# Composable Labeling Trades One Standard for Parallel Capacity ## Claim Letting any account publish moderation labels and any app choose whether to honour them makes moderation capacity grow with the number of communities rather than with the central operator's headcount, at the cost of there no longer being one consistent standard or one party answerable for a bad call. <!-- atomicity: 1 idea ✓ --> ## Scope & conditions - The trade is explicit and not obviously wrong: parallel capacity for accountability concentration. - "No throat to choke" cuts both ways — it also means no single point that can be pressured into a bad removal. - Requires enough labelers to exist; early on, the composable system has less capacity than the centralized one it replaces. - Does not eliminate central moderation for illegal content, which still lands on the infrastructure operator. ## Evidence From the Decoder interview with Bluesky CEO Toni Schneider, where composable labeling is described as the moderation architecture. ## Links - [[Bootstrap-Then-Divest Is Only Credible With a Named Recipient]] — the same interview's governance claim, of which this is the operational half - [[Decentralization Without Audit Distributes Risk Not Accountability]] — the caution — distributing the work is not the same as distributing accountability - [[3. Resources/MOCs/Open Source Governance MOC]] — the governance thread on distributed moderation architecture. - [[3. Resources/MOCs/Commons & Community Governance MOC]] — the commons thread on distributing governance work.