Into the Plativerse… Through Fiddleware?

Richard Reisman is a nonresident senior fellow at Lincoln Network.

Poor performance at Meta and Twitter’s self-destruction remind us that the dustbin of online history is littered with once dominant platforms. Many see this as akin to the extinction of dinosaurs, to be replaced with smaller, more nimble mammals, like the Mastodon. Yet, as David Carroll and Alex Tarkowski each observe in Tech Policy Press, it is becoming apparent that neither highly centralized platforms nor the highly decentralized ‘fediverse’ can solve the problem of the scalable, participatory governance of digital town squares.

The sudden Musk-triggered rush of newbies from Twitter across the chasm to the “fediverse”of Mastodon is finding the decentralized alternative generally attractive, but perhaps not quite ready for prime time. The appeal of the fediverse is its distribution of control as a federation (or perhaps more accurately a confederation) of independently controlled server ‘instances.’ Improvements to this still-primitive federation seem likely to come quickly, but others involve more fundamental changes in mindset among its developers, and open questions about just how control should be distributed – for what aspects of functionality, to serve what objectives. The only thing clear is that an era of experimentation to find better structures for online discourse is badly needed.

The fediverse supports a live-and-let-live diversity of models. Mastodon instances generally use variations of the same open-source code, each having their own user base and moderation functions, much like a small social media platform, but with the interoperability to bridge posting globally to users on other instances. This is based on the open ActivityPub protocol, which functions much like an email transfer protocol. The great appeal of the fediverse is its openness and lack of central control, but that is also a source of weakness.

There are already moves toward improving the fediverse with better user experience, connectivity, and search. But how will instance administrators sustain themselves financially?Ad Age reports that advertisers are already wondering whether this is a medium for them, and concluding not quite yet. While many view advertising as a problematic business model, Mastodon is open source and its license seems not to preclude forks that support ads or other revenue model integrations. Given challenges with the sustainability of the current model of volunteer labor and funding, if entrepreneurs build it, and a sufficiently connected audience is attracted, advertisers will come. It would seem surprising if there are not already commercial Mastodon startups in stealth mode.

The Plativerse (A Fediverse That Includes Platforms)

A related transition step has already begun organically, in limited form, by applying adversarial interoperability. In the spirit of going where the users are, a fediverse that includes centralized platforms is emerging. As Mastodon grew, the market sensed a need to offer services that provide for cross-posting in either direction between the Mastodon fediverse and Twitter. Because these bridges are still crude, Twitter is effectively a huge instance (platstance?) that is poorly federated.

It seems inevitable that those beginnings of a hybrid fediverse/plativerse can be improved on to enable full interoperability between the fediverse and Twitter (or any other platform). Twitter and other platforms might block it, but the tide seems to be toward federating at least to some degree. The BlueSky project Jack Dorsey spun out of Twitter is pursuing similar goals, using a different protocol called AT. Why not an AT-ActivityPub bridge? Maybe only transitional, but some users might prefer the walled garden paternalism of a platform, while others would opt for more freedom and local control. Some might prefer to keep these communities cleanly separated, while some seem to want a level of interconnection. A mature hybrid might enable both.

The appeal of an open fediverse/plativerse model seems obvious, in spite of current limitations in function and ease of use. The challenge of sustainability will motivate commercial versions to fork off the current code base for revenue model experimentation – not only advertising, but subscriptions, sponsorships, donations, and public funding – and participatory pricing/donation innovations (such as FairPay). And of course there will be improvements in both local and federated user experience.

But as Carroll and Tarkowski each explain, perhaps the most fundamental problem with our current fediverse model is the “lack of participatory governance.” Now, instance operators are benevolent dictators when it comes to content moderation and other policies. As the fediverse scales and goes beyond skilled early-adopters, that will collide directly with the challenges of moderation at scale for service providers, and how hard it is for users to specify complex and nuanced preferences.

Carroll suggests this portends “an epic power struggle over who controls the means of social media production and distribution.” Similarly from Tarkoff: “Ultimately, what is at stake is our capacity, as societies, to control and manage planetary-scale communication spaces that are, in the broadest sense, democratic.” Chris Riley and I have explored this more broadly in a series of essays in Tech Policy Press, drawing on ideas of distributed control as a both/and framing that transcends the false binary of centralization or decentralization, and applies a principle of subsidiarity as suggested by Divya Siddarth, Danielle Allen, and E. Glen Weyl.

Fiddleware (Federated Middleware)

I have long advocated for user choice in how our online feeds are organized and moderated as the only effective way for a democratic society to deal with this complexity and nuance. Enabling such choice has recently gained advocates who see a role for “middleware” services that act as user-agents between users and their media distribution systems. I envision this not as choosing a single middleware service to be granted sole control, but as composing and steering combinations of services to blend a range of algorithms that distill selected sources of human judgments – and to use them to draw from a multiplicity of what I called confederated systems as far back as 2003.

The importance of that level of flexibility in middleware has been little recognized, but the fediverse/plativerse may provide just the environment for it to emerge organically. If users can be given powerful tools to manage their navigation of the fediverse, shouldn't they be able to shape these tools to feed them what they want, drawing from any of a multiplicity of instances/platstances, in whatever ways they choose – rather than being under the control of any single home instance with its home community and single benevolent dictator? Shouldn’t they be able to compose multiple ranking services to generate composite rankings, and shift the gears – weighting and steering those systems as their moods, tasks and domains change? Shouldn’t middleware be federated? Call it fiddleware.

The Bluesky initiative is already working toward that kind of flexibility. A Bluesky blog post announcing the AT protocol specifically refers to the objective of algorithmic choice: “Algorithms dictate what we see and who we can reach. We must have control over our algorithms if we're going to trust in our online spaces. The AT Protocol includes an open algorithms mode so users have more control over their experience.”

Shaping a Diverse Information Ecosystem

A fediverse offers a flexible, open market seedbed for such innovations to emerge and develop. If, as Steve Jobs said, technology now offers “bicycles for the mind,” then we need freedom to choose which kind of bicycle to use for a given journey – and freedom to steer the handlebars and shift the gears.

Federated, composable middleware can provide a community/institutional ecosystem layer between us and the underlying systems in a way that hides their complexity. Few users are able or willing to specify detailed criteria for ranking what should be fed to them. But over centuries humans have become skilled at selecting a diversity of communities and institutions to mediate their information ecosystem as suits their needs, values, and tastes. Extending those skills to selecting a diversity of social media media middleware services would be a natural evolution.

The fediverse is surging in reaction to the platforms’ abuse of our attention and failure to scale moderation well. But scalable participatory governance is the crucial failing of the fediverse as well as the platforms. A plativerse can allow platforms to interoperate with less centralized systems – and can create an open marketspace in which shared infrastructure services such as middleware can emerge and find their place organically. Federated middleware that augments individual choice with social support and guidance can apply appropriate degrees of subsidiarity – distributing control to the level where it is best managed: by the individual user.