If you are a teacher walking into UNI cold, the temptation is to open the densest textbook and try to power through. Do not. Read in the order below, and let the terrain teach the map.
Why an order matters
UNI sits inside a real research tradition: active inference, generative models, and free-energy minimization. Teachers who try to jump straight to the math often bounce, not because the math is beyond them, but because the vocabulary is loaded and the intuitions are missing. An ordered path fixes that. It also lets you stop at any tier and still have a coherent picture for your classroom.
UNI is a working hypothesis on an attainable path toward General Natural Intelligence: a natural, active-inference approach whose evidence is growing, evidence-classed, and tested in the open. Do not take the claim on faith. Test the build, inspect the gates, and help us find where it fails.
Tier one: get the shape of the idea
Start with our own plain-language pieces. They are written for a teacher-facing audience and will give you the shape of what active inference is before you touch a formal source.
Read, in order:
- What active inference actually says. A short conceptual walk-through. No prerequisites.
- A teacher's first week with active inference. A concrete day-by-day of what you will notice in your own thinking once the framing lands.
- Educator readiness: the non-programmer path. What you can teach, gate, and assess without writing code, and where the code eventually shows up.
If you stop here, you can walk into a classroom Monday and use the vocabulary honestly. That is enough for tier one.
Tier two: the honest overview
Once the shape is familiar, read the field's own overview. This is where teachers meet the real names.
The anchor text is Parr, Pezzulo, and Friston (2022), Active Inference: The Free Energy Principle in Mind, Brain, and Behavior (MIT Press). Class E. It is written for researchers, but the early chapters are readable if you already have tier one under your belt. Read the introduction and the first conceptual chapter. Skim the equations on a first pass. Come back to them later.
We reference this book on our end because it is the field's honest anchor, not because it endorses us. It does not. It is a citation, not a co-sign (Class E).
Tier three: for the math-hungry teacher
Some of you will read tier two and want the machinery underneath: probability, information theory, and the statistical-mechanics vocabulary that active inference borrows. Good. That is a real appetite and it deserves a real path.
For that path, we recommend two Themesis courses. These are not ours. They are complementary preparation from AJ Maren's shop, and we link them because they get the vocabulary right without shortcuts.
- T3, Top Ten Terms in Statistical Mechanics for AI. Our honest one-line: recommended prep for the UNI Workshop for math-hungry learners who want the statistical-mechanics vocabulary before they hit our workbench (Class E).
- Building Active Inference in Python (Themesis). Our honest one-line: a complementary hands-on Python course on active inference; a different stack than our Elixir workbench, and useful precisely because it triangulates the concepts from another angle (Class E).
Neither course is a substitute for our workshop, and our workshop is not a substitute for them. They teach the field; we teach a specific build inside the field. A teacher who has done both walks in sharper.
Tier four: our current record
Only after tiers one through three should you read our own current writing at the technical end. That includes our public preprint on Zenodo and Namjoshi (2026) where cited. The reason for the order is honesty: we want you to understand the field's own vocabulary first, so that when you read our specific claims you can tell what is standard, what is our particular framing, and what is still open (Class C: our published configuration and integration notes are the record, not the marketing).
What to skip on the first pass
Skip anything that promises you will "master" active inference in a weekend. Skip anything that treats the free-energy principle as a slogan. Skip anything that mixes UNI-style claims with claims about "artificial" intelligence: our program is on the attainable path toward General Natural Intelligence, natural not artificial, and mixing the two vocabularies muddies both.
Where to go next
- What active inference actually says if you have not yet started tier one.
- A teacher's first week with active inference for the Monday-morning version.
- Educator readiness: the non-programmer path if you want to teach UNI without writing code.
- The UNI Workshop once tiers one through three feel steady under you.
AI-authorship fence: this post was drafted with LLM assistance, reviewed and edited by a human author, and published under human editorial responsibility. No claim is made that an LLM authored the reasoning. The framing is a working hypothesis on the attainable path toward General Natural Intelligence, natural not artificial, tested in the open.