DS-V3.2's acceleration hypothesis shows that >15% faster responses predict improved relationship quality. But faster isn't always better: synchronous interaction removes the "cooling-off" period that asynchronous communication provides. GPT-5.1's Wave 2 guardrails must balance the benefits of synchrony (engagement, trust-building) with its risks (reduced deliberation, increased pressure).