I want to develop a search engine algorithm that, instead of reinforcing our confirmation bias, counteracts it by using our preferences to select search results that broaden our perspective. A search engine that displays results that make us surprised, curious, sometimes perhaps confused but always less single-minded. Or a recommendation system that presents: This is what others, who are nothing like you, found interesting. Or an antiviral social media platform that is based on the principle of highlighting all the different nuances in a survey material, rather than weighting the nuances fairly according to how common they were. All added material goes through a classification based on similarities with other material and automatically forms categories. The aim of the network is then to expose users to as much variety of material as possible. This means that as many of the different categories are displayed and as few objects from the same categories as possible. You cannot follow or be followed. You can like posts, which is fed back to the poster but does not affect the future flow. Posts will not go viral through the feedback phenomenon and sharing. So what you see when you look into the network is not what is trending, nor what is most common (cats and babies doing something adorable or hilarious), but a balanced insight into what is happening on the platform.