Speculative-Decoding4/21 - The Memory of a Conversation: KV, Prefix Reuse, Speculation, and ThroughputJuly 4, 20262/20 - Speculative Decoding: Let a Small Model Guess, Let a Large Model JudgeJune 11, 2026Draft Tokens or Smaller Numbers? Speculative Decoding vs Quantization in ProductionApril 16, 2026Speculative Decoding in Production: When Draft Tokens Help and When They HurtFebruary 27, 2026