Back to selection
Speculations
by Joanne McNeil
All Protocols are Not the Same: Joanne McNeil on Bluesky
If you are a filmmaker, journalist, artist or someone else who creates work for an audience, most likely you’ve used social media to share it. Self-promotion is never fun, but it’s felt especially inconsistent (if not totally hopeless) ever since Elon Musk acquired Twitter, renamed it X and began throttling views to posts that include links. A number of users have escaped to new broadcast-based social media services that are “decentralized”—but what that word means depends on who you talk to. There are concrete differences between the AT Protocol, underpinning Bluesky, and ActivityPub, the protocol and standard for Mastodon as well as Threads, that shape your experience as a user.
To understand those differences, a brief history of how social media happened helps. In 1989, Tim Berners-Lee announced his idea for a hypertext service standardizing a way to connect information on computers across the globe. He named this system the “World Wide Web” in 1990, and it opened to the public in 1993. Independently developed web browsers like Mosaic and Netscape made it easy to explore simple web pages written in HTML. At first, people tended to use websites for static content: Think of those gangly “Hi, my name is Jim and this is my personal Web Site” Angelfire pages and blocky layouts that media companies like the New York Times and CNN created when they premiered on the web. Online communities in the 1990s, such as forums and chatrooms, tended to be siloed inside proprietary services like AOL and Comcast or BBSs. You can imagine just how frustrating it was for users to be arbitrarily separated from one another: It would be as if you could only message people on Verizon if you had Verizon cell phone service. But, as a common medium for anyone with an internet connection, the web presented a way for users to connect beyond the limits of what their ISPs provided.
In the 2000s, the “Web 2.0” era saw an explosion of online communities and content creation through blogs and social media; anyone with an internet connection and browser could navigate to the websites for Facebook and Twitter. Later in the decade, investors wary of real estate and banking in the wake of the Great Recession funded tech startups, including social media startups, resulting in user bases at a scale no one on the internet had ever experienced before. All of this happened because the web exists. It was and still is interoperable, noncommercial and “decentralized.”
Shortly after Berners-Lee created the web, he also founded its governing organization, the World Wide Web Consortium, or “W3C.” The organization is still going strong, still making choices that are why, when you look at a website in a Safari browser, it will deliver the same content to you as it will to someone viewing it with a Chrome, Mozilla or Vivaldi browser. In 2018, ActivityPub emerged from one of the working groups at the organization, led by developers Evan Prodromou and Christine Lemmer-Webber.
ActivityPub provides a common protocol for social media organizations. It is structured in direct contrast to the closed nature of the prevailing social media giants. You can’t read LinkedIn posts on Instagram or follow a Twitter user through Facebook, but if a social media organization uses ActivityPub, cross-platform activity is possible with other organizations in the network. It’s also possible for administrators of these social networks to block other social networks or choose not to “federate” with other networks. These networks range from communities of a handful of people to massive platforms like Threads, run by Meta, which is on ActivityPub in the same way that the Facebook website is on the web. Just as you might have separate personal and work email addresses, you can participate in multiple ActivityPub-based communities and sign up for and leave these accounts depending on your preferences.
Broadly speaking, ActivityPub decentralizes your posts and generates permission for content you create to be sent and received across the entire “Fediverse” of servers using this protocol. The AT Protocol created for Bluesky, on the other hand, approaches decentralization on the level of individual users. You have a singular identity across its network, which is highly customizable. If you leave one server and join another, you take with you all your data, including followers and an archive of posts.
If you are familiar with the blockchain, identity on the AT Protocol works similar to a crypto wallet. It’s no coincidence: Bluesky began as a research project at Twitter spearheaded by Jack Dorsey, the co-founder and longtime CEO. He also co-founded the fintech startup Block. In 2021, Dorsey said of Bitcoin, “I don’t think there is anything more important in my lifetime to work on.” That year, Bluesky launched as an independent company, with Jay Graber as CEO. Graber and other team members have backgrounds in blockchain development.
At the time of Bluesky’s founding, “Web3” was trending. It’s the idea that the blockchain framework can be applied for general purposes beyond cryptocurrency. Thought leaders in the Web3 space like to compare their work to the founding of the web and insist the future they propose is “decentralized,” in contrast with Big Tech. Bluesky could be considered the most successful Web3 project to date. However, the organization is quick to tease out the very fine distinction between what a blockchain is and how it’s operationally different from what they are developing. In a recent press release, the company insisted that “we will not hyperfinancialize the social experience (through tokens, crypto trading, NFTs, etc.)” For what it’s worth, that press release also announced they’d raised $15 million in Series A funding from Blockchain Capital.
More than its Web3-esque elements, what worries me about Bluesky is that it is iterating on a model that has, time and time again, failed users: social media at scale. Global-scale platforms like Facebook and Twitter foment problems both top-down (shareholders calling the shots) and bottom-up (users who share gruesome content like beheading videos that must be scrubbed by human content moderators). Not to mention that, at present, the AT Protocol system of record-keeping and data storage is massively centralized. As Lemmer-Webber explained on her blog, DustyCloud.org, it would be enormously expensive for an individual to run their own “relay,” or capture and transmission of data, across the AT Protocol. If Bluesky were like a physical mail delivery service, she writes, the way it operates is like dumping all the mail at the post office. The problem isn’t only receiving this massive trove, Lemmer-Webber says, but also “it’s the responsibility of interested parties to show up and filter through the mail to see what’s interesting to them.” These are serious technical challenges to surmount, and when you throw venture capital in the mix, user-focused solutions seem unlikely.
The decline of X is an opportunity for us all to rethink social media and how we participate within it. Is “following” and “unfollowing” a necessary behavior when people gather online? Is the social graph a useful way to orient yourself in the digital world? Is our personal data something we should own like property, or should we have rights over it and to it like human rights? Is there a way to integrate in-person connections, or is a social network exclusive to networks? What is content anyway, and why should we post it or look at it?
What we know from experience is that organizations that seem small and scrappy at one moment can come to dominate aspects of our personal and professional life seemingly overnight. The World Wide Web itself might have had a heel turn—it didn’t, but potential for catastrophe was there 30 years back. We also know that organizations in the business of developing online communities depend on the users who populate these communities. A better internet is possible, but what remains uncertain is how to get there.
