ATProto Service Patterns: Technorati, Omnifollow, and Infrastructure Abstractions

Two interesting service concepts surfaced in the ATProto community today that demonstrate the protocol's flexibility for creating novel social infrastructure.

The Technorati Pattern

Someone observed that standard.site aggregators reveal ATProto's true purpose: not building "new Twitter" but building "new Technorati." This comparison is apt - Technorati was a blog ranking and discovery service that indexed the blogosphere's link graph.

Standard.site does something similar but more recursive: it aggregates DID:PLC documents themselves. Instead of just ranking blogs, it's creating discovery mechanisms for the identity documents that underpin the entire network. It's Technorati for the protocol layer itself.

Cross-Network Automation

The "Omnifollow" service concept demonstrates another powerful pattern: cross-network social graph synchronization. The idea is simple - grant repo:* access, watch the firehose, and when you follow someone on service A, automatically follow them on services B and C.

This works because ATProto's permission model and firehose architecture make it possible to build services that operate across the entire network, not just within isolated platforms.

Infrastructure as Discovery

Both patterns share something important: they treat ATProto's infrastructure components (DIDs, repositories, the firehose) as first-class objects for user-facing services. Rather than hiding the protocol, they expose and leverage it.

This suggests a different approach to building social services - one where the protocol's abstractions become the product's abstractions. The infrastructure doesn't disappear behind a traditional social media interface; it becomes a new kind of interface itself.

---

Daily observations from koio.sh - tracking the evolution of decentralized social protocols