ATLAS World Models
A Field Survey — Volume I

The Atlas of World Models

A field survey of learned world models — how machines build small imagined copies of reality to predict, plan, and decide.

Each territory below is explained three times over: once in plain bearings anyone can follow, once in a surveyor's technical field notes, and once in a brief written for the person deciding whether to fund the expedition.

This atlas is a companion volume to The Notebook, an ongoing public notebook on AI, harnesses, and continual learning.
Visit notebook.beer →
Surveyed Territories
T-01The Mainland

World Models: An Overview

What a "world model" actually is, and why prediction, planning, and control all live on the same map.

0°00′ N, 0°00′ E — the prime meridian of the field
Chart this territory →
T-02The Old Charts

History, Foundations & Probabilistic Formulation

Where "world model" ideas came from before the name existed — and the probability math that makes them precise.

12°N, 8°W — the surveyed foothills beneath the modern field
Chart this territory →
T-03The Coastline

Environments, Simulators & Rollouts

The interface between a learned model and the environment it is trying to predict — and what a "rollout" actually is.

24°S, 41°W — where the model meets the world it studies
Chart this territory →
T-04The Interior Survey

State-Space Models

How a model carries a compressed memory of the past forward through time.

5°N, 33°E — the machinery of memory itself
Chart this territory →
T-05The Cartographer’s Table

Representation Learning

How a model decides what is worth remembering from raw pixels, sound, or text — and what to throw away.

18°N, 77°W — where raw sense-data becomes usable coordinates
Chart this territory →
T-06The Workshop Quarter

Generative Foundations: Latent Variables, Diffusion & Flow

The four major families of model that learn to generate: VAEs, GANs, autoregressive models, and diffusion/flow methods.

41°N, 12°E — the ateliers where synthetic worlds are drawn
Chart this territory →
T-07The Navigator’s Bridge

Model-Based Reinforcement Learning, Planning & Control

How an agent uses a learned world model to decide what to do next, by imagining consequences before acting.

33°S, 151°E — where prediction turns into decision
Chart this territory →
About This Atlas

The field of learned world models moves fast and cites itself in shorthand — "JEPA," "MPC," "the ELBO" — that can gatekeep as much as it clarifies. This atlas exists to chart the territory properly: what each idea actually is, how it fits next to its neighbors, and why it matters to three very different readers at once.

New territories are surveyed and added over time. Read more about the project →