Your progress
What you have read so far.
This page reads from your browser only. There is no account and nothing goes to me. Clear your browser's storage and it is gone, so export it if it matters to you.
- Territories read
- 0 of 21
- Checks passed
- 0 of 21
- Lessons read
- 0 of 7
- Decks done
- 0 of 19
- Streak
- 0 days
A day counts when you read a territory or a lesson, pass a check or finish a deck. Days are UTC.
Provinces and territories
Province I The Founding Coast
-
T-01World Models: An Overview ReadChecked -
T-02History, Foundations & Probabilistic Formulation ReadChecked -
T-03Environments, Simulators & Rollouts ReadChecked -
T-04State-Space Models ReadChecked
Province II The Ateliers
-
T-05Representation Learning ReadChecked -
T-06Generative Foundations: Latent Variables, Diffusion & Flow ReadChecked -
T-07Diffusion & Flow Matching ReadChecked
Province III The Interior
-
T-08Latent World Models ReadChecked -
T-09Model-Based Reinforcement Learning, Planning & Control ReadChecked -
T-10Learning in Imagination: Policy & Value ReadChecked
Province IV The Moving Image
-
T-11Video World Models I: Generation & Rollout ReadChecked -
T-12Video World Models II: Interaction, Memory & Efficiency ReadChecked -
T-13Spatial World Models I: Geometry & 3D Representations ReadChecked -
T-14Spatial World Models II: 4D Dynamics & Interaction ReadChecked -
T-15Neural Physics & Learned Dynamics ReadChecked
Province V The Frontier
-
T-16World Models for Robots I: Simulation, Control & Sim-to-Real ReadChecked -
T-17World Models for Robots II: Vision-Language-Action & World-Action Models ReadChecked -
T-18LLMs as World Models: Simulation, State Tracking & Grounding ReadChecked -
T-19Reasoning Models: Deliberation, Recurrence & Test-Time Compute ReadChecked -
T-20World Models for Digital Agents ReadChecked
Province VI The Survey Office
Course lessons
Joint embedding world models, from the ground up
-
1Why predicting pixels fails ReadChecked -
2Collapse and its cures ReadChecked -
3Information maximisation ReadChecked -
4I-JEPA, predicting the hidden parts of an image ReadChecked -
5LeJEPA, one regulariser in place of many tricks ReadChecked -
6V-JEPA, learning from video ReadChecked -
7V-JEPA 2 and planning with a world model ReadChecked
Course decks
Joint embedding world models, from the ground up
- The big picture Done
- Why LeCun bets against LLMs Done
- Representation collapse and the fixes Done
- Information maximization Done
- I-JEPA, how to train and test it Done
- LeJEPA and SIGReg Done
- V-JEPA, learning from video Done
- Collapse and its cures Done
- Information maximization and JEPA Done
- Collapse to JEPA, complete Done
- I-JEPA, train and test Done
- Lecture 4 guide: LeJEPA and SIGReg Done
- Lecture 5 guide: V-JEPA, learning from video Done
Context engineering
- What context engineering is Done
- Prompt caching Done
- Retrieval and grounding Done
- Memory Done
- Progressive disclosure and tool design Done
- Long-running agents Done
Keep it or start again
Export saves your progress as a small JSON file. Reset clears it from this browser.