Welcome to our new blog series exploring some of the research we've done internally on decentralized social media systems.
Social media users are not sovereign: they do not own their content, data, identities and even devices. Big tech oligopolies control the hardware, the software and the online platforms used by the vast majority of users. In exchange for their access, users are presented with draconian Terms of Service. These are abused to censor users, erase content, push propaganda, silence dissenting voices and extract the value that users create by posting and participating. Therefore, millions are looking for alternatives.
At this very point in time, we are witnessing the adoption of decentralized social media. These protocols can offer unprecedented sovereignty and features to social media users: identity portability, censorship resistance, choice of content policies and choice and transparency of algorithms as well as native value transfer and fair attribution of value to those who create the content.
At 'embed, we are honored to be a part of this next paradigm shift in human communication. Our mission is to make decentralized social media usable by democratizing algorithm creation and bringing the value flows going through those algorithms on-chain. Our recommendation system already builds the feeds of tens of thousands of users, and we continue to improve it through our research and development, exploring topics like decentralized moderation and spam. Furthermore, we also know a high-performance model alone is not decentralized, so we are researching how to make it open, credibly neutral and incentive-aligned, sharing revenue with users whose data contributes to its success.
Besides recommender systems we’re researching what we see as two of today’s biggest decentralized social media problems: insufficient moderation and rampant spam. We have done and continue to do research on credibly neutral moderation protocols, where users can be confident that the moderation policies they choose classify content as advertised. We are also exploring fact-checking protocols, and we’re building our own social-media spam classifier.
This research is what we’ll be discussing in this series of blog posts. We welcome the help and participation of the community, and invite you to join us in our passion for connecting humans in truthful, constructive and rights-respecting ways.
In our next post, we will analyze the Birdwatch model, which underpins Community Notes, and explain why it is not suitable for a decentralized setting. We will also introduce the concept of Social Epistemology, which permeates the design and analysis of decentralized social media protocols.
Reach out
If you are building and researching decentralized social media feel free to reach out to us on [email protected] or on X.