https://ai.fandom.com/ is a wiki project hosted at the fandom wiki farm. It is about Artificial Intelligence and contains only a little amount of articles. According to the changelog the Wiki has only a small amount of contributers, so it is in an early stage. The main advantage is, that it's possible for a normal user to register an account at Fandom and then he can contribute. I've tried out and made the first edit yesterday.
The most important question around each wiki project is, how to increase the article count. In a classical wiki this is the bottleneck. The first 10 articles are created by enthusiasts but then the projects slows down and after a while nobody writes new content. How to avoid such a workflow is simple. The wiki has to be realized as a social wiki. The idea is that the minimal contribution is similar to post an URL link to a social network.
That means, the users of the wiki aren't asked to invest their spare time in article writing and content production but they are asked to drop down URLs from their favorite youtube videos and their own projects which they want to push forward. The funny thing is, that many users out there are searching for a place to post their backlinks. And the AI Wiki is the perfect place for doing so.
The more important question is, what is the meaining of a website which collects URLs? Does this website has serious readers? Oh yes, because a manually curated playlist about usefull AI ressource is something which is not available right know. The problem is, that so many great content is out there in the internet but what is missing is a peer review hub which is collecting all the URLs and commenting about the quality. Somebody may argue, that the Google search engine is the clearing house for the internet. But google is only a search engine it is not a social network.
There is a need for a place in which users can paste their favorite URLs and comment the URLs from other users. This allows to discuss in a relaxed environment current trends in robotics, machine learning and AGI. A minimal newly created Wiki article contains of:
date, URL
The next user can add more information for example a single sentence comment or perhaps a two paragraph text. Then the article looks similar to a Wikinews article. The typicle wikinews contains of two url plus two paragraph of text.
It is important ot understand that social networks are working in the reverse direction. The idea is not to create first an article about neural network and in the second step the given article is extended with ressources, but the existing ressource stands at the beginning which is extended by comments.
Let us go a step backward. The idea of a content wiki is, that no other ressources are available. The content wiki is the first place in which an author can write an article. So it is a blog but only in a group. This kind of understanding results into self-blocking groups. As a result the users stay away from the wiki because they fear the judgment of the group. The more easier way is to assume that lots of existing ressouces are available and the aim of the wiki is to collect these URL. A single operating in this wiki is a cheap task. Cheap means, the the amount of time which is invested is less than 1 minute. This is the time it takes until a user copy&pasted a link into the wiki and pressed the submit button.
In case of doubt a chatbot is able to post the URLs to a wiki. The bot takes as input stream an RSS feed which contains of:
Date, URL, title
and creates for each item in the feed a new Wiki article. If the RSS feed contains of 100 items, this results into 100 newly created articles. Each of them is ultrashort and contains only of the URL plus the title. Such an automatic autoposting bot is able to increase the article count in the AI wiki from now 89 to 200 and more without much effort.
The prediction is, that after the article counter has increased some new users gets interested into the project and post their URLs in response to the links which are already in the system. A social network has the natural tendency to become attractive to large variety of users, especially users who are not motivated to contribute content but only want to have a bit fun while interacting with other users.
Wiki vs content management system
Wikis are sometimes described as the ultimate content management system. But let us go a step backward, which kind of CMS systems are already available apart from Wikis? Their number is huge. There are blogs, forums, pdf repositories, video platforms, static HTML content, powerpoint presentations, printed books, landing pages of academic journals and even diskmags are available. In all of the content systems fulltext is stored. Why do we need additional an wiki which can hold more content? No we don't. The classical Content management systems are working great. A classical ftp server which contains of pdf papers is a great choice as a long term repository for serious information. There is no need to convert the pdf papers into the wiki syntax.
What is not available right now and which can be served by wikis very well is a peer review system. That is a hub in which a group of people judge about the content and compare the content to each other. Such a system can't be realized within a blog nor with an FTP server. The underlying needed structure is called a social network. And the technical side is called a social network software.
No comments:
Post a Comment