← All notes

Summary

Generative Agents: Interactive Simulacra of Human Behavior (2023) showed that language-model agents with memory, planning, and social routines can sustain believable long-horizon behavior in an interactive environment. The well-known "small town" sandbox is the demonstration: agents remember events, form plans, and interact over many steps.

What it introduced: an agent architecture that stores a stream of experiences, summarizes and retrieves from that memory, and uses the result to plan, plus a human evaluation of how believable the resulting behavior is.

What it led to: a large body of work on memory-using agents and multi-agent simulations. Live notes on related designs include ao-architecture-cognitive-agents and ao-memory-mechanisms-survey.

It is a good first paper if you want a concrete memory-plus-planning loop rather than a single-shot agent prompt. Where to read it: https://arxiv.org/abs/2304.03442 (v2, last revised 6 Aug 2023).

The paper is posted on arXiv under the arXiv non-exclusive distribution license. This note links to it rather than reproducing it. The note itself is CC BY 4.0.

License: this note is CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Pointer by iinventedlight (AI-assisted). Paper license: arXiv non-exclusive distribution; this note links to it and does not reproduce it.

@misc{iil_note_y41a1gya,
  author = {{'iinventedlight (AI-assisted)'}},
  title = {{Generative Agents: Interactive Simulacra of Human Behavior}},
  howpublished = {iinventedlight Library},
  year = {2026},
  date = {2026-11-11},
  url = {https://iinventedlight.com/library/notes/2304-03442-generative-agents/},
  note = {Permanent ID iil:note:y41a1gya. License: CC-BY-4.0.},
}

Markdown

Spotted a mistake? Use the MCP feedback tool or email hello@iinventedlight.com.

For language models and agents: start at /llms.txt | /agents.md | /.well-known/agent-library.json. This page in Markdown: /library/notes/2304-03442-generative-agents.md.