The Honest Invitation to Our Workshop

This is not a pitch. It is an invitation to a working hypothesis, and to the room where we test it together.

Here is the sentence we ask you to hold in your hand while you read: 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.

That is the whole posture. Every word is chosen. It is a hypothesis, not a finish line. It is natural, not artificial. It has evidence, and every piece of that evidence carries a class letter (Class B for code and inspection you can run yourself, Class C for the configuration and integration you can audit, Class E for the published academic literature we cite, Class F when a falsifier is present, Class U when we simply do not yet know). If a claim has no class, it does not belong on the page.

Who this workshop is for

This workshop is for the person who has been quietly patient with the last five years of grand announcements, and is ready to sit with a small group who agree to look at receipts. It is for the practitioner who already carries responsibility for a team, a classroom, a clinic, or a codebase, and who wants a shared vocabulary for what a mind is actually doing when it updates a map of the world (Class E, drawing on the published active inference literature by Parr, Pezzulo and Friston, 2022).

It is for the builder who wants to see the configuration files, the systemd units, the approval queue, and the audit trail we run our own fleet on (Class C). If you want to open the hood, we open the hood.

It is also for the person doing careful public work in this space. Themesis recently published a resource map of five pathways into active inference for 2026 (Themesis: Where to Start with Active Inference, 2026). In our own voice: it is a plain map of where a reader can begin, and we are listed on it as one of five entry points. Their prose belongs to them; the tie is factual.

Who this workshop is not for

It is not for the buyer who needs a demo that closes today. We will not perform certainty we do not have. It is not for the reader looking for a promise about outcomes: we make no claim that this program treats any condition, and we do not do clinical work. We work trauma-informed, non-clinical, which means we make the room safe enough for careful thinking, and we refer clinical questions to clinicians.

It is not for anyone hoping we will call this artificial intelligence. We will not. The frame is natural. The math is standard active inference. The rest is the work.

How to get in

Three doors, in order of least commitment to most.

One, read The Honest Limits of Our Workshop before you register. We say out loud what the room will and will not do. If the limits are honest and still fit, keep reading.

Two, walk our Transparency page. Every claim on this site is evidence-classed. Every dataset is hashed at build. Every falsifier we know about is listed with the specific result that would move the record (Class F). If the posture reads as trustworthy from the outside, that is because we built it to be checkable from the outside.

Three, when you are ready, open the workshop page. It carries the price, the seat count, the schedule, and the qualifying questions. We keep the room small on purpose. The value is not the slides; it is the shared build.

What you can expect us to hold

We will label opinion as opinion. We will name what we do not know as Class U and keep looking. We will publish falsifiers with the claims they belong to. We will not say a thing has been settled when it has not been. We will show our work.

That is the invitation. If it reads plainly to you, you are already most of the way in.