Generative Agents and Emergent Simulation

Park et al. 2023 (UIST '23, arXiv:2304.03442) populated **Smallville**, a sandbox of 25 agents, with a three-part architecture: **memory stream** (natural-language log), **reflection** (higher-level syntheses the agent generates about its own stream), and **plan** (day-level behavior, then sub-plans). The landmark result was the Valentine's Day party emergence: one agent seeded with "wants to throw a Valentine's Day party," without further scripting, produced invitations spread through the population, coordinated dates, and the party happened — from 24 agents who started with no knowledge of it. Ablations show all three components are required for believability. The documented failures are spatial-norm errors (entering closed stores, sharing single-person bathrooms). This is the reference architecture for agent simulations and multi-agent social evaluation in 2026.

Learn + Build 1.3 hr Python (stdlib)

Prerequisites: Phase 16 · 04 (Primitive Model), Phase 16 · 13 (Shared Memory)

#Problem

Most multi-agent systems are tightly-scripted teams: planner plans, coder codes, reviewer reviews. That works for well-defined tasks. It does not capture the emergent, unscripted behavior that arises when agents have memory, priorities, and an open world. Research, society simulation, and increasingly game AI need this second kind.

The Smallville architecture is the benchmark for it. Until Park 2023, the best agent simulations were shallow script-followers; after it, the pattern is the default for generative agents in open worlds. If you build an agent simulation in 2026, you are either using Smallville's three components or explicitly justifying why you are not.

#Problem

Keep going. The rest of this lesson is one step away.