A searchable index for websites, that is structered and labeled by humans
We design a chatbot, that helps you find the website you are looking for. The chatlog, that helped the user find the correct website, will be interpreted as a label for that same website, and stored in an index, under that newly labeled category.
With more and more users of the chatbot, the MemX index will grow. As it gets bigger, it gets more useful. You can start using it as an alternative search engine. You would put in a very broad description of what you are looking for, as you can do with the google search engine. The MemX then gives you a selection of topics that are more concrete, but still very broad. You choose the one that seems closest to the result you are looking for. The process repeats itself, until you reach the final subcategory, that has the most descriptions of all the categories above. In that last category, there is a selection of all the websites that have been found by users of the chatbot, using the description of the category they are stored in now. Those few websites are now ranked by importance the same way google already does.
The MemX should help the user to explore the internet, much like one would explore a natural landscape. It organises websites after their actual content. It gives you an overview of what is going on in the world, of what is out there.