Protocol Wars: The History and Potential Future of Decentralized Social Networks
How decentralized social media is being shaped, and what the past tells us about the future.
“History is a relay of revolutions.” – Saul Alinsky
With Elon Musk’s recent acquisition of Twitter in October, the world is watching how the social media giant is changing overnight by going private. With massive layoffs and a planned change of direction, it is clear that Elon’s intentions long-term are going to shake things up at the company, and may completely change how Twitter is utilized. Potential alternatives to Twitter became the hot new topic online, with Mastodon, Hive, and Farcaster receiving more attention.
The push for something new is clear. Nearly all microblogging alternatives place an emphasis on decentralization, utilization of open-source protocols, interoperability, and getting rid of advertising. On-chain communication and web5-based protocols have also taken center stage, along with Ethereum-based solutions like Lens Protocol.
Spatial Awareness is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.
So, will the solution that society gravitates towards be web3 or web5 based? Will cryptocurrency inevitably be a part of the next big microblogging platform? Or, will we meet somewhere in the middle, and adopt platforms that are free, open-source, and actively employing protocols like ActivityPub? What is the fediverse, and why does the history of the push for decentralized social media over the last decade help frame the conversation as we move forward?
Let’s make sense of the complexity and find a signal in the noise. We’ll briefly touch on the history of Twitter, dig into earlier attempts at decentralized social networks, compare Ethereum-based social media solutions to DID-centric ones and finish things off with X, Musk’s proposed ‘everything’ app.
The Birth of the Blue Bird
By the time Twitter was invented by Jack Dorsey and other founders at Odeo in 2006, blogging was already incredibly popular online. Microblogging (initially known as tumblelogs) is, by definition, blogging that is limited to short posts. The origin of Twitter (like most other social media apps that gained dominance in the mid-2000’s) can be traced back to a brainstorming session between undergraduate students attending various prestigious universities in the United States. The initial intent of Twitter (briefly called twttr) was to be a simple SMS-based service to share status updates between friends. But, like every major successful social media network, the choice of how to grow quickly grew out of the founder’s hands once popularity and outside influence took hold.
Since 2006, Twitter has gone from being an app used internally at Odeo to post jokes to a worldwide company that services over 330 million monthly active users. Twitter has become an important part of daily life for many people, and has various purposes depending on who you ask. Like most social networks, Twitter is used for connecting people and sharing ideas. How do such simple ideas become so nuanced and political?
Decentralization and Data Ownership
If you stuck to reading only recently released articles and headlines, you might presume that decentralization and data ownership for users was something that has only recently gained traction online, but that simply isn’t the case. Data privacy, open-source, and Linux advocates were some of the first to sound the alarm regarding the usage of Facebook, and by 2010 there was a vocal minority who wanted to forge a new path for social media usage.
Here’s what people are finding principally wrong about modern (web2) social media:
Ownership is centralized. Central servers owned by an organization maintain and track data, and sell that data to other people. You give away valuable information in exchange for using the network (often allowing for the network to operate without a fee for usage).
Identity. We need a better way to verify identity, and we want alternatives available to operate exclusively under aliases.
Money. Advertising within a social network changes how people operate within it, and there are visible and hidden economic incentives to posting and creating content online. Value can and should be exchanged online, but it ideally shouldn’t be the core focus of the application.
Disinformation. The spread of misinformation is a real problem, with convincing material being nearly indistinguishable from the real deal. Advancements in AI have made content increasingly easier to generate, leading to an endless stream of potentially damaging material.
At the heart of all of this is the true lack of understanding and transparency surrounding the usage and influence of algorithms. What started out as something that is helpful (creating algorithms to help users find the content they’re looking for) quickly can devolve into doom-scrolling, attentional issues, and proverbial echo-chambers of content. These problems are common among the most popular social media applications today, but a sea change is on the horizon. People are waking up to the reality of how programming can program you, and can have profound implications on how you process information and engage both online and off.
The things that are clearly wrong with the web are largely moralistic and are undeniably tied to the production of capital. Think about how the usage and further employment of algorithms is directly tied to the production of capital. This becomes exemplified to a greater degree when talking about web3, which is an ultra-hyperactive version of our already hyperactive capitalistic global economy. If web3 is the future, will we not find ourselves back where we started? Will web3 lead to the same algorithm-centric dystopia?
Web3 takes a stab at ownership, identity, money, and disinformation. But before we get there, we need to talk about the middle way: distributed social networks that are not cryptocurrency-centric.
Distributed Social Networks
User-owned, distributed (federated) social networks are not a new idea. Usenet was a worldwide distributed discussion system, established in 1980, which served as a precursor to modern internet forums. By 1983 there were over 500 hosts for Usenet all over the United States, doubling in size by 1984. Usenet was utilized by early internet proponents in the same ways that Discord and Twitter are used today, existing as the internet townsquare by the late 80’s through the mid 90’s. The most notable projects being created at the time were announced first on Usenet, with Linux, Mosaic browser, and the World Wide Web taking center stage. The first notable Usenet email list for LGBT activism was introduced in the early 90’s on the platform as well. Many of the core ideologies surrounding the usage of Usenet are strongly correlated with GNU and copyleft, with freedom to use software and an individual’s privacy being a top concern.
Once the world wide web took off, it was only a matter of time before emergent centralization due to economies of scale would take place, leading us to the early 2010s. The idealization and preference for distributed/federated social networks has been floating around the web for over a decade now, and the same proponents of open standards, EFF, and W3C are often pushing for the embrace of interoperability and federated communication.
Diaspora is the best example of the first modern push towards a user-owned, distributed social network. It was an idea spawned from four students (most notably Illya Zhitomirskiy, Max Salzberg and Daniel Grippi) listening to a speech by Eben Moglen in 2010 regarding “Freedom in the Cloud”. The founders of Diaspora wanted to address the core issues of Facebook, namely privacy, by creating their own social media network. The network consists of nodes, called pods, that are hosted by various institutions and individuals. Each node runs a separate copy of the software, and accounts are restricted to usage within a pod. Importantly, users maintain ownership of their data, allowing them to download all text and images they’ve contributed at any point. Diaspora isn’t hosted in any single place, and it isn’t controlled by anyone.
Diaspora is still a project, and you can host a pod or join one. But, despite the backlash in the early 2010’s for dwindling levels of privacy within social media, the world did not embrace Diaspora. It is free, and embraces decentralization, privacy, and data-ownership. Instagram, Twitter, Facebook, and eventually TikTok would continue to evolve with little interest in embracing the ideologies that made Diaspora stand out. Money ended up influencing everything, and privacy wasn’t enough of a concern for the mass majority of people using social media. This largely remains the case today.
With Twitter’s future uncertain, the world has turned its attention to Mastodon, the only federated social network where messages used to be known as ‘toots’.
Today, Mastodon is the most popular Twitter alternative, and it is particularly popular in China because of increased censorship in the mainland. Because Mastodon moderation is isolated to each node, groups that otherwise would be unable to operate on Twitter or Facebook have chosen Mastodon as their base of operations.
Mastodon is free and open-source software, and as a result the code has been forked and used for controversial projects, most notably for alt-tech. Gab, a far-right social network, changed its software to a fork of Mastodon and quickly became the largest node in the network in July 2019. This was also the case with Truth Social, created by Trump Media & Technology Group. In both cases, far-right projects forked the code and capitalized on Mastodon’s usage of an AGLv3 license.
All of the servers that run Mastodon or are forks of the original software utilize ActivityPub and operate within the fediverse network. Why is this important?
What is the Fediverse and ActivityPub?
The fediverse is an ensemble of federated (interconnected) servers that are used for web publishing (i.e. social networking, microblogging, blogging, or websites) and file hosting, but which, while independently hosted, can communicate with each other.
Many of the initial ideas for a federated social network were brought to fruition by GNU advocates. Starting in 2008, multiple open standards were created with the intention of allowing users on one website to send and receive status updates with users on other sites. The first open standard that was widely adopted was Ostatus, with several GNU social instances embracing this standard. In 2018, W3C, the international standards organization for the web, presented ActivityPub, which improves upon Ostatus and has since become the dominant protocol used between fediverse compliant servers.
ActivityPub works as intended, and is considered a success by those active within the GNU social community. With GNU social compliant software like Mastodon massively growing in popularity by the day, one might wonder how else social media networks can evolve.
Concurrently, there are two other strains of these ideas being pushed online, Bitcoin (or web5-based) and Ethereum (or web3-based) social protocols. If software can already be FOSS, why do we need cryptocurrency added into the mix? Is sufficient decentralization possible? As Varun Srinivasan (former director of Coinbase) points out, “ActivityPub doesn’t provide a way to use a managed host without compromising decentralization.” Smart contracts can make decentralized registries a possibility in a way that isn’t currently possible with the most popular GNU-social-based protocols.
So, let’s take things to the next level. Let’s talk about sufficiently decentralized social media networks that utilize smart-contracts. To start, let’s turn our attention to how Ethereum projects are approaching the problem of decentralizing social media.
Ethereum Social Networks
Because of the readership’s presumed familiarity, I won’t dive into the specifics about the nature of Ethereum’s decentralization. Ethereum is decentralized and open-source, and trails Bitcoin for being the most popular public blockchain. While Ethereum can be utilized for DeFi and is the most popular chain for trading and utilizing NFTs, it also has the potential of being the next hub for social media. Like ActivityPub distributed social networks, Ethereum is censorship resistant, immutable, and has payments built into its design.
Since Ethereum has become more popular, there have been a number of Ethereum-enabled social networks created for microblogging, as well as LinkedIn/Facebook clones. Using tokenization to incentivize behavior is a theme consistent with crypto-centric decentralized social networks.
To start, here’s two notable earlier Ethereum-enabled social networks, Peepeth and Minds.
Peepeth: A decentralized Twitter clone that utilizes IPFS, an important tool that powers the distributed web.
Minds: An open-source social network founded in 2011 that uses an ERC-20 token (MINDS) to reward users for creating content, referring friends, or providing liquidity. Minds utilizes Arweave for permanent backups, and hosts MindsChat, an E2E encrypted messaging application similar to Signal.
However, I really think it is worth giving Lens Protocol and Farcaster more attention, so let’s check them out.
Lens Protocol is a web3 social graph that runs off of Polygon’s PoS blockchain (EVM-compatible). The idea of Lens is to empower creators to “own the links between themselves and their community, forming a fully composable, user-owned social graph”. The protocol is modular, and seeks to solve some of the major issues with current web2 networks.
Because Lens was built from the ground-up with interoperability in mind, any application can be built to plug into Lens. It is community-owned, and puts the decision about how the social graph will be built into your hands, not a corporation.
The concept of an open graph is not new. Facebook’s Graph API allows websites to get information about people, events, pages, and how they are all related. The concept of a social graph extends beyond the connection of people, and includes everything people do and how they are related, etc.
Lens builds upon many of the ideas that were initially fleshed out by Facebook, taking everything a step further by allowing posts, follows, comments, profiles, and governance to reside on-chain in the form of NFTs and on-chain digital footprints. With built-in on-chain governance, voting or building a DAO is as simple as deploying a contract.
Essentially, Lens is similar to Diaspora or Mastodon, but instead of utilizing ActivityPub, it has created its own protocol. This is very similar to the approach that Farcaster is taking.
The team at Farcaster intends to build a sufficiently decentralized protocol that is an open social network that scales. On the surface, Farcaster looks like Twitter and other Twitter-clones. Funding for Farcaster has been led by a16z, and is seeing active support by many proponents of Ethereum-based social networks.
You’ll see the same pattern here again; the next-iteration of social networking have the same things in common: interoperability and protocols that are neutral and vetted by the communities they are meant to serve. Adding your Ethereum address adds the ability to show off your NFTs, and there are plans to tie-in visible interests on the network based around what you can ‘prove’ on-chain. The username registry is also smart-contract based, and is designed to be fully decentralized. Data on Farcaster is cryptographically signed, but stored off-chain in user-controlled servers that are called Farcaster Hubs. Although having all data stored on either an L1 or L2 blockchain sounds ideal, it would be expensive and difficult to maintain in practice. The failures of version 1 of Farcaster have led it to improvements that may make it primed for mass adoption as version 2 begins to roll out.
There have been a lot of questions about how Farcaster stores data within Hubs. Hubs are always-on servers that validate, store, and replicate signed messages. As Farcaster’s Github goes on to explain, “Conceptually, Hubs form an L2 network for storing social data, though the network has different properties from blockchain-based L2s. Its consensus model has weaker consistency guarantees but stronger scalability guarantees because the network data is shardable down to the FID level.”
Each account within Farcaster has an ID, called an FID (Farcaster ID). FIDs are secure, decentralized identifiers similar to UUIDs (universally unique identifiers). As we begin to discuss web5 (Bitcoin + social networks), you’ll note that UUIDs and FIDs are similar conceptually to DIDs, as they do not require a centralized registration authority.
Twitter users will find Farcaster’s UI to be familiar and intuitive. There are so many incredible things about Farcaster, but the potential for having developers build apps that support the Farcaster network may be the most exciting. A list of apps that can do things using the Farcaster protocol is available here. If you can imagine a tool or app similar to anything that’s been done in the past (Instagram, Tik-Tok clone, you name it), it is likely possible with the current setup in Farcaster.
When Dan Romero and team set out to build Farcaster, they conceived of the idea of RSS+. RSS+ is an enhancement to RSS, giving it a social graph along with social actions. Going into more details about RSS+ would make this article even longer, so if you’re interested in Dan’s ideas around how RSS+ may work moving forward, check it out here.
With all of this talk around Ethereum-based solutions, one might wonder if Jack Dorsey’s web5 initiative would have any movement surrounding decentralizing social media. Let’s turn to how technologists are thinking of utilizing the consensus layer of Bitcoin to create a social network.
Bitcoin’s Turn: Bluesky, AT Protocol, and web5
For starters, we need to define web5. Web5 is a decentralized, peer-to-peer network (using Bitcoin as a consensus layer) that aims to create an internet free from centralization and censorship. The network is based on web5 nodes, essentially decentralized data storage and message relaying mechanisms. Important concepts within web5 infrastructure are DIDs (decentralized identifiers) and DWNs (decentralized web nodes).
For a refresher (or first time crash course) on how DIDs and DWNs work and why they matter, I encourage you to check out my previously published article here.
Web5 primitives/building-blocks are supported by W3C, but where does social networking come into play?
Bluesky is an initiative started in 2019 by Jack Dorsey and his team to develop a decentralized social network. He anticipated the trends that are gaining traction now, and because Bluesky was created as a nonprofit organization by the team developing it, it is not impacted by Elon Musk’s recent acquisition of Twitter.
Most recently, the Bluesky team announced it is building out AT Protocol, which is essentially a new foundation that incorporates the usage of DIDs and DWNs. If you take a look at AT Protocol, you will again see the same decentralized social media points emphasized. Federated social networks, with portable accounts, sufficiently decentralized, with algorithmic choice. While Bluesky is not yet available, AT Protocol’s developer documentation is online here and you’ll find many similarities when compared to Farcaster and Lens Protocol.
We’ve seen distributed social networks, Ethereum and Bitcoin-based social networking solutions, and the various protocols associated with each. But what about ‘X’? How does Elon Musk’s hypothetical Twitter 2.0 compare to everything else we’ve listed above?
X (Twitter 2.0): The Everything App
X is Elon Musk’s dream app. Elon wants to get away from the United States government moderating and controlling content. He shares the same aims as many other multi-polarity advocates for creating a world where the United States is not the primary dominant superpower. Elon has a love for WeChat, and views it as a combination of Twitter, plus Paypal, plus ‘everything else’.
With Musk controlling Twitter, he has stated that the platform will eventually merge with the everything app, X. Musk recently re-purchased X.com, having used the domain in the past as a hub for a financial services start-up that eventually merged with PayPal in 2000. The future app X may look similar to Zion and Bluesky, with Bitcoin being used as the base consensus layer for both on-chain messages and financial transactions.
Musk has also been public about his desire for converting Twitter to a blockchain-based messaging service similar to the projects and protocols we’ve mentioned above. To that point, it is unclear how X would function or what protocols it would utilize, but the focus would likely be less on decentralization on more on implementing a blockchain-based messaging service where users make micropayments with cryptocurrency to post, like, and interact. As Twitter evolves over the coming months, we will see how it transforms, and if Twitter will indeed become X. We’re being taken along for the rollercoaster ride regardless of what we think.
Social Networks Moving Forward
There appears to be a lot of options for the potential future of decentralized social networks. We’ve compared Diaspora, Mastodon, Lens Protocol, Farcaster, Bluesky, and X, and learned about the history of distributed networks and decentralization, ActivityPub, the fediverse, and the various protocols that govern Ethereum and Bitcoin-based projects. These projects are the ones to watch as the future of the social web evolves.
While blockchain technology allows for sufficiently decentralized networks in a fashion that is stronger than without the technology, it remains unclear if pros will outweigh the cons long-term. Unlike many other technology stacks, blockchain comes with baggage, and many consumers will outright reject applications if they make any mention of Bitcoin or cryptocurrency in general. Ethereum advocates are convinced that Ethereum will be the future of the social web, and web5 enthusiasts remain confident that W3C’s embrace of DIDs and DWNs will make or break the future of the web. As with many other protocol and format wars, the embrace of these technologies may boil down once again to funding and big names behind the solutions.
If we see a scalable decentralized social network that is interoperable and widely adopted by developers, that may be the solution the world goes with. Such a solution may look something like one of these projects mentioned, or it may be created in a few years as a spiritual successor to something like Farcaster or Mastodon. If the world can create novel solutions for sufficiently decentralized social networks that solve the hard problems of authenticity, bots, fake news, and moderating content, we’ll be in a good spot.
So, where do you think our future is taking us? Let us know in the comments below, and be sure to check back in next week for a discussion on the future of GPT3-based writing generation.
Need help with your web3 go-to-market strategy? Eager to dive into web3 and NFTs, but not sure how to approach it? Reach out to immutablelabs.io and schedule a chat with one of our experts!
Spatial Awareness is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.