A Comparative Analysis of the Decentralized Social Protocols Nostr and Farcaster

Original Title: 《The moral imperative of a distributed social layer》

Author: pourteaux

Compilation: The Way of DeFi

For a long time, I have criticized the view that Twitter is a "town square," where moderation is called "censorship," and users have the right to be members of the platform and become a controlled audience. Supporters of this position almost see Twitter as a practical tool and believe that limiting the platform's activities is, in fact, an unconstitutional restriction on speech. Surprisingly, this is the perspective of modern "new wave" libertarianism, whose proponents claim that Twitter has a monopoly due to its insurmountable network effects and therefore does not deserve the typical protections afforded to private companies.

Private companies have never been town squares, but protocols can be.

I align with the views of old-school libertarians: Twitter is a private company that sets its own terms of service, and the government should not make operational decisions on behalf of Twitter. No one should have the right to access the services of a private company, and there exist unlimited venues for free speech without demanding anything from a specific platform's audience. Clearly, platforming hate speech is detrimental to Twitter's business, as they would lose advertisers. The idea that you cannot compete with Twitter is absurd; it has many social network competitors, such as Facebook and TikTok, and it also has direct competitors like Gab, Mastodon, Truth Social, and others.

New wave libertarians argue that those banned from Twitter "cannot just start their own social network," but this is exactly what the former president did after being banned on January 6th with Truth Social. Moreover, many who claim "censorship" actually have significant platforms and often use the notion of "censorship" as a populist growth tactic.

As a private company, Twitter not only has the right to moderate speech in ways they see fit, but a carefully curated platform can make Twitter a better user experience and a more profitable business. Almost no one wants 4chan content on their Twitter timeline. Musk did us all a favor when he took over Twitter; he quickly proved to the right that the man who promised to "free Twitter" moderates speech just as much, if not more, than the previous leadership. He does it more impulsively, with fewer procedures, and out of a center-right bias, but the key is that he did it. An unmoderated Twitter has always been untenable; a private company cannot create a truly unfettered speech platform. When they try, the platform becomes a cesspool that almost everyone ignores, with no business model, just like 4chan.

Given this, you might wonder—why do I advocate for an open, permissionless, and censorship-resistant social network?

The new social layer allows us to cater to those who want a completely open "town square" and those who prioritize a carefully managed and curated experience. These two perspectives may seem opposing, but they are actually completely consistent. In fact, I do not believe there is any disagreement between the groups represented by these two statements; they are simply noticing different parts of the dysfunction of modern social networks. Over the past few years, private companies have struggled to be defenders of free speech while also trying to be places with basic etiquette rules, but both efforts have failed.

You cannot have both. Government, advertisers, employees, and public pressure will force private companies to limit speech. Moreover, any moderation action may be abused in the future and could deprive many of their rights.

Private companies cannot be public squares, but protocols can.

Protocols like Nostr or Farcaster can be permissionless and censorship-resistant because, in a distributed network, no one can send a takedown request. Of course, this is an imperfect solution, because if you stop here, all you get is a distributed 4chan, which hardly anyone would be interested in. You need a second-layer private company and clients to act as curators and moderators. Users can choose to join and pay for this moderation layer to provide a better experience. Depending on your worldview, you can pay for curation from The New York Times or The Joe Rogan Experience at the second layer. Twitter itself could be a content curator from Nostr or Farcaster (or both) and provide ads alongside the content to keep the platform free. All these different L2 operators' moderation decisions may overlap somewhat, but there will also be many non-overlapping moderation decisions. Therefore, no one is at risk of being "de-platformed" unless you accomplish the impossible task of being banned simultaneously by all different L2 moderators. Even so, for those who can endure it, such a person's posts can still be seen on the unmoderated L1.

Thus, the "social layer" addresses the needs of both sides of the political spectrum: a social platform that does not have to worry about "censorship" and "de-platforming," while also allowing users to choose to join curated experiences by moderators they trust!

The social layer allows for all the benefits of permissionless and censorship-resistant L1, along with all the benefits of optional curated experiences in L2.

Let’s delve into the two attempts at L1: Nostr and Farcaster. I am pleased to use both simultaneously and will provide my non-technical experiences with both, as well as the potential and limitations I see. I also want to thank fiatjaf (who built Nostr) for helping me understand Nostr, and Dan Romero (the founder of Farcaster) for inviting me to join Farcaster. I also want to thank Maciek Laskus, who provided thorough analysis on every question and discussed the topic with me. Let’s start with a table:

Farcaster and Nostr both aim to decentralize communication but make different trade-offs in the process.

Nostr and Farcaster both use relays (called "hubs" in Farcaster) to store social data (like posts) and transmit it to users (the software users run is called "clients"). Farcaster uses Ethereum architecture (for example, accounts use Ethereum key pairs), while Nostr does not use a blockchain (although you see many Lightning Network transactions, it could theoretically support any blockchain).

It is worth noting that Farcaster is a VC-backed company with a well-known team and an excellent product, while Nostr is not a company at all; it was built by an anonymous person without funding (though it later received a grant of 14 BTC from Jack Dorsey) and is clearly in an earlier stage of development. In conversations with fiatjaf, I confirmed that the only funding for Nostr is this Bitcoin grant provided by Jack, and fiatjaf subsequently parted ways with jb55, who is developing the Damus client on iOS and macOS. Both Farcaster and Nostr have already built clients and services on top of them (examples: Nostr, Farcaster), although Farcaster seems to have progressed further in this process. In fact, some design choices made by Farcaster make development easier, while the simplicity of Nostr is a trade-off between recent difficult development and more potential future use cases.

For example, Farcaster currently requires complete synchronization of all hubs in the network, which makes client development easier. In contrast, Nostr allows relays to host or delete any content, making client development more challenging in the short term but allowing for many potential business models for these relays and more open potential use cases. Relays on Nostr can then charge subscription fees or run ads or otherwise choose what content to include and exclude.

Farcaster has already prepared a beautiful product for ordinary users, which is as easy to use as Twitter and may even be better.

Nostr's current user base is primarily Bitcoin users, while Farcaster's user base is mainly Ethereum users and tech entrepreneurs. In many ways, this is also reflected in the design choices of each platform. Like Bitcoin, Nostr prioritizes simplicity, neutrality, and the convenience of running your own relay (like nodes in Bitcoin). Like Bitcoin, Nostr avoids venture capital, forming companies, or having well-known founders. Running a hub on Farcaster is more difficult and expensive, but Farcaster asserts that certain design trade-offs still allow for "sufficient decentralization" while prioritizing more recent use cases and optimizing products that are almost ready for retail users.

I have accounts on both platforms (Nostr at @pourteaux.com, Farcaster at @px). Simply put, Nostr seems to follow the spirit of cypherpunks and Bitcoin, where simplicity and robust protocols are paramount. I am impressed by fiatjaf, jb55, and other contributors. Farcaster follows a more traditional tech startup model, with their product polished and looking delightful. Both efforts aim to create a native social layer for the internet: I believe this is a moral imperative. So much social dysfunction is a result of our current corporate social layer trying to provide everything for everyone. The layered approach to social media combines the cypherpunk principles of L1 with the importance of institutions in L2 to serve social experiences that meet everyone's needs.