Valkey / Redis | February 4, 2026 | Martin Visser Semantic Caching for LLM Apps: Reduce Costs by 40-80% and Speed up by 250x