walk — a system for exploring and finding content on the internet

walk — a system for exploring and finding content on the internet

Walk is a system that provides mechanisms to disperse the mainstream by providing individuals with new, diverse and tailored content when it is needed.


Social Media has in the past encouraged to follow trends, idealize influencers and push personal interests forward, especially by not being bound to moral standards and assigned to a real identity as in the offline world. However, this did not assist the dissemination of different ideas but instead reinforced a mainstream dominated by media-effective content. Ratings and likes act as alleged objective representations of worth, while they are in fact highly subjective. A user does not find the content that might suit them best when browsing the Web—instead, they will find the results, that are most popular among many people. Due to the pace of information exchange on the Internet, these trends spread even faster than they did years ago, creating an ever more comfortable platform for monopolies.

Walk aims to provide value within large amounts of unstructured data, while also enabling self-controlled exploration. It should do so without being a channel for judgement and harrasment like many social networks are today. To address these goals we developed three interlinked core concepts as groundwork for the system.

Humans as authorities for declaring qualitative content

An algorithm alone cannot decide what value is. It can filter for specific parameters of objective quality, but to judge the subjective quality, emotions or preferences humans—due to their unique characters—are more suited. Similar to social networks, a model of following and followers allows users to share and receive recommendations from people, the user connects with emotionally or people whose opinion they value. As trust is not unconditional, people are assigned to interests to decrease the amount of irrelevant content.

Alterable filter bubbles supporting the user’s intention and context

Usually, platforms either hardcode fixed parameters for topical closeness directly into their algorithms, or they try to assume the user’s intention using keywords and context. That, reversely, has the effect of recommendations seeming either random or are limiting. If the algorithm decides for the user to look only at content that is similar to what they usually favor, it takes away the opportunity for people to branch out of their comfort zones and limits the possibility of exploring different mind-sets.

Diversity of content through cross-platform linking

Our goal is not to restrict content providers, so instead of interfering with content on a website, a layer on top of it creates new connections among platforms. Neither are those thereby entirely managed by the website creators, nor are they organized by biased platform users. Instead, they are linked through the topics they address.