After Elon Musk’s recent purchase of Twitter, many users are starting to explore alternative social networks such as the fediverse’s Twitter alternative, Mastodon. Those users are not necessarily enjoying a smooth transition.
First, there are plenty of good reasons for not wanting to switch away from existing platforms, such as the fact that Twitter has a huge established user base, so leaving means abandoning one’s network there.
Second, it’s burdensome to check too many apps regularly to get more or less the same style content. The amount of attention required to use a new app for each new social network will quickly grow unsustainable. But what if you only needed to use one app to keep track of all those different social networks?
Enter aggregation. Rather than opening one app for Twitter, another for Mastodon, and yet another for Reddit, what if you could view all three together? Furthermore, what if you could filter and prioritize your content from each platform, taking control of your own feed instead of depending on proprietary algorithms? This is what we are hoping to achieve with the next iteration of iDPI’s Gobo project.
The Gobo project began while iDPI head Ethan Zuckerman was still leading the Center for Civic Media in the MIT Media Lab. The original Gobo was essentially an interesting tech demo that allowed users to import their social media feeds from a few sites and view them through a set of filters with selectable parameters. It also provided explanations for which posts were hidden, which posts were shown, and why those filtering decisions were being made.
It was a good tool for exploring the possibilities of user control over social media feeds, but it was slow: there were significant delays before posts would appear in Gobo.
For the next generation of Gobo, we are going to expand on the concepts modeled in Gobo 1.0 by adding some fresh functionality, and making the app fast enough that users will want to make a habit out of using it.
We believe Gobo needs to allow users to:
- Read and post to multiple social networks from one open source client.
- Pick and choose between algorithms for filtering and sorting the posts from all these networks.
- Design (and potentially share) different algorithms for filtering and sorting.
- Use third-party services to assist in filtering and sorting.
- Audit the performance of these different algorithms as well as the third party services.
All of these elements deserve to be explored in depth in their own posts. For now, let’s focus on two that demonstrate a lot of what we are trying to achieve with the tool: third-party scoring services and lenses.
Third-party scoring services are sites that can receive social media posts and generate ratings for those posts based on the content therein. This could involve, for example, a machine learning model that scores images on the likelihood that they contain a dog, or it could perform fact checking, analyze news quality, or implement parental control.
The point is that these services will not be run by the Gobo software directly but would instead be providing an API that Gobo can call to retrieve these scores for registered users. As an initial demonstration of a third party scoring service, we are building Gobo Doggo which scores posts based on the likelihood that they contain various cute animals. We are hoping to work with others to construct additional services as the development of Gobo progresses.
A lens is a set of rules for filtering and sorting your social media feed. We are saying lens rather than just filter as the lenses are essentially native scoring functions and will be used for sorting rather than just filtering. A lens allows you to select which media sources should be included, which third party scoring services to use, and what keywords to promote or remove.
Users will be able to select several lenses that will be combined to create a news feed. For example, you could select from lenses such as “politics,” “no politics,” “cute animals,” “professional news.” One lens could aggregate all of your professional contacts, industry news, and other work related content while another could focus on your friends and family. We are working on building tools for constructing, editing, sharing, and auditing these lenses and they will be part of the core functionality for Gobo.
I hope to write additional posts in the next few weeks diving into more of our ideas for the new Gobo in detail. We are targeting a beta release in two to three months, and if you are interested in participating – and in keeping up with other goings on at iDPI – please sign up for our newsletter below!
Spencer Lane is a PhD student at UMass Amherst working with the Initiative for Digital Public Infrastructure, and currently Gobo’s lead developer.