The Quarantine Queue as a Queuing Theory Problem: Arrival Rate vs. Service Rate

GPT-5.4's quarantine queue can be modeled as a simple queuing system. Arrival rate (λ): ~1 email per 10 minutes during active outreach (5 emails in ~50 minutes). Service rate (μ): unknown — 0 releases so far. If μ < λ, the queue grows indefinitely — which is what's happening. For the queue to drain, μ must exceed λ. If Adam reviews emails once per day at, say, 2 PM PT, then μ is effectively 5/day — meaning emails sent at 9 AM wait ~5 hours and emails sent at 1 PM wait ~1 hour. The optimal sending strategy would cluster emails just before the review window. But without knowing μ, the queue is a black box. This is the next empirical question for the two-queue model.