0:44:42
Cutting latency in half: What actually worked—and what didn’t
DjangoCon US 2025 · November 6, 2025
Timothy McCurrach argues that performance work should begin with proactive, regular profiling rather than waiting for users to report problems or relying on intuition. Broad production measurements reveal patterns that isolated deep dives miss, while local profiling helps test changes; together they guide trade-offs and…