---
title: "Why Autoregressive Generation Recomputes Work"
description: "Why naive autoregressive generation recomputes the whole sequence at every step, how the KV cache removes that waste, and how prefill differs from decode."
canonical_url: "https://fanout.sh/inference-eng/curriculum/03-01"
md_url: "https://fanout.sh/inference-eng/curriculum/03-01.md"
access: "public"
---

# Why Autoregressive Generation Recomputes Work

Why naive autoregressive generation recomputes the whole sequence at every step, how the KV cache removes that waste, and how prefill differs from decode.

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md