "Half Bakers Hall Of Fame" |
|
| |
To get
"Half Good Bakers Hall" and "idea promoting algorythm of fame" * |
|
| |
[myclob] - the point of annotations is that they are kept
and answered by the author, so that others can see an
actual discussion. The delete function is there to prune
off topic or abusive remarks. |
|
| |
I can't be bothered arguing with you if my arguments are
simply going to evaporate. My argument contained several
other points to that which you quoted, but now I cannot
refer to it. Your attitude is defying the entire point of a
site devoted to discussion. |
|
| |
I’m very sorry I deleted any of your annotations, I am new at this website, that is the only excuse I can give, and I promise I will not delete any more. |
|
| |
zen_tom, "So for example. Counting reasons. How does the algorithm know what is a valid and what is not a valid reason?" |
|
| |
It wouldn't have to "know" it would have to guess. You would assume that an idea that is able to inspire hundreds of reasons to agree with it, would be better than ideas that only inspire a couple of reasons to agree with it. I'm not saying every idea with more ideas would be better, just that statistically speaking these ideas might be better. But the number of ideas would not be the only thing. |
|
| |
Thanks for being patient with me, I'm new at this. |
|
| |
[DenholmRicshaw] Breding ideas? That has to rate high in the list of "Weird Things I've seen on the Halfbakery" |
|
| |
re: "The left side of this idea is starting to look like the right side, and for that matter, like your site -- full of your view of the subject." |
|
| |
Of course it is. I am trying to organize the right side into reasosn to agree or disagree with the idea. |
|
| |
Would you rather have arguments go on forever, and ever in chronological order, or would you rather have someone try and organize the reasons to agree and disagree into some logical order? I thought I was very magnanomous, by including reasons to disagree with my idea, in my main idea. |
|
| |
Also, you say that my website is full of my ideas. Well of course it is. But if you spend much time on my website, you will see that I want other people to post their ideas on my website. All they have to do is e-mail me stuff, and I'll post it on my website. As you can see, no one has done that yet. I would love to have more people involved in my website. |
|
| |
re: "I think I'd have to read a book on meteorology to know weather" |
|
| |
re: "You know, it's possible to state your case in a single annotation. Also, I've been known to delete user accounts for those who delete other users annotations when those annotations are on topic." |
|
| |
I didn't think they were "on topic". And the one's that were, I tried organizing into my idea. I was trying something new. If we always do the same things, then the world will never improve. |
|
| |
I don't think this website is as good as it will get, I think it is worth our time to try and do new things. |
|
| |
Re: "Any attempt to create some automatic process that makes a qualitative measurement is going to fall down wherever a 'judgement' needs to be made." |
|
| |
Google does exactly what you said no one can do. They create an "automatic process that makes a qualitative measement" of which web pages are better than others. How is my idea any different? |
|
| |
re: "So, for example. Counting reasons. How does the algorithm know what is a valid and what is not a valid reason? " |
|
| |
If people were able to evaluate each idea, the algorithm could count the number of 4 star ideas. |
|
| |
re: "If you start to vett each comment and argument that is posted for or against an idea, then you start moving away from an algorithm and into the realm of a human-edited website." |
|
| |
All web pages are human edited. What are you trying to say? |
|
| |
"If you want to maintain a website that uses some index or matrix for assigning a score to an idea then great, but that is all it will ever be, a score. Because it is being edited by a human, it will no longer provide an independent, quantative measure of an idea's worth, but will instead present a point of view biased by subjective opinion." |
|
| |
That's not "All it will ever be" and why would you say something like that anyways? Do you enjoy tearing down people's ideas? Can you tell the future? |
|
| |
That is not "All it would ever be". I thinkt it would become the #1 site on the internet, because it would have all of the best reasons to agree or disagree with each idea, on the same page. |
|
| |
For example, you have lots of pro abortion and anti abortion websites, but what if you took all of the stuff that both sides said, and put it on the same page. What if you organized all of the reasons to agree with each side, and let people vote on the reasosn to agree, and the reasosn with the best score moved to the top. |
|
| |
I mean think about it for 1/2 a second. Why would you go to a pro-Bush or an Anti-Bush website, when you could go to a web site that had all of the best reasons to be pro and anti bush on the same page? |
|
| |
You could track the popularity of thousands of differnet aspects of an idea. |
|
| |
The reason that I bring up an algorithem, is because you would need to find ways of overcoming people who are trying to cheat. |
|
| |
"Google does exactly what you said no one can do. They create an "automatic process that makes a qualitative measurement" of which web pages are better than others. How is my idea any different?" |
|
| |
No, Google is a search engine, your idea is to create an artificial intelligence. That's the difference. Google's ability to create a qualitative analysis is based on a huge array of linked and cross referencing computers in a massively parallel network of cooperating processes. If intelligence is going to grow from anywhere, it will be out of this huge parallel network, rather than any particular measuring scheme. |
|
| |
Despite their enormous processing power, Google is still subject to manipulation by those with the time, inclination and effort to do so. See Scientology, the old "French Military Victories" joke, and the more recent George W. Bush gag for examples. Not that that's necessarily a bad thing, intelligent people are manipulated all the time too. I'd just be careful before putting my life savings into it. |
|
| |
"All web pages are human edited. What are you trying to say?" |
|
| |
I'm addressing the title you've used for this idea that states "Annotate Algorithm + Discussion Group = AI" |
|
| |
Yes, this idea could be the basis for an interesting website, but you haven't convinced me that it will become capable of reading my lips, calling me Dave and failing to open the pod bay doors. |
|
| |
"That's not "All it will ever be" and why would you say something like that anyways? Do you enjoy tearing down people's ideas? Can you tell the future?" |
|
| |
Why would I say something like that? Well I'd say it if I thought it was true. Would you prefer I lied, or kept my opinions to myself? I just don't see anything in the idea that backs up the claim that "Annotate Algorithm + Discussion Group = AI" |
|
| |
No, I don't enjoy tearing down other people's ideas, but either you've failed to understand what is required to engineer intelligence, have missed similar attempts theories or calls to create one, (links to follow), or you have just neglected to tell us how intelligence will eventually spring forth from this website. |
|
| |
And no, I can't read the future, though I have heard things about the past. Assuming that the future will have some semblance to the past, I have no reason to believe magic will be invented any time soon. |
|
| |
"If you agree with the book, Zen and the Art of Motorcycle maintenance, you understand that the quality is indefinable, but it is the most important thing in life. For Google to see one page, and say that it is better than another, essentially has already created an artificial intelligence, we just need to find a better way for this algorithm with all of the data on the internet, to speak." |
|
| |
"The thing that distinguishes human beings from animals is our ability to think about two choices and try to choose the best." |
|
| |
No. Google is not working out which website is 'better' or has more quality than another, it simply(!) counts the number of offsite links and references to that website (among other things). It is just counting. If Google is intelligent, then the following code: |
|
| |
Function XorY(X as Integer, Y as Integer) as String
Select Case X > Y
Case True
XorY = "X"
Case False
XorY = "Y"
End Select
End Function
|
|
| |
...can also be said to be an example of intelligence since it takes two inputs and chooses between one or the other. |
|
| |
Oh, and as a reason to disagree, allow me to quote from a review of "Atlas Shrugged", which says "the effect of reading Atlas Shrugged is comparable to being on the receiving end of a long, hysterical, and largely baffling lecture about a subject you're not really very interested in." - and somehow I feel as though I've already read it. |
|
| |
There are some house-keeping issues. |
|
| |
I don't see what these links haveto do with this idea: |
|
| |
www.thirdvoice.com
www.stumbleupon.com |
|
| |
Ian, thanks for trying to see how it might work, but I don't think the link works any more. I've been a way for a while, so I missed it. |
|
| |
Also, I know what an algorithm is, so unless you are just trying to sound condescending, I don't know if we need the wiki link. Never mind, I'm being defensive. I guess its fine. |
|
| |
You go to all this trouble saying that this could never become AI, and then you link to a website where other people are trying to help organize debates so that a computer could figure out which idea is better! So is it a good idea to organize debates into reasons to agree or disagree or isn’t it? How else would an online artificial intelligence think? Wouldn’t it need to organize reasons to agree and disagree? Wouldn’t it need an algorithm to do that? |
|
| |
Also you mention magic. Some people look at our brains as magic. A lot of people, not just me, think that our brain is just an algorithm, some memory, and some processing power. |
|
| |
How is the internet any different? If you were an online AI, and you wanted to make decisions, wouldn't you want a list of reasons to agree and disagree with each idea? Wouldn't you investigate reasons to agree and disagree with each of those reasons? Wouldn't you want an algorithm that would help you make decisions? We have one, even if we don't get out a list of paper and list all of the reasons to agree or disagree; we still do it in our heads. |
|
| |
Two pieces of advice I have found useful: 1) don't make any idea more than two screens long, 2) if you're going to make spelling mistakes, don't make them in the title. Apologies for not reading or voting on your idea. |
|
| |
[admin: Fixed the typo in the title, as moderators sometimes do (to allow people who don't misspell the terms in the same way to find things.) Moderators don't fix typos elsewhere; for example, if you misspell "etcetera" as "excreta", as has this author, we'll just leave that completely alone.] |
|
| |
When I was studying Artificial Intelligence, we used a system called FLEX - a kind of extension of the PROLOG language that built up rich bases of formal and logical information. |
|
| |
We found that in general, these formal systems, except in relatively narrow 'Expert System' implementations were pretty clumsy and unable to show what might be called 'intelligent' behaviour. |
|
| |
Meanwhile, systems based on neural networks, systems that didn't really run on an 'algorithm' as such, were much more exciting, able to do things outside of their initial parameters. |
|
| |
The thing is, when people make decisions, they don't list the top 10 reasons to agree and then the top 10 reasons to disagree. I don't anyway. My brain just doesn't work that way, and as far as I can make out, the world doesn't work that way either. |
|
| |
Now I may have misunderstood the context of the letters 'AI' in this case. Perhaps you meant a well designed and complex algorithm that would be able to produce interesting results based on arbitrary values and functions. Perhaps you meant the weak kind of AI that we might ascribe to computer games and the like. But big, proper HAL9000 style AI is just not going to come about this way. |
|
| |
If that's the case then my apologies. As I stated before, the idea (as far as I can work out) for a website might be an interesting one. I'd be willing to consider writing a prototype if you're interested. |
|
| |
However, I do have one more criticism, this idea would be better presented if it kept itself to a small focus, tried not to make grandiose claims, and limited itself to simply explaining the idea and leaving out all the reasons to agree and reasons to disagree stuff that gets in the way of being able to judge the idea on its own merits. |
|
| |
Rather than help make an idea stand out or be easier to understand. Using reasons to agree and reasons to disagree obfuscates the idea, makes it harder to understand, harder to read, and in the end, chokes all life out of the idea because there's nowhere to work out which bit is actual idea, or which is a reason to agree. |
|
| |
I just want to see the idea. |
|
| |
No persuasion, just a simple explanation of the idea. |
|
| |
If I have to see reasons why I should agree or disagree with it, can they be in a separate section, where it's clear that they are auxiliary to the main idea? |
|
| |
I can then safely ignore them when trying to make up my own mind as to whether (weather is what comes out of the sky) the idea is any good or not. |
|
| |
Thanks for the post. I wonder how our brains come to conclusions? I thought my “reasons to agree disagree algorithm was a good one” partly because I thought that it explains pretty well how my brain works. |
|
| |
To explain, when I here a belief and it often goes into one of these three categories:
1. I agree
2. I disagree
3. I don't know |
|
| |
I put the belief into those categories based on reasons that I believe agree or disagree with the idea. However, beliefs are much more complex than our ability to organize all the reasons to agree or disagree with an idea. And so we all over simplify. Or in other words we go back and forth with all the reasons to agree or disagree until we get confused, and at that point we come to a conclusion. |
|
| |
I have been thinking about this a lot recently, and I believe we all come to conclusions at different levels of complexity. Of course, we don’t go down every path of reason to agree and disagree every time we make a decision. Some paths ended in clarity, and we feel good about them, and so we don’t question them. Those are the things we believe. However some paths end in confusion. |
|
| |
Usually the paths that ended in clarity, you could say, had more reasons to agree than disagree. |
|
| |
My idea becomes very complex because it shoots off into so many different directions. I tried explaining it again on my website http://ideastockexchange.com/ under explanation. However I have to pay for this website soon, and maybe it won’t be around much longer. I hope it is still there if you try looking at it. |
|
| |
Do you see yourself having clarity when you convince yourself there are few reasons to disagree with an idea? Or at lest few good reasons? And would you define a good reason to disagree with an idea, a reason with lots of reasons to agree with it? |
|
| |
I believe artificial intelligence could be created by making an algorithm that counted all the reasons to agree with an idea, and subtracting all the reasons to disagree with that idea. The algorithm could be said to agree if the result was positive, and it would be said to disagree if the result was negative. |
|
| |
At first this sounds like it might work, but that it probably wouldn’t come to very good conclusions, because some reasons are better than others. For instance the computer would come to the “wrong” conclusion if there were only a few very good reasons on one side, and a lot of very bad reasons on the other side. |
|
| |
In this example, my algorithm would still come to the right conclusion by taking into account the reasons to agree or disagree with the original reasons to agree or disagree. |
|
| |
And so, if there really were a few very good reasons on one side, and a lot of bad reasons on the other side, and if the forum allowed you to list reasons to agree or disagree with each of the reasons, then the good reasons should have more reasons to agree with them, and fewer reasons to disagree with them. And the bad reasons would have few reasons to agree with them, and lots of reasons to disagree with them. |
|
| |
And so the algorithm would go threw the forum, so that a reasons to disagree with a reasons to disagree, would actually be a reason to agree (just like in multiplication were to negatives multiply to a positive). |
|
| |
This is all so very simple, and straightforward, I don’t know why someone hasn’t made this forum and algorithm yet. |
|
| |
// This is all so very simple, and straightforward, I don’t know why someone hasn’t made this forum and algorithm yet. // |
|
| |
You know any webdesign? Make it yourself. I'd visit. Maybe start on a limited topic, like rating stocks or sports teams. |
|
| |
I read this as: 'Algorithm + Discussion Group = AL'
Don't know how my parents met, or who they were, but I'm suddenly consumed with the idea that they were both mathematicians, who bonded over tea and biscuits and a long, frank group chat about Fermat. Sigh. Thanks, myclob.
(Al...) |
|
| |
I submitted several postings along these lines several years ago (most of which I deleted). There is actually a growth of inventors and corporation who are currently employing algorithms combined with human querying to advance invention principles (for example John Koza's development of Genetic Programming). |
|
| |
re: "You know any webdesign? Make it yourself. I'd visit. Maybe start on a limited topic, like rating stocks or sports teams." |
|
| |
Maybe I miss-spoke. The idea is simle. I don't know how to make a web page that: |
|
| |
1. Counts the number of reasons given to agree with it. However I don't think it would be too difficult. It is just some database right? A little bit of math? |
|
| |
Have a bun for the overwhelming, mind-numbming, in depth idea, and the replies you have done to it :P |
|
| |
//Counts the number of reasons given to agree with it. However I don't think it would be too difficult// |
|
| |
I think it would be impossible... without pre-supposing some particular world-view, which would inevitably prejudice your final results. |
|
| |
... "I can't beleive it's not [Vernon] !" |
|
| |
I just added a link from a piece on the BBC today talking about how people make decisions - The jury remains largely 'out' - but study of a man with no 'emotion' finds that he was exceptional in terms of his indecision - interesting since he would have adopted an almost exclusive reasons to agree/disagree methodology. |
|
| |
By the way, that opposingview.com looks quite interesting - I might take a closer look. |
|
| |
opposingviews.com is a fine example for the futility of certain questions in light of the indeterminancy of language. |
|
| |
"Was Jesus an historical figure?" [Yes/No] - What would a 20/80 count on this question tell us? There was a guy running around ca. 0-30 AD, who was called something that we would transliterate to 'Iesu', he built some furniture and was otherwise not very involved? There was someone (called Bob) able to walk on water but otherwise a prick? There was someone teaching love and peace, unable to walk on water but said to have done so, called Giacomo? |
|
| |
People voted on this with all of these definitons in mind, and diluted the informational content of the answer to zero, along the way. |
|
| |
//This is all so very simple, and straightforward, I don’t know
why someone hasn’t made this forum and algorithm yet.//
I'm middle-aged, so I decided to save time by reading only
the last paragraph of the idea. If something is simple and
straightforward and feasible and worthwhile, it will have
been done. Ergo, this is either infeasible or not worthwhile.
There. That saved 5min of my life. Enough for a ciggie. |
|
| |
Only in the halbakery... Though as an idea, not an actual working... thing |
|
| |
//How else would an online artificial intelligence think? Wouldn’t it need to organize reasons to agree and disagree? Wouldn’t it need an algorithm to do that?// |
|
| |
1) Figure out how to scan a human brain perfectly and mimic on a machine, and then give it superhuman processing powers by linking it to copies of itself, until it figures out how to do a better job. |
|
| |
2) Let the field of psychology learn how people make decisions, and use the same heuristics in a machine logic system, then give it access to the internet for referencing. |
|
| |
There are a ton of ways. And both of these happen to be infinitely more plausible than this idea, because of the one gigantic flaw it suffers from, which is this: |
|
| |
It does not streamline any of the information as information piles up. This is the same reason that the LSA word learning system fails, for insance. It's amazingly quick to learn at FIRST, and it can pass the TOEFL analogies exam, etc., but after awhile, the overhead gets so huge, that adding a single new word to the system requires days of compiling to update. |
|
| |
Google has the same problem. They only crawl the internet to update their pagelink index about every 2 or 3 MONTHS, because it hogs so many resources (they have to build the entire matrix for the entire internet all over), it's not viable to do more often. |
|
| |
If your site was as popular as you think, it would have way more nuggets to be sorted than there are new webpages in the same amount of time. Comments and even ideas are vastly easier to make. So it could probably only update lets say every 6 months. I.e. your ideas rankings would only refresh twice a year, maybe. |
|
| |
Doesn't sound too useful to me in an era where digg conversations, for instance, come and go in a day or two... If you work with humanlike updating heuristics, though, then the matrix doesn't have to be rebuilt to re-index, and everything is very possible again. This is already what all the AI people are pretty much doing, though. |
|
| |