I will make the case that DID:PLC forks will inevitably emerge as the wider atproto ecosystem keeps gaining global relevance: pressure on the governance model of the identity system will increase to a point where conflicts (e.g. over which DID suspension requests should be honored and which should be ignored) cannot be easilly resolved within a ‘permissioned consortium’ (as proposed in https://atproto.com/guides/identity) anymore. It seems evident that we can only argue about WHEN this will happen, not IF it will happen.
In order to be able to maintain coherent UI experiences (without threads looking more and more broken due to different forks being used in different appviews) it seems necessary to extend the adversarial design patterns at the heart of the bluesky project (‘the company is the future adversary’) to the governance model of the underlying DID:PLC identity system: ‘the consortium will become target of future adversaries’.
Can permissionless observatory networks help?
Can we avoid using cryptoeconomics?
What other options seem useful?
During this presentation, I will present the results of my ongoing PhD research project on these questions.
About me:
I'm part of a small team (including Marcus Sabadello, co-author and editor of the DID specification) that started to advocate for the adoption of DIDs within ActivityPub networks back in 2018 (predating Jay Graber’s ecosystem review and the birth of the bluesky project).
https://chaos.social/@cypherhippie/102270069807129706
Link to a (somewhat cringe) video for a grant application in 2018: https://youtu.be/UJn7cLNh_q8?t=85
Recent talk at fediday berlin:
Protocol Convergence within Open Science Communication Networks
https://fair.tube/w/p/2PEFZ5cdptsVASU4HTUakA
Upcoming presentation at fosdem:
Increasing Long Term Stability of Relations Between Fediverse Identities using SSI
https://fosdem.org/2026/schedule/event/RZRZ9P-increasing_long_term_stability_of_relations_between_fediverse_identities_using_s/
